├── README.md ├── paper_figure ├── 1522721106458.png ├── 1522938283913.png ├── 1523241524030.png ├── 1523329925719.png ├── 1526027787768.png ├── 1526305327024.png ├── 1526389148656.png ├── 1531322452717.png ├── 1531501985565.png ├── 1533545264630.png ├── 1533558772424.png ├── 1533559666806.png ├── 1533632589963.png ├── 1533635551383.png ├── 1533713900014.png ├── 1533716802051.png ├── 1533785674947.png ├── 1533801963837.png ├── 1533803970846.png ├── 1533832651101.png ├── 1533909218190.png ├── 1533978994514.png ├── 1533991845908.png ├── 1533992512318.png ├── 1534147681755.png ├── 1534231184704.png ├── 1534231209288.png ├── 1534322692743.png ├── 1534324237488.png ├── 1534491739567.png ├── 1534491996294.png ├── 1534492052150.png ├── 1534492144193.png ├── 1534493095149.png ├── 1534493102024.png ├── 1534514760021.png ├── 1534820931384.png ├── 1534843008014.png ├── 1534859524747.png ├── 1534927153524.png ├── 1534941877505.png ├── 1535025690952.png ├── 1535080941866.png ├── 1535118132918.png ├── 1535424966773.png ├── 1535509143693.png ├── 1535529964676.png ├── 1535530049894.png ├── 1535630992315.png ├── 1535718654925.png ├── 1536304812187.png ├── 1536304821166.png ├── 1536311837927.png ├── 1536314355540.png ├── 1536411514566.png ├── 1536422986790.png ├── 1537101426419.png ├── 1537113167500.png ├── 1537156395887.png ├── 1537170968620.png ├── 1537502009449.png ├── 1537519750596.png ├── 1537523583894.png ├── 1537620570242.png ├── 1537856398909.png ├── 1537856407481.png ├── 1538214740187.png ├── 1538237284824.png ├── 1538364574274.png ├── 1539263939506.png ├── 1539311866259.png ├── 1546917876953.png ├── 1546918149839.png ├── 1546932031713.png ├── 1546937172649.png ├── 1546937279046.png ├── 1547042229954.png ├── 1547437042911.png ├── 1547437239206.png ├── 1547449080194.png ├── 1547719418272.png ├── 1547724806425.png ├── 1548039339341.png ├── 1550132357894.png ├── 1550132390502.png ├── 1550133510605.png ├── 1550145505105.png ├── 1550738798721.png ├── 1551686443341.png ├── 1551708715146.png ├── 1551708732566.png ├── 1552535525593.png ├── 1552545363095.png ├── 1552878709205.png ├── 1552893983832.png ├── 1552911506553.png ├── 1552912503121.png ├── 1553325398564.png ├── 1553325419379.png ├── 1553335463524.png ├── 1553335487753.png ├── 1553497729383.png ├── 1553516953870.png ├── 1554376060947.png ├── 1554865684235.png ├── 1555763466457.png ├── 1555765487674.png ├── 1555769164180.png ├── 1555770525391.png ├── 1556250344815.png ├── 1556250623415.png ├── 1556280764754.png ├── 1557370885604.png ├── 1557642537365.png ├── 1557673345371.png ├── 1557997523898.png ├── 1557997548672.png ├── 1558079145575.png ├── 1558494722437.png ├── 1558513853943.png ├── 1558519202625.png ├── 1558614526597.png ├── 1558614554155.png ├── 1558623033915.png ├── 1558962484711.png ├── 1558962790911.png ├── 1559037152377.png ├── 1559037213718.png ├── 1559037217373.png ├── 1559318184036.png ├── 1559318207399.png ├── 1559372865123.png ├── 1559378652026.png ├── 1559549453140.png ├── 1559551516676.png ├── 1559566271088.png ├── 1559919212428.png ├── 1560000228134.png ├── 1560135727490.png ├── 1560174931404.png ├── 1560175081156.png ├── 1560263818057.png ├── 1560263846164.png ├── 1560321278256.png ├── 1560321327144.png ├── 1560321404408.png ├── 1560328097317.png ├── 1560328134020.png ├── 1560329084025.png ├── 1560441378608.png ├── 1560668911327.png ├── 1560672462328.png ├── 1560684390577.png ├── 1560764288518.png ├── 1560779509995.png ├── 1561021159924.png ├── 1561195550827.png ├── 1561365990754.png ├── 1561383117771.png ├── 1561383159912.png ├── 1561385818759.png ├── 1561621592254.png ├── 1561707845616.png ├── 1561708816690.png ├── 1561714450649.png ├── 1561970185538.png ├── 1561973214857.png ├── 1562167485203.png ├── 1562229396844.png ├── 1562232295413.png ├── 1562405070274.png ├── 1562591190121.png ├── 1562596227030.png ├── 1564405750463.png ├── 1564457932100.png ├── 1564909153784.png ├── 1565180832447.png ├── 1565186648789.png ├── 1565186677190.png ├── 1565942057872.png ├── 1565946348305.png ├── 1568705054880.png ├── 1569071462061.png ├── 1569076126086.png ├── 1569251146102.png ├── 1570072565254.png ├── 1570362198248.png ├── image-20191106114728286.png ├── image-20191118003954646.png ├── image-20191122113840773.png ├── image-20191122113917775.png ├── image-20191123134134016.png ├── image-20191124230639274.png ├── image-20191126100315199.png ├── image-20191127151115458.png ├── image-20191128142955079.png ├── image-20191128153143279.png ├── image-20191128160533652.png ├── image-20191205195055060.png ├── image-20191209162702042.png ├── image-20191217204932349.png ├── image-20191217205029441.png ├── image-20191223162054823.png ├── image-20191224214842931.png ├── image-20191224214916387.png ├── image-20200130153405675.png ├── image-20200306224943511.png ├── image-20200308183445934.png ├── image-20200714012030378.png ├── image-20200714142649199.png ├── image-20200717161721279.png ├── image-20200717175301020.png ├── image-20200717223501471.png ├── image-20200718163048858.png ├── image-20200731155149662.png ├── image-20200914021803963.png ├── image-20200916145707212.png ├── image-20200916150327642.png ├── image-20200916152327699.png ├── image-20200918215222896.png ├── image-20200918215434370.png ├── image-20200918224123859.png ├── image-20200919014910600.png ├── image-20200923153618230.png ├── image-20200923153649390.png ├── image-20201029020046029.png ├── image-20201110121332322.png ├── image-20201110154834949.png ├── image-20201128184758932.png ├── image-20201128194505744.png ├── image-20201220224404400.png ├── image-20201221014051897.png ├── image-20210223233822225.png ├── image-20210223234507807.png ├── image-20210302202326671.png ├── image-20210304160351467.png ├── image-20210529210747569.png ├── image-20210604111024132.png ├── image-20210604120837514.png ├── image-20210928000837995.png ├── image-20210929143733993.png ├── image-20210929143735643.png ├── image-20210929150541422.png ├── image-20211103001651759.png ├── image-20211122232836344.png ├── image-20211123000856695.png ├── image-20211129000800270.png ├── image-20211129001332809.png ├── image-20211213011550369.png ├── image-20211213022458420.png ├── image-20211213022829073.png ├── image-20220116023316997.png ├── image-20220116030809631.png ├── image-20220225024908395.png ├── image-20220316134336877.png ├── image-20220316134407739.png ├── image-20220526171832531.png ├── image-20220615220333487.png ├── image-20220615221046364.png ├── image-20220912232446270.png ├── image-20220924003433635.png ├── image-20221127004323847.png ├── image-20221127011509981.png ├── image-20230402232037481.png ├── image-20230409234734929.png ├── image-20230410002227037.png ├── image-20230805235704150.png ├── image-20230806152854636.png ├── image-20230806200735934.png ├── image-20240305223437449.png ├── image-20240306000128456.png ├── image-20240306004919044.png ├── image-20240319012231775.png ├── image-20240321001530743.png ├── image-20240321002025742.png ├── image-20240321204347685.png ├── image-20240321210826877.png ├── image-20240805014222417.png └── image-20240807011652739.png └── storage_paper_note ├── CRUSH-SC'06.md ├── Ceph-OSDI'06.md ├── IDEA-FAST'24.md ├── Useful Matrial ├── ceph_dedup-ICDCS'18.md ├── deduplication ├── CodePlugin-HotCloud'15.md ├── cache_dedup │ ├── AustereCache-ATC'20.md │ └── LIPA-MSST'19.md ├── chunking │ ├── FastCDC-ATC'16.md │ ├── FrequencyBasedChunking-MASCOTS'10.md │ ├── RapidCDC-SoCC'19.md │ └── SSCDC-SYSTOR'19.md ├── dedup_estimation │ ├── ContentBasedSampling-ATC'13.md │ ├── EstimateDedupRatio-MSST'12.md │ ├── Rangoli-SYSTOR'13.md │ ├── SketchDeduplication-FAST'19.md │ └── UnseenDeduplication-FAST'16.md ├── dedup_in_nvm │ └── LightDedup-ATC'23.md ├── dedup_system_design │ ├── CloudTier-ATC'19.md │ ├── Dedup-ATC'11.md │ ├── DedupDesignTradeoff-FAST'15.md │ ├── DiskBottleneck-FAST'08.md │ ├── ExtremeBining-MASCOTS'09.md │ ├── InftyDedup-FAST'23.md │ ├── MFDedup-FAST'21.md │ ├── Redesigning-ATC'18.md │ ├── SparseIndex-FAST'09.md │ ├── dedupv1-MSST'10.md │ ├── iDedup-FAST'12-zryang-HP-Z440-Workstation-2.md │ └── iDedup-FAST'12-zryang-HP-Z440-Workstation.md ├── distributed_dedup │ ├── ClusterSingle-ToS'18.md │ ├── EDP-IWQoS'15.md │ ├── GoSeed-FAST'20.md │ ├── MigrationGame-FAST'22.md │ ├── Produck-SoCC'12.md │ └── TradeoffDataRouting-FAST'11.md ├── gc │ ├── AcceleratingRestore-ATC'14.md │ ├── GC-FAST'17.md │ └── MemorySanitization-FAST'13.md ├── mem_block_dedup │ ├── CAFTL-FAST'11.md │ ├── HintsDeduplication-FAST'16.md │ └── UKSM-FAST'18.md ├── metadate_management │ ├── FileRecipeCompression-FAST'13.md │ ├── MetadataHarmful-HotStorage'15.md │ ├── Metadedup.md │ └── REED-DSN'16.md ├── post_dedup │ ├── CompressionEst-FAST'13.md │ ├── DUPEFS-FAST'22.md │ ├── DedupSearch-FAST'22.md │ ├── DeepSketch-FAST'22.md │ ├── DeltaCompression-DCC'14.md │ ├── Finesse-FAST'19.md │ ├── Gdelta-CLUSTER'20.md │ ├── IDEA-FAST'24.md │ ├── LPC-SYSTOR'21.md │ ├── MeGA-ATC'22.md │ ├── MigratoryCompression-FAST'14.md │ ├── deltaStore-HotStorage'12.md │ └── deltaWAN-FAST'12.md ├── restore_performance │ ├── ALACC-FAST'18.md │ ├── HiDeStore-Middleware'20.md │ ├── ImproveRestore-FAST'13.md │ ├── LookBackWindow-FAST'19.md │ └── RevDedup-APSys'13.md ├── secure_dedup │ ├── BloomFilterDedup-ICSP'19.md │ ├── CAONT-RS-HotStorage'14.md │ ├── CDStore-ATC'15.md │ ├── ClearBox-CCS'15.md │ ├── DUPEFS-FAST'22.md │ ├── DistributedKeyGen-CCSW'14.md │ ├── DupLESS-Security'13.md │ ├── EnhancedThreshold-TDSC'16.md │ ├── FeatureSync-ICDCS'21.md │ ├── FrequencyAnalysis-DSN'17.md │ ├── IndependentServer-CCS'15.md │ ├── InformationLeakage-CCSW'16.md │ ├── MLE-EuroCrypto'13.md │ ├── MinHash-DSN'17.md │ ├── MitigatingSideChannel-IPDPS'18.md │ ├── PoWSGX-ACNS'20.md │ ├── PrivacyPreservingDedup-CLOUD'17.md │ ├── RARE-INFOCOM'18.md │ ├── S2Dedup-SYSTOR'21.md │ ├── SideChannel-S&P'10.md │ ├── SideChannelTradeOffs-AsiaCCS'17.md │ ├── TappingPotential-CNS'18.md │ └── UWare-DSC'18.md ├── summary │ └── 99DeduplicationProblem-HotStorage'16.md └── workload_analysis │ ├── BackupWorkloads-FAST'12.md │ ├── CapacityForecasting-LISA'11.md │ ├── PracticalDedup-FAST'11.md │ └── SimRedundancy-TC'18.md ├── erasure_coding ├── Background │ ├── Availability-OSDI'10.md │ ├── Erasure-Code-Summary.md │ ├── Facebook Warehouse Cluster-HotStorage'13.md │ └── Raft-ATC'14.md ├── Comparison │ └── LRC-Comparison-ATC'18.md ├── EC Update │ ├── CAU-ICPP'18.md │ └── ParIX-ATC'17.md ├── EC-System │ └── Giza-ATC'17.md ├── Erasure Coding Basics │ ├── Dimakis-TIT'09.md │ ├── Plank-FAST'09.md │ └── Plank-USENIX login'13.md ├── Erasure Coding Deployment │ ├── Alpha-Entanglement-Codes-DSN'18.md │ ├── ButterflyCode-FAST'16.md │ ├── DRC-ISIT'16.md │ ├── Galloper-ICDCS'18.md │ ├── Hitchhiker-SIGCOMM'14.md │ ├── LRC-ATC'12.md │ ├── NCCloud-FAST'12.md │ ├── Pyramid Codes-ToS'13.md │ ├── XORing-PVLDB'13.md │ └── product-matrix-MSR-FAST'15.md ├── JunLi's paper │ ├── Beehive-HotStorage'15.md │ ├── Carousel-ICDCS'17.md │ ├── JunLi-INFOCOM'14.md │ ├── Pipelined_Regeneration-NetCod'11.md │ └── Zebra-IWQoS'16.md └── Techniques for Erasure Coding │ ├── CAR-DSN'16.md │ ├── CORE-TC'15.md │ ├── Clay Codes-FAST'18.md │ ├── Degraded-First-Scheduling-DSN'14.md │ ├── EC Store-ICDCS'18.md │ ├── ECPipe-ATC'17.md │ ├── Encoding-aware-DSN'15.md │ ├── Lazy Means Smart-SYSTOR'14.md │ ├── OpenEC-FAST'19.md │ ├── ProactiveLatency-SoCC'17.md │ ├── RAFI-ATC'18.md │ ├── RotatedReed-SolomonCodes-FAST'12.md │ ├── SpeedupXOR-MSST'12 .md │ └── Two_Tale_in_HDFS-FAST'15.md ├── general_storage ├── BlockStorage │ ├── BlockHead-HotOS'21.md │ └── FEMU-FAST'18.md ├── Cache │ └── Cache-HotStorage'20.md ├── CloudStorageSystem │ ├── Cumulus-FAST'09.md │ ├── Duplicacy-ToCC'20.md │ ├── HDFS-MSST'10.md │ ├── Kurma-SYSTOR'19.md │ └── SPANStore-SOSP'13.md ├── Hash │ └── CompareByHash-ATC'06.md ├── OC_ZNS │ └── ZNS_SSD_Cache-HotStorage'24.md ├── PM │ ├── NoveLSM-ATC'18.md │ └── SLMDB-FAST'19.md ├── ceph_related │ └── BlueStore-SOSP'19.md ├── hpc_storage │ └── Lustre_BIO_DIO-FAST'24.md └── io_optimization │ └── BPFStorage-HotOS'21.md ├── paper_figure ├── 1522721106458.png ├── 1522938283913.png ├── 1523241524030.png ├── 1523329925719.png ├── 1526027787768.png ├── 1526305327024.png ├── 1526389148656.png ├── 1531322452717.png ├── 1531501985565.png ├── 1533545264630.png ├── 1533558772424.png ├── 1533559666806.png ├── 1533632589963.png ├── 1533635551383.png ├── 1533713900014.png ├── 1533716802051.png ├── 1533785674947.png ├── 1533801963837.png ├── 1533803970846.png ├── 1533832651101.png ├── 1533909218190.png ├── 1533978994514.png ├── 1533991845908.png ├── 1533992512318.png ├── 1534147681755.png ├── 1534231184704.png ├── 1534231209288.png ├── 1534322692743.png ├── 1534324237488.png ├── 1534491739567.png ├── 1534491996294.png ├── 1534492052150.png ├── 1534492144193.png ├── 1534493095149.png ├── 1534493102024.png ├── 1534514760021.png ├── 1534820931384.png ├── 1534843008014.png ├── 1534859524747.png ├── 1534927153524.png ├── 1534941877505.png ├── 1535025690952.png ├── 1535080941866.png ├── 1535118132918.png ├── 1535424966773.png ├── 1535509143693.png ├── 1535529964676.png ├── 1535530049894.png ├── 1535630992315.png ├── 1535718654925.png ├── 1536304812187.png ├── 1536304821166.png ├── 1536311837927.png ├── 1536314355540.png ├── 1536411514566.png ├── 1536422986790.png ├── 1537101426419.png ├── 1537113167500.png ├── 1537156395887.png ├── 1537170968620.png ├── 1537502009449.png ├── 1537519750596.png ├── 1537523583894.png ├── 1537620570242.png ├── 1537856398909.png ├── 1537856407481.png ├── 1538214740187.png ├── 1538237284824.png ├── 1538364574274.png ├── 1539263939506.png ├── 1539311866259.png ├── 1546917876953.png ├── 1546918149839.png ├── 1546932031713.png ├── 1546937172649.png ├── 1546937279046.png ├── 1547042229954.png ├── 1547437042911.png ├── 1547437239206.png ├── 1547449080194.png ├── 1547719418272.png ├── 1547724806425.png ├── 1548039339341.png ├── 1550132357894.png ├── 1550132390502.png ├── 1550133510605.png ├── 1550145505105.png ├── 1550738798721.png ├── 1551686443341.png ├── 1551708715146.png ├── 1551708732566.png ├── 1552535525593.png ├── 1552545363095.png ├── 1552878709205.png ├── 1552893983832.png ├── 1552911506553.png ├── 1552912503121.png ├── 1553325398564.png ├── 1553325419379.png ├── 1553335463524.png ├── 1553335487753.png ├── 1553497729383.png ├── 1553516953870.png ├── 1554376060947.png ├── 1554865684235.png ├── 1555763466457.png ├── 1555765487674.png ├── 1555769164180.png ├── 1555770525391.png ├── 1556250344815.png ├── 1556250623415.png ├── 1556280764754.png ├── 1557370885604.png ├── 1557642537365.png ├── 1557673345371.png ├── 1557997523898.png ├── 1557997548672.png ├── 1558079145575.png ├── 1558494722437.png ├── 1558513853943.png ├── 1558519202625.png ├── 1558614526597.png ├── 1558614554155.png ├── 1558623033915.png ├── 1558962484711.png ├── 1558962790911.png ├── 1559037152377.png ├── 1560441378608.png ├── 1562405070274.png ├── 1564405750463.png ├── 1564457932100.png └── image-20201029020046029.png ├── security ├── CompareByHash-ATC'06.md ├── Data-Encryption │ └── BlockEnc-HotStorage'22.md ├── DeRef-TrustCom'11.md ├── DifferentialPrivacy-INFOCOM'18.md ├── FADE-TDSC'12.md ├── FrequencyHiding-CCS'15.md ├── FrequencySmoothing-ICAR'17.md ├── NetworkSecurity │ └── DeRef-TrustCom'11.md ├── ORAM-CCSW'16.md ├── OverheadConfidentiality-UCC'19.md ├── QuantifyingInformationLeakage-CCSW'19.md ├── SGX-Storage │ ├── EnclaveCache-Middleware'19-zryang-HP-Z440-Workstation.md │ ├── SPEED-ICDCS'19.md │ ├── SPEICHER-FAST'19.md │ ├── SeGShare-DSN'20.md │ ├── ShieldStore-EuroSys'19.md │ └── StorageDataPathSGX-arxiv.md ├── SGX-Technique │ ├── SGXExplained-IACR'16.md │ ├── SGXPerf-Middleware'18.md │ ├── SGXPerformance-SIGMETRICS'19.md │ ├── SwitchLess-SysTEX'18.md │ └── talos-arxiv'17.md ├── Secret-Sharing │ ├── AONT-RS-FAST'11.md │ ├── SSSS-ACMCommunication'79.md │ └── SecretSharing-Summary.md ├── SecretSharing │ ├── AONT-RS-FAST'11.md │ ├── SSSS-ACMCommunication'79.md │ ├── SecretSharing-Summary.md │ └── ShareBigSecret-SYSTOR'18.md ├── SecureDeletion │ └── FADE-TDSC'12.md ├── ShareBigSecret-SYSTOR'18.md └── Splinter-NSDI'17.md ├── sketch └── Count-MinSketch.md ├── survey └── Deduplcation-Survey'18.md └── template.md /paper_figure/1522721106458.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1522721106458.png -------------------------------------------------------------------------------- /paper_figure/1522938283913.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1522938283913.png -------------------------------------------------------------------------------- /paper_figure/1523241524030.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1523241524030.png -------------------------------------------------------------------------------- /paper_figure/1523329925719.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1523329925719.png -------------------------------------------------------------------------------- /paper_figure/1526027787768.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1526027787768.png -------------------------------------------------------------------------------- /paper_figure/1526305327024.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1526305327024.png -------------------------------------------------------------------------------- /paper_figure/1526389148656.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1526389148656.png -------------------------------------------------------------------------------- /paper_figure/1531322452717.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1531322452717.png -------------------------------------------------------------------------------- /paper_figure/1531501985565.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1531501985565.png -------------------------------------------------------------------------------- /paper_figure/1533545264630.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533545264630.png -------------------------------------------------------------------------------- /paper_figure/1533558772424.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533558772424.png -------------------------------------------------------------------------------- /paper_figure/1533559666806.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533559666806.png -------------------------------------------------------------------------------- /paper_figure/1533632589963.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533632589963.png -------------------------------------------------------------------------------- /paper_figure/1533635551383.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533635551383.png -------------------------------------------------------------------------------- /paper_figure/1533713900014.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533713900014.png -------------------------------------------------------------------------------- /paper_figure/1533716802051.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533716802051.png -------------------------------------------------------------------------------- /paper_figure/1533785674947.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533785674947.png -------------------------------------------------------------------------------- /paper_figure/1533801963837.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533801963837.png -------------------------------------------------------------------------------- /paper_figure/1533803970846.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533803970846.png -------------------------------------------------------------------------------- /paper_figure/1533832651101.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533832651101.png -------------------------------------------------------------------------------- /paper_figure/1533909218190.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533909218190.png -------------------------------------------------------------------------------- /paper_figure/1533978994514.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533978994514.png -------------------------------------------------------------------------------- /paper_figure/1533991845908.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533991845908.png -------------------------------------------------------------------------------- /paper_figure/1533992512318.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1533992512318.png -------------------------------------------------------------------------------- /paper_figure/1534147681755.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534147681755.png -------------------------------------------------------------------------------- /paper_figure/1534231184704.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534231184704.png -------------------------------------------------------------------------------- /paper_figure/1534231209288.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534231209288.png -------------------------------------------------------------------------------- /paper_figure/1534322692743.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534322692743.png -------------------------------------------------------------------------------- /paper_figure/1534324237488.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534324237488.png -------------------------------------------------------------------------------- /paper_figure/1534491739567.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534491739567.png -------------------------------------------------------------------------------- /paper_figure/1534491996294.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534491996294.png -------------------------------------------------------------------------------- /paper_figure/1534492052150.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534492052150.png -------------------------------------------------------------------------------- /paper_figure/1534492144193.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534492144193.png -------------------------------------------------------------------------------- /paper_figure/1534493095149.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534493095149.png -------------------------------------------------------------------------------- /paper_figure/1534493102024.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534493102024.png -------------------------------------------------------------------------------- /paper_figure/1534514760021.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534514760021.png -------------------------------------------------------------------------------- /paper_figure/1534820931384.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534820931384.png -------------------------------------------------------------------------------- /paper_figure/1534843008014.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534843008014.png -------------------------------------------------------------------------------- /paper_figure/1534859524747.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534859524747.png -------------------------------------------------------------------------------- /paper_figure/1534927153524.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534927153524.png -------------------------------------------------------------------------------- /paper_figure/1534941877505.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1534941877505.png -------------------------------------------------------------------------------- /paper_figure/1535025690952.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535025690952.png -------------------------------------------------------------------------------- /paper_figure/1535080941866.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535080941866.png -------------------------------------------------------------------------------- /paper_figure/1535118132918.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535118132918.png -------------------------------------------------------------------------------- /paper_figure/1535424966773.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535424966773.png -------------------------------------------------------------------------------- /paper_figure/1535509143693.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535509143693.png -------------------------------------------------------------------------------- /paper_figure/1535529964676.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535529964676.png -------------------------------------------------------------------------------- /paper_figure/1535530049894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535530049894.png -------------------------------------------------------------------------------- /paper_figure/1535630992315.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535630992315.png -------------------------------------------------------------------------------- /paper_figure/1535718654925.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1535718654925.png -------------------------------------------------------------------------------- /paper_figure/1536304812187.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536304812187.png -------------------------------------------------------------------------------- /paper_figure/1536304821166.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536304821166.png -------------------------------------------------------------------------------- /paper_figure/1536311837927.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536311837927.png -------------------------------------------------------------------------------- /paper_figure/1536314355540.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536314355540.png -------------------------------------------------------------------------------- /paper_figure/1536411514566.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536411514566.png -------------------------------------------------------------------------------- /paper_figure/1536422986790.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1536422986790.png -------------------------------------------------------------------------------- /paper_figure/1537101426419.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537101426419.png -------------------------------------------------------------------------------- /paper_figure/1537113167500.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537113167500.png -------------------------------------------------------------------------------- /paper_figure/1537156395887.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537156395887.png -------------------------------------------------------------------------------- /paper_figure/1537170968620.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537170968620.png -------------------------------------------------------------------------------- /paper_figure/1537502009449.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537502009449.png -------------------------------------------------------------------------------- /paper_figure/1537519750596.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537519750596.png -------------------------------------------------------------------------------- /paper_figure/1537523583894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537523583894.png -------------------------------------------------------------------------------- /paper_figure/1537620570242.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537620570242.png -------------------------------------------------------------------------------- /paper_figure/1537856398909.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537856398909.png -------------------------------------------------------------------------------- /paper_figure/1537856407481.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1537856407481.png -------------------------------------------------------------------------------- /paper_figure/1538214740187.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1538214740187.png -------------------------------------------------------------------------------- /paper_figure/1538237284824.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1538237284824.png -------------------------------------------------------------------------------- /paper_figure/1538364574274.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1538364574274.png -------------------------------------------------------------------------------- /paper_figure/1539263939506.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1539263939506.png -------------------------------------------------------------------------------- /paper_figure/1539311866259.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1539311866259.png -------------------------------------------------------------------------------- /paper_figure/1546917876953.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1546917876953.png -------------------------------------------------------------------------------- /paper_figure/1546918149839.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1546918149839.png -------------------------------------------------------------------------------- /paper_figure/1546932031713.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1546932031713.png -------------------------------------------------------------------------------- /paper_figure/1546937172649.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1546937172649.png -------------------------------------------------------------------------------- /paper_figure/1546937279046.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1546937279046.png -------------------------------------------------------------------------------- /paper_figure/1547042229954.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547042229954.png -------------------------------------------------------------------------------- /paper_figure/1547437042911.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547437042911.png -------------------------------------------------------------------------------- /paper_figure/1547437239206.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547437239206.png -------------------------------------------------------------------------------- /paper_figure/1547449080194.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547449080194.png -------------------------------------------------------------------------------- /paper_figure/1547719418272.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547719418272.png -------------------------------------------------------------------------------- /paper_figure/1547724806425.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1547724806425.png -------------------------------------------------------------------------------- /paper_figure/1548039339341.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1548039339341.png -------------------------------------------------------------------------------- /paper_figure/1550132357894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1550132357894.png -------------------------------------------------------------------------------- /paper_figure/1550132390502.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1550132390502.png -------------------------------------------------------------------------------- /paper_figure/1550133510605.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1550133510605.png -------------------------------------------------------------------------------- /paper_figure/1550145505105.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1550145505105.png -------------------------------------------------------------------------------- /paper_figure/1550738798721.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1550738798721.png -------------------------------------------------------------------------------- /paper_figure/1551686443341.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1551686443341.png -------------------------------------------------------------------------------- /paper_figure/1551708715146.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1551708715146.png -------------------------------------------------------------------------------- /paper_figure/1551708732566.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1551708732566.png -------------------------------------------------------------------------------- /paper_figure/1552535525593.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552535525593.png -------------------------------------------------------------------------------- /paper_figure/1552545363095.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552545363095.png -------------------------------------------------------------------------------- /paper_figure/1552878709205.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552878709205.png -------------------------------------------------------------------------------- /paper_figure/1552893983832.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552893983832.png -------------------------------------------------------------------------------- /paper_figure/1552911506553.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552911506553.png -------------------------------------------------------------------------------- /paper_figure/1552912503121.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1552912503121.png -------------------------------------------------------------------------------- /paper_figure/1553325398564.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553325398564.png -------------------------------------------------------------------------------- /paper_figure/1553325419379.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553325419379.png -------------------------------------------------------------------------------- /paper_figure/1553335463524.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553335463524.png -------------------------------------------------------------------------------- /paper_figure/1553335487753.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553335487753.png -------------------------------------------------------------------------------- /paper_figure/1553497729383.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553497729383.png -------------------------------------------------------------------------------- /paper_figure/1553516953870.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1553516953870.png -------------------------------------------------------------------------------- /paper_figure/1554376060947.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1554376060947.png -------------------------------------------------------------------------------- /paper_figure/1554865684235.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1554865684235.png -------------------------------------------------------------------------------- /paper_figure/1555763466457.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1555763466457.png -------------------------------------------------------------------------------- /paper_figure/1555765487674.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1555765487674.png -------------------------------------------------------------------------------- /paper_figure/1555769164180.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1555769164180.png -------------------------------------------------------------------------------- /paper_figure/1555770525391.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1555770525391.png -------------------------------------------------------------------------------- /paper_figure/1556250344815.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1556250344815.png -------------------------------------------------------------------------------- /paper_figure/1556250623415.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1556250623415.png -------------------------------------------------------------------------------- /paper_figure/1556280764754.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1556280764754.png -------------------------------------------------------------------------------- /paper_figure/1557370885604.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1557370885604.png -------------------------------------------------------------------------------- /paper_figure/1557642537365.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1557642537365.png -------------------------------------------------------------------------------- /paper_figure/1557673345371.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1557673345371.png -------------------------------------------------------------------------------- /paper_figure/1557997523898.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1557997523898.png -------------------------------------------------------------------------------- /paper_figure/1557997548672.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1557997548672.png -------------------------------------------------------------------------------- /paper_figure/1558079145575.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558079145575.png -------------------------------------------------------------------------------- /paper_figure/1558494722437.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558494722437.png -------------------------------------------------------------------------------- /paper_figure/1558513853943.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558513853943.png -------------------------------------------------------------------------------- /paper_figure/1558519202625.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558519202625.png -------------------------------------------------------------------------------- /paper_figure/1558614526597.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558614526597.png -------------------------------------------------------------------------------- /paper_figure/1558614554155.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558614554155.png -------------------------------------------------------------------------------- /paper_figure/1558623033915.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558623033915.png -------------------------------------------------------------------------------- /paper_figure/1558962484711.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558962484711.png -------------------------------------------------------------------------------- /paper_figure/1558962790911.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1558962790911.png -------------------------------------------------------------------------------- /paper_figure/1559037152377.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559037152377.png -------------------------------------------------------------------------------- /paper_figure/1559037213718.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559037213718.png -------------------------------------------------------------------------------- /paper_figure/1559037217373.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559037217373.png -------------------------------------------------------------------------------- /paper_figure/1559318184036.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559318184036.png -------------------------------------------------------------------------------- /paper_figure/1559318207399.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559318207399.png -------------------------------------------------------------------------------- /paper_figure/1559372865123.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559372865123.png -------------------------------------------------------------------------------- /paper_figure/1559378652026.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559378652026.png -------------------------------------------------------------------------------- /paper_figure/1559549453140.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559549453140.png -------------------------------------------------------------------------------- /paper_figure/1559551516676.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559551516676.png -------------------------------------------------------------------------------- /paper_figure/1559566271088.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559566271088.png -------------------------------------------------------------------------------- /paper_figure/1559919212428.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1559919212428.png -------------------------------------------------------------------------------- /paper_figure/1560000228134.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560000228134.png -------------------------------------------------------------------------------- /paper_figure/1560135727490.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560135727490.png -------------------------------------------------------------------------------- /paper_figure/1560174931404.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560174931404.png -------------------------------------------------------------------------------- /paper_figure/1560175081156.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560175081156.png -------------------------------------------------------------------------------- /paper_figure/1560263818057.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560263818057.png -------------------------------------------------------------------------------- /paper_figure/1560263846164.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560263846164.png -------------------------------------------------------------------------------- /paper_figure/1560321278256.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560321278256.png -------------------------------------------------------------------------------- /paper_figure/1560321327144.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560321327144.png -------------------------------------------------------------------------------- /paper_figure/1560321404408.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560321404408.png -------------------------------------------------------------------------------- /paper_figure/1560328097317.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560328097317.png -------------------------------------------------------------------------------- /paper_figure/1560328134020.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560328134020.png -------------------------------------------------------------------------------- /paper_figure/1560329084025.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560329084025.png -------------------------------------------------------------------------------- /paper_figure/1560441378608.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560441378608.png -------------------------------------------------------------------------------- /paper_figure/1560668911327.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560668911327.png -------------------------------------------------------------------------------- /paper_figure/1560672462328.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560672462328.png -------------------------------------------------------------------------------- /paper_figure/1560684390577.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560684390577.png -------------------------------------------------------------------------------- /paper_figure/1560764288518.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560764288518.png -------------------------------------------------------------------------------- /paper_figure/1560779509995.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1560779509995.png -------------------------------------------------------------------------------- /paper_figure/1561021159924.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561021159924.png -------------------------------------------------------------------------------- /paper_figure/1561195550827.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561195550827.png -------------------------------------------------------------------------------- /paper_figure/1561365990754.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561365990754.png -------------------------------------------------------------------------------- /paper_figure/1561383117771.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561383117771.png -------------------------------------------------------------------------------- /paper_figure/1561383159912.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561383159912.png -------------------------------------------------------------------------------- /paper_figure/1561385818759.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561385818759.png -------------------------------------------------------------------------------- /paper_figure/1561621592254.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561621592254.png -------------------------------------------------------------------------------- /paper_figure/1561707845616.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561707845616.png -------------------------------------------------------------------------------- /paper_figure/1561708816690.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561708816690.png -------------------------------------------------------------------------------- /paper_figure/1561714450649.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561714450649.png -------------------------------------------------------------------------------- /paper_figure/1561970185538.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561970185538.png -------------------------------------------------------------------------------- /paper_figure/1561973214857.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1561973214857.png -------------------------------------------------------------------------------- /paper_figure/1562167485203.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562167485203.png -------------------------------------------------------------------------------- /paper_figure/1562229396844.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562229396844.png -------------------------------------------------------------------------------- /paper_figure/1562232295413.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562232295413.png -------------------------------------------------------------------------------- /paper_figure/1562405070274.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562405070274.png -------------------------------------------------------------------------------- /paper_figure/1562591190121.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562591190121.png -------------------------------------------------------------------------------- /paper_figure/1562596227030.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1562596227030.png -------------------------------------------------------------------------------- /paper_figure/1564405750463.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1564405750463.png -------------------------------------------------------------------------------- /paper_figure/1564457932100.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1564457932100.png -------------------------------------------------------------------------------- /paper_figure/1564909153784.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1564909153784.png -------------------------------------------------------------------------------- /paper_figure/1565180832447.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1565180832447.png -------------------------------------------------------------------------------- /paper_figure/1565186648789.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1565186648789.png -------------------------------------------------------------------------------- /paper_figure/1565186677190.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1565186677190.png -------------------------------------------------------------------------------- /paper_figure/1565942057872.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1565942057872.png -------------------------------------------------------------------------------- /paper_figure/1565946348305.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1565946348305.png -------------------------------------------------------------------------------- /paper_figure/1568705054880.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1568705054880.png -------------------------------------------------------------------------------- /paper_figure/1569071462061.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1569071462061.png -------------------------------------------------------------------------------- /paper_figure/1569076126086.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1569076126086.png -------------------------------------------------------------------------------- /paper_figure/1569251146102.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1569251146102.png -------------------------------------------------------------------------------- /paper_figure/1570072565254.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1570072565254.png -------------------------------------------------------------------------------- /paper_figure/1570362198248.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/1570362198248.png -------------------------------------------------------------------------------- /paper_figure/image-20191106114728286.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191106114728286.png -------------------------------------------------------------------------------- /paper_figure/image-20191118003954646.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191118003954646.png -------------------------------------------------------------------------------- /paper_figure/image-20191122113840773.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191122113840773.png -------------------------------------------------------------------------------- /paper_figure/image-20191122113917775.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191122113917775.png -------------------------------------------------------------------------------- /paper_figure/image-20191123134134016.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191123134134016.png -------------------------------------------------------------------------------- /paper_figure/image-20191124230639274.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191124230639274.png -------------------------------------------------------------------------------- /paper_figure/image-20191126100315199.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191126100315199.png -------------------------------------------------------------------------------- /paper_figure/image-20191127151115458.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191127151115458.png -------------------------------------------------------------------------------- /paper_figure/image-20191128142955079.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191128142955079.png -------------------------------------------------------------------------------- /paper_figure/image-20191128153143279.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191128153143279.png -------------------------------------------------------------------------------- /paper_figure/image-20191128160533652.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191128160533652.png -------------------------------------------------------------------------------- /paper_figure/image-20191205195055060.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191205195055060.png -------------------------------------------------------------------------------- /paper_figure/image-20191209162702042.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191209162702042.png -------------------------------------------------------------------------------- /paper_figure/image-20191217204932349.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191217204932349.png -------------------------------------------------------------------------------- /paper_figure/image-20191217205029441.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191217205029441.png -------------------------------------------------------------------------------- /paper_figure/image-20191223162054823.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191223162054823.png -------------------------------------------------------------------------------- /paper_figure/image-20191224214842931.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191224214842931.png -------------------------------------------------------------------------------- /paper_figure/image-20191224214916387.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20191224214916387.png -------------------------------------------------------------------------------- /paper_figure/image-20200130153405675.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200130153405675.png -------------------------------------------------------------------------------- /paper_figure/image-20200306224943511.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200306224943511.png -------------------------------------------------------------------------------- /paper_figure/image-20200308183445934.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200308183445934.png -------------------------------------------------------------------------------- /paper_figure/image-20200714012030378.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200714012030378.png -------------------------------------------------------------------------------- /paper_figure/image-20200714142649199.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200714142649199.png -------------------------------------------------------------------------------- /paper_figure/image-20200717161721279.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200717161721279.png -------------------------------------------------------------------------------- /paper_figure/image-20200717175301020.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200717175301020.png -------------------------------------------------------------------------------- /paper_figure/image-20200717223501471.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200717223501471.png -------------------------------------------------------------------------------- /paper_figure/image-20200718163048858.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200718163048858.png -------------------------------------------------------------------------------- /paper_figure/image-20200731155149662.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200731155149662.png -------------------------------------------------------------------------------- /paper_figure/image-20200914021803963.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200914021803963.png -------------------------------------------------------------------------------- /paper_figure/image-20200916145707212.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200916145707212.png -------------------------------------------------------------------------------- /paper_figure/image-20200916150327642.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200916150327642.png -------------------------------------------------------------------------------- /paper_figure/image-20200916152327699.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200916152327699.png -------------------------------------------------------------------------------- /paper_figure/image-20200918215222896.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200918215222896.png -------------------------------------------------------------------------------- /paper_figure/image-20200918215434370.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200918215434370.png -------------------------------------------------------------------------------- /paper_figure/image-20200918224123859.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200918224123859.png -------------------------------------------------------------------------------- /paper_figure/image-20200919014910600.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200919014910600.png -------------------------------------------------------------------------------- /paper_figure/image-20200923153618230.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200923153618230.png -------------------------------------------------------------------------------- /paper_figure/image-20200923153649390.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20200923153649390.png -------------------------------------------------------------------------------- /paper_figure/image-20201029020046029.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201029020046029.png -------------------------------------------------------------------------------- /paper_figure/image-20201110121332322.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201110121332322.png -------------------------------------------------------------------------------- /paper_figure/image-20201110154834949.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201110154834949.png -------------------------------------------------------------------------------- /paper_figure/image-20201128184758932.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201128184758932.png -------------------------------------------------------------------------------- /paper_figure/image-20201128194505744.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201128194505744.png -------------------------------------------------------------------------------- /paper_figure/image-20201220224404400.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201220224404400.png -------------------------------------------------------------------------------- /paper_figure/image-20201221014051897.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20201221014051897.png -------------------------------------------------------------------------------- /paper_figure/image-20210223233822225.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210223233822225.png -------------------------------------------------------------------------------- /paper_figure/image-20210223234507807.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210223234507807.png -------------------------------------------------------------------------------- /paper_figure/image-20210302202326671.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210302202326671.png -------------------------------------------------------------------------------- /paper_figure/image-20210304160351467.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210304160351467.png -------------------------------------------------------------------------------- /paper_figure/image-20210529210747569.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210529210747569.png -------------------------------------------------------------------------------- /paper_figure/image-20210604111024132.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210604111024132.png -------------------------------------------------------------------------------- /paper_figure/image-20210604120837514.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210604120837514.png -------------------------------------------------------------------------------- /paper_figure/image-20210928000837995.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210928000837995.png -------------------------------------------------------------------------------- /paper_figure/image-20210929143733993.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210929143733993.png -------------------------------------------------------------------------------- /paper_figure/image-20210929143735643.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210929143735643.png -------------------------------------------------------------------------------- /paper_figure/image-20210929150541422.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20210929150541422.png -------------------------------------------------------------------------------- /paper_figure/image-20211103001651759.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211103001651759.png -------------------------------------------------------------------------------- /paper_figure/image-20211122232836344.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211122232836344.png -------------------------------------------------------------------------------- /paper_figure/image-20211123000856695.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211123000856695.png -------------------------------------------------------------------------------- /paper_figure/image-20211129000800270.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211129000800270.png -------------------------------------------------------------------------------- /paper_figure/image-20211129001332809.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211129001332809.png -------------------------------------------------------------------------------- /paper_figure/image-20211213011550369.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211213011550369.png -------------------------------------------------------------------------------- /paper_figure/image-20211213022458420.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211213022458420.png -------------------------------------------------------------------------------- /paper_figure/image-20211213022829073.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20211213022829073.png -------------------------------------------------------------------------------- /paper_figure/image-20220116023316997.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220116023316997.png -------------------------------------------------------------------------------- /paper_figure/image-20220116030809631.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220116030809631.png -------------------------------------------------------------------------------- /paper_figure/image-20220225024908395.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220225024908395.png -------------------------------------------------------------------------------- /paper_figure/image-20220316134336877.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220316134336877.png -------------------------------------------------------------------------------- /paper_figure/image-20220316134407739.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220316134407739.png -------------------------------------------------------------------------------- /paper_figure/image-20220526171832531.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220526171832531.png -------------------------------------------------------------------------------- /paper_figure/image-20220615220333487.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220615220333487.png -------------------------------------------------------------------------------- /paper_figure/image-20220615221046364.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220615221046364.png -------------------------------------------------------------------------------- /paper_figure/image-20220912232446270.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220912232446270.png -------------------------------------------------------------------------------- /paper_figure/image-20220924003433635.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20220924003433635.png -------------------------------------------------------------------------------- /paper_figure/image-20221127004323847.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20221127004323847.png -------------------------------------------------------------------------------- /paper_figure/image-20221127011509981.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20221127011509981.png -------------------------------------------------------------------------------- /paper_figure/image-20230402232037481.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230402232037481.png -------------------------------------------------------------------------------- /paper_figure/image-20230409234734929.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230409234734929.png -------------------------------------------------------------------------------- /paper_figure/image-20230410002227037.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230410002227037.png -------------------------------------------------------------------------------- /paper_figure/image-20230805235704150.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230805235704150.png -------------------------------------------------------------------------------- /paper_figure/image-20230806152854636.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230806152854636.png -------------------------------------------------------------------------------- /paper_figure/image-20230806200735934.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20230806200735934.png -------------------------------------------------------------------------------- /paper_figure/image-20240305223437449.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240305223437449.png -------------------------------------------------------------------------------- /paper_figure/image-20240306000128456.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240306000128456.png -------------------------------------------------------------------------------- /paper_figure/image-20240306004919044.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240306004919044.png -------------------------------------------------------------------------------- /paper_figure/image-20240319012231775.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240319012231775.png -------------------------------------------------------------------------------- /paper_figure/image-20240321001530743.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240321001530743.png -------------------------------------------------------------------------------- /paper_figure/image-20240321002025742.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240321002025742.png -------------------------------------------------------------------------------- /paper_figure/image-20240321204347685.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240321204347685.png -------------------------------------------------------------------------------- /paper_figure/image-20240321210826877.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240321210826877.png -------------------------------------------------------------------------------- /paper_figure/image-20240805014222417.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240805014222417.png -------------------------------------------------------------------------------- /paper_figure/image-20240807011652739.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/paper_figure/image-20240807011652739.png -------------------------------------------------------------------------------- /storage_paper_note/CRUSH-SC'06.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | # CRUSH: Controlled, Scalable, Decentralized Placement of Replicated Data 5 | 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | SC'06 | Ceph | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | - distributing petabytes of data among tens or hundreds of thousands of storage devices 15 | - efficiently **map data objects to storage devices** 16 | - without relying on a **central** directory 17 | - main problem 18 | - facilitate the **addition and removal of storage** while **minimizing unnecessary data movement** 19 | 20 | ### CRUSH 21 | 22 | ### Implementation and Evaluation 23 | 24 | ## 2. Strength (Contributions of the paper) 25 | 26 | ## 3. Weakness (Limitations of the paper) 27 | 28 | ## 4. Some Insights (Future work) 29 | 30 | - object-based storage 31 | - unlike conventional block-based hard drives 32 | - object-based storage devices (**OSDs**) manage disk block allocation **internally**, exposing an interface that allows others to **read and write to variably-sized, named objects** 33 | - replace large block lists with **small object lists** and distributing the low-level block allocation problem 34 | -------------------------------------------------------------------------------- /storage_paper_note/Ceph-OSDI'06.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | # Ceph: A Scalable, High-Performance Distributed File System 5 | 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | OSDI'06 | deployed systems | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | ### Method Name 15 | 16 | ### Implementation and Evaluation 17 | 18 | ## 2. Strength (Contributions of the paper) 19 | 20 | ## 3. Weakness (Limitations of the paper) 21 | 22 | ## 4. Some Insights (Future work) 23 | -------------------------------------------------------------------------------- /storage_paper_note/Useful Matrial: -------------------------------------------------------------------------------- 1 | link: http://web.eecs.utk.edu/~plank/plank/www/software.html 2 | link: https://bitbucket.org/jimplank/ 3 | -------------------------------------------------------------------------------- /storage_paper_note/ceph_dedup-ICDCS'18.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | # Design of Global Data Deduplication for A Scale-out Distributed Storage System 5 | 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | ICDCS'18 | Ceph | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | ### Method Name 15 | 16 | ### Implementation and Evaluation 17 | 18 | ## 2. Strength (Contributions of the paper) 19 | 20 | ## 3. Weakness (Limitations of the paper) 21 | 22 | ## 4. Some Insights (Future work) 23 | -------------------------------------------------------------------------------- /storage_paper_note/deduplication/dedup_estimation/ContentBasedSampling-ATC'13.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Estimating Duplication by Content-based Sampling 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----: | :------------------: | 8 | | ATC'13 | Deduplication Sample | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | The benefit of deduplication in a primary storage system varies for different workloads. 14 | > For a certain workload that have a low level of deduplication, one would to turn off the deduplication feature to avoid its effect on **I/O performance** and to avoid the **metadata overhead** of deduplication. 15 | 16 | It is necessary for the estimator to allow customers to quickly estimate the deduplication benefit on their primary data. 17 | > existing deduplication estimators are either not fast enough or not accurate enough. 18 | 19 | ### Content-based Sample 20 | 21 | - Theory 22 | Assume a hash function that generates a fingerprint for each data block. The proposed **content-based sampling** applies a modulo-based filter to all the block fingerprints of a data set. 23 | > A block fingerprint passes the filter and is added to the sample 24 | > $Fingerprint$ Mod $M$ == $X$, $M$ is the filter divisor. 25 | 26 | **Idea**: split the fingerprint space into $M$ partitions, and to use one of the partitions in the estimation. 27 | 28 | > The estimation of distinct block size in the whole data set: $S^* = M \times S_{sample}$ 29 | > can also consider the case where the size of block is different. 30 | 31 | ![1557370885604](paper_figure/1557370885604.png) 32 | 33 | - How to set $M$? 34 | Define a $\alpha-\beta$ accuracy notion, to prove how to choose $M$ under the given $\alpha$ and $\beta$. 35 | 36 | ### Implementation and Evaluation 37 | - Changing logging module: samples data blocks during the consistency point. (for update) 38 | - Disk Scanner: sample existing blocks in the volume. 39 | - All the sample will store in a fingerprint sample file (FPS). 40 | - Estimation Operation: 41 | merges the sample from change logging to the FPS, and update the estimate accordingly. 42 | 43 | - Experiment 44 | 1. Accuracy 45 | 2. Performance degradation: increase the rate of random access for counting module. 46 | 47 | 48 | ## 2. Strength (Contributions of the paper) 49 | 1. this work provides a **single-scan-pass** method for estimating deduplication ratio in a large dataset, also, with statistically guaranteed accuracy. 50 | 2. This work also consider to tackle the rapid change of dataset. 51 | 3. It implements this technique in a real commercial storage system for real world dataset. It also verifies the impact of performance (less than a few percent). 52 | ## 3. Weakness (Limitations of the paper) 53 | 1. Theory of proving the correctness bound of is not very clear, I cannot understand the rationale behind them. 54 | 55 | ## 4. Future Works 56 | This paper provides a very simple sample method based on fingerprint, i.e., content-based, this is not very novel from my perspective. And this paper also presents the whole theory to support its idea. The insight is for large dataset, it is possible to estimate the deduplication ratio with sampling. -------------------------------------------------------------------------------- /storage_paper_note/deduplication/post_dedup/DUPEFS-FAST'22.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | DUPEFS: Leaking Data Over the Network With Filesystem Deduplication Side Channels 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | FAST'22 | secure deduplication | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | - 15 | 16 | ### Method Name 17 | 18 | ### Implementation and Evaluation 19 | 20 | ## 2. Strength (Contributions of the paper) 21 | 22 | ## 3. Weakness (Limitations of the paper) 23 | 24 | ## 4. Some Insights (Future work) 25 | 26 | -------------------------------------------------------------------------------- /storage_paper_note/deduplication/secure_dedup/CAONT-RS-HotStorage'14.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Convergent Dispersal: Toward Storage-Efficient Security in a Cloud-of Clouds 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :-----: | :------------------: | 8 | | HotStorage'14 | dispersal, deduplication | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | The keyless security of existing dispersal algorithms relies on embedded random information which breaks data deduplication of the dispersed data. This paper proposes convergent dispersal, which replaces the original random information with deterministic cryptographic hash which is derived from the original data. 14 | 15 | The Summary of existing dispersal algorithm 16 | ![1552893983832](paper_figure/1552893983832.png) 17 | 18 | 19 | ### Convergent Dispersal 20 | This paper contains two convergent dispersal algorithms 21 | > CRSSS and CAONT-RS 22 | > both of them augment existing dispersal algorithms with the deduplication property. 23 | 24 | - Target Deployment Scenario 25 | Each compute server performs the cross-user deduplication. 26 | 27 | The security of existing dispersal algorithms depends on the embedded **random information**. Due to randomness, distinct secrets with identical content lead to different sets of shares (impede data deduplication) 28 | 29 | - Key idea 30 | replace the original random information in existing dispersal algorithms with **deterministic hashes** generated from the secret. 31 | - Convergent RSSS (CRSSS) 32 | ![1552911506553](paper_figure/1552911506553.png) 33 | Given a secret, it splits it into a number of words which has the **same size** as hashes (e.g., SHA-256 $\rightarrow$ 32 bytes word size). 34 | Each time, it processes $k - r$ words, and generates $r$ hashes from the $k - r$ words uising $r$ different hashes. 35 | $$ 36 | h_i = H(D, i), for \space i = 0,1,...,r-1 37 | $$ 38 | 39 | - Convergent AONT-RS (CAONT-RS) 40 | Key idea: 41 | replace the random key employed in the AONT step of AONT-RS with a cryptographic hash generated from the secret. 42 | ![1552912503121](paper_figure/1552912503121.png) 43 | $$ 44 | c_i = d_i \oplus E(h_{key}, i), for \space i = 0, 1, ..., s-1 45 | $$ 46 | 47 | ### Implementation and Evaluation 48 | - Investigate the throughput of CRSSS and CAONT-RS 49 | > measure the total amount of processed secret data divided by the computational time of generating all shares. 50 | > SHA-256: for default hash 51 | > AES-256: for default encryption function 52 | 53 | - Evaluate this performance 54 | > In local machine 55 | 56 | Insight: CRSSS can achieve a more flexible tradeoff between security and performance than CAONT-RS. 57 | 58 | ## 2. Strength (Contributions of the paper) 59 | 60 | ## 3. Weakness (Limitations of the paper) 61 | 62 | ## 4. Future Works 63 | This paper does not analyze how the secret size affects the deduplication ratios for different dispersal algorithms. For CRSSS and CAONT-RS, since there are some constrains on the its secret sizes. 64 | -------------------------------------------------------------------------------- /storage_paper_note/deduplication/secure_dedup/DupLESS-Security'13.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | DupLESS: Server-Aided Encryption for Deduplicated Storage 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----------------: | :------------------: | 8 | | USENIX Security'13 | Secure Deduplication | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper: 13 | In cloud deduplication, Message-locked encryption (MLE) is inherently subject to brute-force attacks that can recover files falling into a know set. This paper proposes an key-server based architecture that can provides secure deduplicated storage resisting brute-force attacks. 14 | > The weakness of **convergent encryption** (CE): deterministic and keyless, security only is possible when the target message is too **large** to exhaust. (unpredictable) 15 | 16 | - Secure deduplication: Dedup with strong security against untrusted storage 17 | - Compromise resilience: under client compromise 18 | 19 | ### DupLESS (Duplicateless Encryption for Simple Storage) 20 | - It contains a group of affiliated clients, a key server 21 | > Goal: make DupLESS work **transparently** with existing Storage Service system. (Site as a layer on top of existing simple storage service interfaces) 22 | 23 | - Oblivious PRF (OPRF) protocol (Server-aided encryption) 24 | $F$: A pseudorandom function (PRF) 25 | $K \leftarrow F(K_s, H(f))$ ($f$ is the file) 26 | > OPRF: server learns nothing, client learns on $K$ 27 | 28 | - Against online brute-force attack: 29 | Goal: slow down online brute-force trials from attacker controlled clients 30 | > limit clients to send $q$ queries per epoch. 31 | 32 | - System Overview 33 | ![1551708732566](paper_figure/1551708732566.png) 34 | > Encrypt and decrypt files 35 | > Handle file names and paths 36 | > Run Transparently: low overhead, still can run when KS is down, no client-side state 37 | 38 | ### Implementation and Evaluation 39 | - DupLESS client: python, support **Dropbox** and **Google Drive**. (command-line interface) 40 | - Evaluation 41 | > 1. Bandwidth overhead: DupLESS bandwidth overhead compared to plain Dropbox 42 | > 2. Retrieval latency: vary file size, compare with plain Dropbox and Covergent Encryption. 43 | > 3. Storage Overhead: DupLESS storage overhead compared to dedup over plaintexts 44 | 45 | ## 2. Strength (Contributions of the paper) 46 | 1. The system design is very considerable. When authors design DupLESS, they consider deduplication, compromise resilience and brute-force attack resilience. 47 | 2. To achieve low performance overhead, this paper also considers the optimization in client-to-KS OPRF protocol. 48 | 3. To guarantee the performance, DupLESS aims to work as a transparent layer over the storage service system. 49 | ## 3. Weakness (Limitations of the paper) 50 | 1. Compared with plain Dropbox or Google Drive, DupLESS incurs the higher overhead on storage overhead (encrypted user key from user) so as to gain the improvement in security. 51 | ## 4. Future Works 52 | 1. How about do the deduplication over the metadata (e.g., the *key* used to encrypt data and be encrypted by user's private key)? Although the size of this part of data may be very small, they also store in back-end storage service. 53 | 2. Can it leveage the characteristic of the data workload to further optimize the deduplication process? 54 | -------------------------------------------------------------------------------- /storage_paper_note/deduplication/secure_dedup/FrequencyAnalysis-DSN'17.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Information Leakage in Encrypted Deduplication via Frequency Analysis 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----: | :------------------: | 8 | | DSN'17 | Secure Deduplication | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | Existing MLE implementations still cannot fully protect against content leakage, mainly because their encryption approaches are deterministic. 14 | Previous stduies have addressed the possibility of launching frequency analysis MLE-based storage and also proposed cryptographic mechanisms to mitigate the issue. 15 | > Those investigations are theoretically driven. 16 | 17 | ### Locality Based Attack 18 | - Chunk Locality 19 | This is very prevalent in backup workloads that chunks are likely to re-occur together with their neighboring chunks across backups. 20 | > rationale: changes to backups often appear in few clustered regions of chunks. 21 | 22 | In security perspective: if a plaintext chunk $M$ corresponds to a ciphertext chunk $C$, then the neighboring plaintext chunks of $M$ are likely to correspond to the neighboring ciphertext chunks of $C$. 23 | 24 | - Threat Model 25 | An attacker may obtain the auxiliary information (as the plaintext chunks of a prior backup). 26 | > Ciphertext-only mode: can only access the ciphertext chunks 27 | > Known-plaintext mode: a more powerful adversary that knows a small fraction of the ciphertext-plaintext chunks pairs about the latest backup. 28 | 29 | 30 | - Attack Model 31 | Message, ciphertext model: 32 | > $M_1, M_2, ....$, $C_1, C_2, ...$ shows the logical order for message chunks and ciphertext chunks. 33 | 34 | **Basic Attack**: 35 | It takes in $C$ and $M$ as input, and returns the result set of all inferred ciphertext-plaintext chunk pairs. 36 | > just counting the frequency of each kind of chunks, and sort them to generate ranking, and find the message and ciphertext chunks with the same ranking. 37 | 38 | This kind of attack is sensitive to data updates that occur across different versions of backups over time. 39 | > it exists many ties, i.e., chunks have the same frequency. How to break ties during sorting also affects the frequency rank. 40 | 41 | ### Implementation and Evaluation 42 | 43 | ## 2. Strength (Contributions of the paper) 44 | 45 | ## 3. Weakness (Limitations of the paper) 46 | 47 | ## 4. Future Works 48 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Background/Availability-OSDI'10.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Availability in Globally Distributed Storage System 5 | @OSDI'10 @Background 6 | [TOC] 7 | 8 | ## Summary 9 | 10 | ***Availability*** 11 | **Part 1** 12 | ![1537856407481](paper_figure/1537856407481.png) 13 | The vast majority of such unavailability events are transient and do not result in permanent data loss. Less than 10% of events last longer than 15 minutes. 14 | 15 | - Two metrics throughout this paper: 16 | 17 | 1. **The average availability** of all $N$ nodes in a cell 18 | $$ 19 | A_N = \frac{\sum_{N_i \in N}uptime(N_i)}{\sum_{N_i \in N}(uptime(N_i)+downtime(N_i))} 20 | $$ 21 | $uptime(N_i)$ and $downtime(N_i)$ refer to the lengths of time a node $N_i$ is available or unavailable. 22 | 23 | 2. **Mean time to failure (MTTF)**: 24 | the measurements of availability 25 | $$ 26 | MTTF = \frac{uptime}{number \quad failures} 27 | $$ 28 | 29 | **Failure Bursts** 30 | This paper presents a simple time-window-based method to group failure events into failure bursts which, despite its simplicity, successfully identifies bursts with a common cause. 31 | 32 | **Key findings from the fleetwide data** 33 | 1. Most unavailability periods are transient 34 | 2. Unavailability durations differ significantly by cause 35 | > Median duration between 1 and 30 minutes depending on the cause 36 | 37 | 3. Correlated failures matter 38 | > 37% of all failures of at least 15 minutes in duration in Google datacenters are part of a correlated failure burst. 39 | 40 | 4. Largest correlated failures share common failure domains 41 | > Large failure bursts almost always have significant **rack-corelation**. 42 | 43 | ## Feedback and Recommendations 44 | 1. Determining the acceptable rate of successful transfers to battery power for individual machines upon a power outage. 45 | 2. Focusing on reducing reboot times, because planned kernel upgrades are major source of related failures. 46 | 3. Moving towards a dynamic delay before initiating recoveries, based on failure classification and recent history of failures on the cell. 47 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Background/Facebook Warehouse Cluster-HotStorage'13.md: -------------------------------------------------------------------------------- 1 | # A Solution to the Network Challenges of Data Recovery in Erasure-coded Distributed Storage Systems: A study on the Facebook Warehouse Cluster 2 | @HotStorage'13 @Study on Data-center Network @Note 3 | [TOC] 4 | 5 | ## Summary 6 | ***Background from Facebook's warehouse cluster***: 7 | - The warehouse cluster at Facebook employs an RS code with parameters ($k=10, r=4$), thus resulting in a $1.4 \times$ storage requirement, as compared to $3 \times$ under conventional replication, for a similar level fo reliability. 8 | - The most frequently accessed data is stored as 3 replicas, to allow for efficient scheduling of the map-reduce jobs. 9 | - For the data which has not been accessed for more than three months is stored as a $(10,4)$ RS code. 10 | 11 | ***Data Recovery*** 12 | - In Facebook warehouse cluster, the median is more than 50 machines-unavailability events per day. 13 | > This reasserts the necessity of redundancy in the data for reliability and availability. 14 | 15 | - Number of missing blocks in a stripe: 16 | > one block missing: 98.08% (most common) 17 | > two blocks missing: 1.87% 18 | > three or more blocks missing: 0.05% 19 | 20 | - Cross-rack bandwidth consumed: 21 | > A median of more than 180TB of data is transferred through the TOR switches every day for RS-coded data recovery. -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/EC Update/ParIX-ATC'17.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | PARIX: Speculative Partial Writes in Erasure-Coded Systems 5 | -------------------------------- 6 | 7 | @ATC'17 @Erasure Coded writes 8 | [TOC] 9 | 10 | ## 1. Summary 11 | ### Motivation of this paper: 12 | - This paper aims to solve the problem of partial write in EC. The main problem of partial write of EC is the high I/O amplification. Although this can can be mitigate by leveaging parity logging to append incremental change log (update the delta of parity), it agrues the cost of **in-place read-and-write** is still expensive (its latency is equivalent to that of random seek). 13 | 14 | - This paper wants to further optimize parity logging by eliminating the read. (i.e., the key to improve the performance of partial writes is to reduce the number of reads) 15 | 16 | ### PariX 17 | - The key idea of PariX 18 | > 1. Each parity records $m$ series of changes logs of $m$ data blocks. 19 | > > The rationale: $d^{(0)}_i \leftarrow d^{(1)}_i, d^{(2)}_i, d^{(3)}_i, ... ,d^{(r)}_i​$, 20 | > > $p_j^{(r)} = p_j^{(0)} + a_{ij} \times (d_i^{(r)} - d_i^{(0)})​$ 21 | > 2. The Speculation: 22 | > > In this paper, its speculation means whether the parities need $d_i^{(0)}$ or not. 23 | > > it speculates that 24 | > > 1) Assume $d_i^{(0)}$ is NOT needed (mostly right) 25 | > > 2) Send $d_i^{(0)}$ only when it actually needed (sometimes only) 26 | 27 | - PariX's overhead is a disk write, while previous parity log schemes's overhead is **a disk write after a disk read**, assuming the read cache is missed. 28 | 29 | - The whole workflow of PariX: 30 | 31 | ![1547042229954](paper_figure/1547042229954.png) 32 | 33 | **Note that**: For the case in replication, it is also (1 loop + 1 RTT) 34 | 35 | ### Implementation and Evaluation 36 | - Based on a previous block store of this group, master server is implemented by **MySQL** and **Redis**. (refer to Usra (Open Source partly)) 37 | - Random I/O lantency 38 | - Recovery time 39 | 40 | ## 2. Strength (Contributions of the paper) 41 | - the idea of this paper is not very complex, but can really work. Reducing the time of read is an intitutive method to impove the performance of writing. 42 | - the method can truly work in the scenario of 43 | ## 3. Weakness (Limitations of the paper) 44 | - When there are too many one-shot write in a workload, PariX cannot work well. Because it cannot fit its speculation. 45 | - When PariX speculation fails, it needs to send $d_i^{(0)}$ to all parities, this penatly would introduce high network traffic overhead. (Although it argues that parities writes are small) 46 | - In its evaluation, it just test this prototype in a small testbed (with 10 machines), with EC (4, 2). This setting cannot prove the scalability of it. 47 | 48 | ## 4. Future Works 49 | - In this method, its speculation is very simple, which cannot fit the one-shot write well. How to handle the large sequential one-shot write workload? Maybe it is a point can be extended. 50 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Basics/Plank-FAST'09.md: -------------------------------------------------------------------------------- 1 | #A Performance Evaluation and Examination of Open-Source Erasure Coding Libraries For Storage 2 | @James S. Plank @FAST'09 3 | [TOC] 4 | 5 | ## Summary 6 | - This paper's goals are to 7 | > 1. compare codes and implementations to discern whether **theroy matches practice** 8 | > 2. demonstrate how parameter selection can lead to good coding performance 9 | 10 | It firstly demonstrates the **Nomenclature** of a Erasure Coded Storage System, then gives brief introductions of different kinds of Erasure Codes 11 | > 1. *Reed-Solomon (RS) Codes*: strip unit is a $w-bit$ word, where $w$ must be large enough that $n \leq 2^w+1$ (Most implementations find $w=8$ performs the best). The multiplication in $GF(2^w)$ implemented with multiplication tables or discrete logarithm tables so that it is more **complex**. For this reason, Reed-Solomon codes are considered expensive. 12 | > 2. *Cauchy Reed-Solomon (CRS) Codes*: this code extends RS codes in two points. First, the Generatir matrix uses Cauchy matrices instead of Vandermonde matrices. Second, 13 | it eliminates the expensive multiplications of RS codes by converting them to extra XOR operations. 14 | 15 | Next, this paper tests five open source erasure coding libraries (i.e., **Luby**, **Zfec**, **Jerasure**, **Cleversafe**, **EBENODD/RDP**) and do the **Encoding Experiment** and **Decoding Experiment** to measure their performances. When it does the experiment, it uses speeds of **memcpy()** and **XOR** as the baseline. 16 | > 1. Encoding Time VS. Data Buffer Size: using **gettimeofday()** to measure the times of all coding activities. 17 | > **Conclusion**: the size of the data buffer can impact performance, with the I/O removed, the effects of modifying the buffer size are negligible. 18 | > 2. Encoding Speed VS. Packet Size: **lower** packet sizes have less tight XOR loops, but better cache behavior. **higher** packet sizes perform XORs over larger regions, but cause more cache misses. 19 | > **Conclusion**: higher packet sizes perform better than lower ones. It also performs **a search algorithm** to find optimal packet size. 20 | 21 | In addition, it also mentions that **the unit of XOR** used by the encoding/decoding software should match the the largest possible XOR unit of the machine. 22 | 23 | In Conclusion, it summaries: 24 | > 1. how to choose $w$ in RAID-6 code, CSR, RS. 25 | > 2. the packets sizes of the codes should be chosen to yield **good cache behavior**. 26 | > 3. the implementation must pay attention to **memory and cache**. It will be paramount for high performance. 27 | 28 | ## Strength 29 | 1. This paper presents a full evaluation and examination of the current Open-Source Erasure Coding Libraries in Storage System, which can help the readers get a great perspective of erasure codes. 30 | ## Weakness 31 | Because this paper just evaluuates the Open-Source Erasure Coding Libraries, so it cannot say that there exists obvious weakness in this paper. 32 | ## Future Work 33 | 1. This paper mentions that the modern architectures shift more universally toward **multicore**. So it can further investigate how open source libraries to exploit the opportunities of multiple processors on a board. (this is an additional challenge) 34 | 2. More research needs to be performed on **special-purpose codes** beyond RAID-6, and implementations need to take advantage of the special-purpose codes that already exist. -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Deployment/Alpha-Entanglement-Codes-DSN'18.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Alpha Entanglement Codes: Practical Erasure Codes to Archive Data in Unreliable Environments 5 | @DSN'18 @ Redundancy Propagation 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: this paper intends to design flexible and practical erasure codes with high fault-tolerance to improve data durability and availability even in catastrophic scenarios. 10 | 11 | ***Alpha Entanglement Codes (AEC) $(\alpha, s, p)$***: The idea of AE code is to create redundancy by **tangling (mixing)** new data and blocks with old ones, building entangled data chains that are woven into a growing mesh of interdependent content. 12 | >1. $\alpha$: determine the local connectivity, the number of parities created per data block. 13 | >2. $s$ amd $p$: determin the global connectiviry of data blocks in the grid. $p$ defines the number of helical strands. $s$ the number of horizontal strands 14 | 15 | Each node belongs to $\alpha$ strands each edge belongs to only one strand. 16 | 17 | - Alpha Entanglements permit changes in the paramenters without the need to encode the content again (**dynamic** fault-tolerance, long-term storage system) 18 | - The encoder builds block chains, strands, that alternate data and redundant blocks. 19 | > The entanglement function computes the exclusive-or (XOR) of two consecutive blocks at the head of a strand and inserts the output adjacent to the last block. 20 | 21 | - The interwined graph of AEC: 22 | ![1535630992315](paper_figure/1535630992315.png) 23 | ![1535718654925](paper_figure/1535718654925.png) 24 | 25 | - The decoder repairs a node using two adjacent edges that belong to the same strand, thus, there are $\alpha$ options. 26 | 27 | ***Implementation and Evaluation***: 28 | **Implementation**: It mentions two use cases of entangled storage system 29 | >1. A Geo-Replicated Backup 30 | >2. Disk Arrays 31 | 32 | **Evaluation**: It examines the design of AE codes to understand how the code settings impact on fault tolerance and write performance. 33 | >1. Code Parameters and Write Performance 34 | >2. Code Parameters and Fault Tolerance 35 | 36 | ## Strength (Contributions of the paper) 37 | 1. These mechanisms based on the novel redundancy propagation, which are robust and sound, but they are also efficient and easily implementable in real system. 38 | 2. The encoder and decoder are lightweight, based on exclusive-or operations, and offer promising trade-offs between security, resource usage and performance. 39 | 3. This paper also mentions the use cases for distributed and centralised storage systems. 40 | ## Weakness (Limitations of the paper) 41 | 1. This paper does not mention how to improve the efficiency of repairs. It just focuses on reducing the maximum possible storage overhead. 42 | 43 | ## Future Works 44 | 1. A point that can be considered is how to improve its efficiency of repairs. -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Deployment/ButterflyCode-FAST'16.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Opening the Chrysalis: On the Real Reapir Performance of MSR Codes 5 | @FAST'16 @MSR Code 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: The implementation of RGCs in production systems requires dealing with complex and bug-prone algorithms. In this paper, it focuses on managing the drawbacks of RGC-MSR codes. 10 | 11 | ***Butterfly Codes***: 12 | **Encode**: The key point in Butterfly Codes is its **recursive** construction approach, which can have a simplified design that results in a simpler implementation, and allow for a better reuse of precomputed values, leading to better cache locality. 13 | ![1534324237488](paper_figure/1534324237488.png) 14 | Because of the double vertical flip in equations above, the recursion can be simplified 15 | ![1534322692743](paper_figure/1534322692743.png) 16 | 17 | **Decode**: In the case of one failure, the lost column can be regenerated by communicating an amount of data equal to 1/2 of the remaining data. 18 | > If the lost column is not the butterfly parity, the amount of communicated data is exactly equal to the amount read from suriving disk. (**optimal I/O access**) 19 | 20 | ***Implementation and Evaluation***: 21 | **Ceph**: RADOS uses an erasure code plug-in infrastructure that allows dynamical use of external erasure code libraries. Third-party developers can provide independent erasure code implementation.Efficient integration of Butterfly code with Ceph requires **re-defining** the plug-in interface. It also implements a **proxy plug-in** that dynamically links with existing plug-ins. 22 | 23 | **Evaluation**: In HDFS and Ceph 24 | >1. Repair Throughput: 25 | >2. CPU Utilization 26 | >3. Network Traffic 27 | >4. Storage Traffic 28 | 29 | ## Strength (Contributions of the paper) 30 | 31 | 1. design a recursive Butterfly code construction (two parity MSR code) and implement it in two real-world distributed storage systems (HDFS and Ceph). 32 | 2. compare two major approaches when using erasure codes in distributed storage systems (**online** and **batch-based** encoding) and point the major tradeoffs between the two approaches. 33 | 3. examine the performance of Butterfly code and draw a comparison between the theoretical results of MSR codes and the performance achievable in real systems. 34 | 35 | ## Weakness (Limitations of the paper) 36 | 1. Butterfly Code is not very general, it can just tolerate that any two of the codeword columns are missing 37 | 38 | ## Future Work 39 | 1. How to make Butterfly codes more general is another issue that can be extended in the future. 40 | 41 | 42 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Deployment/DRC-ISIT'16.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Double Regenerating Codes for Hierarchical Data Centers 5 | @ISIT'16 @Regenerating Codes 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: Deploying erasure coding in data center remains challenging due to the hierarchical nature of data centers (multiple racks, each comprising multiple nodes for storage). And the cross-rack bandwidth is heavily oversubscribed. Thus, this paper's goal is to minimize the cross-rack repair bandwidth in hierarchical data center. 10 | 11 | ***Double Regenerating Codes*** 12 | The core idea of DRC is to perform regeneration **twice**: first within a rack and the across multiple racks. So it calls this approach **double regeneration**. 13 | > repair design targets the unbalanced nature of inner-rack and cross-rack capacities in hierarchical data centers. 14 | ![1537620570242](paper_figure/1537620570242.png) 15 | 16 | - Double Regeneration makes two trade-offs: 17 | >1. It can only tolerate a single-rack failure 18 | >2. the sum of the inner-rack and cross-rack repair bandwidth is higher than that in MSR codes. (trade the inner-rack repair bandwidth for the cross-rack repair bandwidth) 19 | 20 | - DRC exploits node cooperation within a rack to minimize the cross-rack repair bandwidth. 21 | 22 | ***Implementation and Evaluation***: 23 | **Implementation** 24 | None 25 | **Evaluation** 26 | Compared with RS, MSR, IEE 27 | 28 | ## Strength (Contributions of the paper) 29 | 1. prove that there exists a DRC construction that minimizes the cross-rack repair bandwidth for a single-node repair , while preserving the MDS property. 30 | 2. show via quantitative comparisons that DRC reduces the cross-rack repair bandwidth of state-of-the-art MSR codes by up to 45.5%. 31 | ## Weakness (Limitations of the paper) 32 | 1. DRC can only tolerate a single-rack failure (as opposed to the nomal code) 33 | ## Future Works 34 | 1. How about leveraging the special topological structure of rack-based data center to further reduce the repair overhead? -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Deployment/NCCloud-FAST'12.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # NCCloud: A Network-Coding-Based Storage System in a Cloud-of Clouds 5 | @FAST'12 @ Regenerating Code 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: It is important to do the repair in multiple cloud storage. This paper's objective is to minimize the cost of storage repair (due to the maintain of data over the clouds) for a permanent sing-cloud failure. 10 | 11 | 12 | ***Functional Minimum-Storage Regenerating Code (FMSR)***: 13 | This paper present a **proxy-based**, multiple-cloud storage system, called NCCloud, which can practically address the reliability of today's cloud backup storage. NCCloud implements FMSR, which eliminates the encoding requirement if storage nodes during repair, while ensuring the new set of stored chunks after each round of repair preserves the required fault tolerance. 14 | ![1534231209288](paper_figure/1534231209288.png) 15 | Code chunk $P_i$ equals to linear combination of original data chunks. For the repair in FMSR, it downloads one code chunk from each surviving node and reconstructs new code chunks (via random linear combination) in new node. 16 | However, FMSR codes is not systematic, FMSR codes are acceptable for **long-term** archival applications, where the read frequency is typically low. 17 | 18 | **Iterative Repairs**: 19 | FMSR codes regenerates different chunks in each repair, so it is necessary to ensure **MDS property** still holds even after iterative repairs. To achieve this, it also proposes a **two-phase** checking method. 20 | >1. **MDS property check**: current repair maintains MDS property. 21 | >2. **Repair MDS property check**: next repair for any possible failure maintains MDS property. 22 | 23 | 24 | ***Implementation and Evaluation***: 25 | It implements NCCloud as a proxy that bridges user applications and multiple clouds including three layers: 26 | >1. file system layer 27 | >2. coding layer (both RAID-6 and FMSR) 28 | >3. storage layer 29 | 30 | **Evaluation**: 31 | Response time: both in Local Cloud (OpenStack Swift) and Commercial Cloud (multiple containers in Azure) with different file sizes. 32 | 33 | ## Strength (Contributions of the paper) 34 | 1. design FMSR code that can be deployed in thin-cloud setting as they do require storage nodes to perform encoding during repair. 35 | 2. implement FMSR code with a two-phase checking scheme, which ensures that double-fault tolerance is maintained in the current and next round of repair. 36 | 3. conduct monetary cost analysis and extensive experiments on both local cloud and commercial cloud settings. 37 | 38 | ## Weakness (Limitations of the paper) 39 | 1. FMSR code is not a systematic code that it only stores encoded chunks formed by the linear combination of the original data chunks, and does not keep the original data. Thus, the overhead of frequent reading is high. 40 | 2. In this paper, it just considers an FMSR code implementation with double-fault tolerance. 41 | 42 | ## Future Work 43 | 1. consider how to reduce the encoding overhead of FMSR code 44 | 2. make the FMSR code more general, beyond double-fault tolerance 45 | 46 | 47 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Erasure Coding Deployment/XORing-PVLDB'13.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # XORing Elephants: Novel Erasure Codes for Big Data 5 | @PVLDB'13 @New Erasure Code @Repair problem 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: This paper wants to overcome the high repair cost of RS code. It presents a novel family of erasure codes that are efficiently repairable and offer higher reliability compared to Reed-Solomon codes. 10 | 11 | ***Locally Repairable Codes (LRC)***: 12 | - The main idea of this code is to sacrifice storage efficiency to gain well performance. And it is inspired by "MDS code with parameters $(k, n-k)$ cannot have locality smaller than $k$". Because it is exactly the cost of **optimal fault tolerance**. Thus, it tries to construct a **near-MDS** code with non-trivial locality. 13 | - LRCs are constructed on the top of MDS codes: MDS encoded blocks are grouped in logarithmic sized sets and then are combined together to obtain parity blocks or **logarithmic** degree. For the distance: 14 | $$ 15 | \lim\limits_{k \rightarrow \infty} \frac{d_{LRC}}{d_{MDS}}=1 16 | $$ 17 | It makes repair efficient by adding additional **local parities**. 18 | ![1534514760021](paper_figure/1534514760021.png) 19 | 20 | ***Implementation and Evaluation***: 21 | Its system is a modification of **HDFS-RAID** that incorporates LRC. The **RaidNode** and **BlockFixer** classes were also subject to modifications. 22 | **Evaluation**: both in Amazon's Elastic Compute Cloud (EC2) and a test cluster in Facebook 23 | >1. HDFS Bytes Read: total amount of data read 24 | >2. Network Traffic: total amount of data communicated from nodes (Amazon's AWS Cloudwatch monitoring tools) 25 | >3. Repair Duration: the time interval between the starting time of the first repair job and the ending time of the last repair job. 26 | 27 | ## Strength (Contributions of the paper) 28 | 1. introduce a now family of erasure code called **Locally Repairable Codes** (LRCs), which are efficiently repairable both in terms of network bandwidth and disk I/O. 29 | 2. present both randomized and explicit LRC constructions starting from generalized Reed-Solomon parities. 30 | 3. design and implement a module that replaces Reed-Solomon codes with LRCs in HDFS-RAID. 31 | 4. The part of **impartance of repair** is well-written 32 | ## Weakness (Limitations of the paper) 33 | 1. The disadvantage of adding these local parities is the extra storage requirement 34 | ## Future Work 35 | 1. For the part of Reliability Analysis, this paper cannot solve the issue fo measuring the availability tradeoffs of coded storage systems. Thus, this can be a future research direction. 36 | 37 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/JunLi's paper/Beehive-HotStorage'15.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Beehive: Erasure Codes for Fixing Multiple Failures in Distributed Storage Systems 5 | @HotStorage'15 @Multiple Failure @MSR code 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: Generally, distributed storage systems will reconstruct multiple failure blocks separately. However, data unavailability events can be **correlated**, specifically, many disks fail at similar ages. To leveage the advantage of the correlated failures, this paper wants to reconstruct multiple missing blocks in batches, to save both network transfer and disk I/O during reconstruction. 10 | 11 | ***Beehive***: 12 | - An instant benefit of Beehive is each block will only be **read once** to reconstruct multiple blocks. 13 | ![1534820931384](paper_figure/1534820931384.png) 14 | - The construction of Beehive codes is built on top of product-matrix MSR codes, based on one particular *product-matrix* construction proposed by Rashmi. 15 | > 1. MSR codes constructed are systematic 16 | > 2. Unlike other constructions that impose constraints on specific values of $d$ or $k$, the construction proposed in product-matrix is much more general by only requiring $d \geq 2k-2$ 17 | 18 | **Encode**: 19 | It divides one generation of the original data into two parts that contains $k$ and $k-1$ blocks. In first part each block contains $d-k+1$ segments. In second par each block contains $t-1$ segments. 20 | From the original data, Beehive computes $n$ blocks. each block contains the $d-k+1$ segments from $g_iF$ and the $t-1$ segments from $\sum_{l=1}^{k-1}a_{i,l}c_l$ 21 | ![1534843008014](paper_figure/1534843008014.png) 22 | 23 | **Decode**: 24 | - The helpers computes the decoding information and sends it to newcomer $j$. 25 | - At the side of newcomers, it can divide their operation into two stages. In the first stage, it will receive $d$ segments from helpers and do the calculation. In the second stage, it will send the calculation result to another newcomer $j^{'}$, and it will receive $t-1$ segments from other newcomers as well. 26 | - During reconstruction, each newcomer will receive $d+t-1$ segments, achieving the optimal network transfer to reconstruct $t$ blocks. 27 | 28 | 29 | ***Implementation and Evaluation***: 30 | Implement Beehive in C++ by using the Intel storage acceleration library (ISA-L) for the finite field arithmetic. 31 | ## Strength (Contributions of the paper) 32 | 1. this paper propose a new family of erasure codes that can reconstruct multiple blocks at same time. And each block will only be read once to reconstrcut multiple blocks. 33 | ## Weakness (Limitations of the paper) 34 | 1. Need additional storage overhead 35 | 2. Its encoding process is based on MSR code, so Beehive's encoding operation is bit slower than MSR codes 36 | ## Future Works 37 | 1. A very improtant issue I consider is how to combine Beehive with practical distributed storage system. 38 | 39 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/JunLi's paper/Carousel-ICDCS'17.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # On Data Parallelism of Erasure Coding in Distributed Storage Systems 5 | @ICDCS'17 @Data Parallelism 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: Data paralleism, which refers to the number of blocks that can be read by different processes simultaneously, is limited by existing **systematic** erasure code. This paper is designed to extend data parallelism from reading $k$ data blocks in parallel to reading all $n$ blocks. Therefore, it can have a higher overall throughput. 10 | ![1535080941866](paper_figure/1535080941866.png) 11 | 12 | ***Carousel Codes***: 13 | - Carousel codes achieve a flexible trade-off between **data parallelism** and **data availability**. 14 | The process of the construction is shown below: 15 | ![1535118132918](paper_figure/1535118132918.png) 16 | - How to achieve flexible parallelism: This paper also allows users to flexibly specify the degree of data parallelism by controlling the number of blocks that contains the original data. 17 | 18 | 19 | ***Implementation and Evaluation***: 20 | This paper implements the Carousel Code in C++. All operations, including encoding, decoding, and reconstruction are implemented by using ISA-L. 21 | - The advantage of the **sparsity** of the generating matrix in Carousel codes 22 | > Even though the size of the generating matrix is expaned the complexity of encoding and output bit does not change. 23 | 24 | **Evaluation** 25 | >1. The comparison of the encoding and decoding throughput for various values of $k$. 26 | >2. Completion time of reconstruction operations for various values of $k$. 27 | >3. Comparison of Hadoop jobs running on data encoded with systematic RS codes and Carousel codes. (terasort and wordcount in Hadoop) 28 | >4. Comparsion of the time of retrieving a 3GB file from HDFS with systematic RS code and Carousel codes 29 | 30 | 31 | ## Strength (Contributions of the paper) 32 | 1. This paper presents **Carousel Codes**, which has a higher data parallelism $\rightarrow$ a higher overall throughput. 33 | 2. It also shows the besides the MDS property and the configurable data parallelism, Carousel codes can achieve such optimal network transfer during reconstruction as well. 34 | 3. It also implemented Carousel codes in C++ and developed its prototype in Apache Hadoop. 35 | ## Weakness (Limitations of the paper) 36 | 1. I think a important issue of this code is the overhead of reconstruction, including the data re-distribution and re-encoding, which is much higher than the systematic RS code. 37 | ## Future Works 38 | 1. For the first weakness, I consider how to decrease this kind of overhead is a potential research direction. 39 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/JunLi's paper/JunLi-INFOCOM'14.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Cooperative Pipelined Regeneration in Distributed Storage Systems 5 | @INFOCOM'14 @Cooperative Regeneration 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: In regenerating codes, it needs to engage a large number of nodes during regeneration, which is not desirable in practical system. Although, some techniques have been proposed to reduce the number of participating nodes during regeneration, they either fail to maintain the recoverability property, or are not designed for regenerating **multiple data losses**. Thus, this paper proposes a cooperative pipelined regeneration process to regenerate multiple data looses in batches. 10 | 11 | ***Cooperative Pipelined Regeneration***: 12 | 1. Newcomers recevive data from other participating nodes and encode received data into partially regenerated coded blocks. 13 | 2. After this round of regeneration, newcomers will be partially regenerated and become **apprentices**. 14 | 3. In the next round of cooperative pipelined regeneration, apprentices and $r$ new newcomers receive data from other participating nodes with another set of $v$ providers. 15 | ![1534941877505](paper_figure/1534941877505.png) 16 | 17 | ## Strength (Contributions of the paper) 18 | 1. This paper illustrates the pipelined regeneration with both **Random Linear Codes** and **Regenerating Codes**. 19 | ## Weakness (Limitations of the paper) 20 | 1. The apprentices introbduce additional storage overhead in this code 21 | 2. This paper just gives the theroical analysis of this code, not includes the implementation 22 | 3. It needs to do the experiments in a practical distribution storage system. 23 | ## Future Works 24 | 1. This work is really theorical, I think it should discuss more on how to implement it in a practical distribution system. 25 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/JunLi's paper/Pipelined_Regeneration-NetCod'11.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Pipelined Regeneration with Regenerating Codes for Distributed Storage Systems 5 | @NetCod'11 @Regenerating Code 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: In regenerating codes, minimum-storage regenerating codes require the newcomer to receive $\frac{M}{k} \times \frac{d}{d-k+1}$. However, in practical distributed storage system, it is not favorable to let a large number of nodes work cooperatively. This paper proposes a pipelined regeneration process that can reduce the number of participating nodes required by regenerating codes **without sacrificing data integrity**. 10 | 11 | ***Pipelined Regeneration***: 12 | The *independent regeneration* and *cooperative regeneration* 13 | ![1534927153524](paper_figure/1534927153524.png) 14 | In some practical distributed storage system, the regeneration process will not be triggered until the number of failed storage node has reached a **certain threshold** and they can be regenerated in batches. (many newcomers) 15 | ![1534859524747](paper_figure/1534859524747.png) 16 | This paper engages much fewer participating nodes by dividing one round of cooperative regeneration into several rounds of pipelined regeneration. In each round of pipelined regeneration, both the newcomer and apprentices are partially regenerated, such that there are fewer than $d$ providers. Suppose there are one newcomer, $\alpha$ apprentices, $v$ providers, during pipelined regeneration, all providers send $\alpha+1$ ($\alpha$ apprentices and one newcomer). A newcomer will be fully regenerated after $\alpha+1$ rounds of pipelined regeneration. 17 | 18 | ## Strength (Contributions of the paper) 19 | 1. This paper proposes the pipelined regeneration process, which can consume a smaller amount of traffic that would have been achieved by the conventional regeneration process with a much higher number of participating nodes. 20 | ## Weakness (Limitations of the paper) 21 | 1. It just consider the case of single failure and fails to utilize multiple newcomers cooperatively, and can only support MSR codes. 22 | 2. I feel the implementation of this code maybe very hard. So how to combine it with a practical distributed storage system is a issue. 23 | ## Future Works 24 | 1. How to utilize multiple newcomers in cooperatice fashion, such that it can further saving the bandwidth consumption 25 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/Clay Codes-FAST'18.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-root-url: paper_figure 3 | 4 | --- 5 | 6 | # Clay Codes: Moulding MDS Codes to Yield an MSR Code 7 | @FAST'18 @erasure code @Ceph @SMSR Codes 8 | [TOC] 9 | 10 | ## Summary 11 | 12 | ***Motivation of this paper***: There is a strong need for erasure codes that can efficiently recovery from single-code. However, the conventional repair of an RS code is inefficient. This paper will deal with Minimum Storage Regeneration (MSR) codes, which has least possible repair bandwidth and three additional properties. It proposes the Coupled-layer (Clay) Code by extending the theoretical construction of Ye and Brag with practical considerations. 13 | 14 | > 1. minimal disk read 15 | > 2. minimize sub-packetization level $\alpha$ 16 | > 3. small field size, low-complexity implementation 17 | 18 | ***Clay Code Construction***: Here is the example with a (4, 2) scalar RS Code. 19 | ![1533545264630](/1533545264630.png) 20 | 21 | ![1533558772424](/1533558772424.png) 22 | 23 | ***Decoding***: For decoding, it introduces the notion of an Intersection Score (IS) to indicate the number of the vertices that correspond to erased bytes and which are at the same time colored red. Decoding is carried out sequentially, layer-by-layer, in order of increasing IS. 24 | 25 | For example: 26 | ![1533559666806](/1533559666806.png) 27 | 28 | 29 | ***Implementation and Evaluation***: This work introduce the notion of **sub-chunking** to enable use of **vector erasure codes** with Ceph. For implementations of various MDS codes and Galois-field arithmetic, it uses **Jerasure** and **GF-Complete** libraries. 30 | In its experiments, it does evaluations in Amazon EC2. Those experiments are carried out on both **fixed** and **variable object-size** workloads. Measurements are made of: 31 | 32 | > 1. repair network traffic 33 | > 2. repair disk read 34 | > 3. repair time 35 | > 4. encoding time 36 | > 5. I/O performance for degraded, normal operations. 37 | 38 | ## Strength (Contributions of the paper) 39 | 1. introduce the construction of Clay codes 40 | 2. do the modification of Ceph to support any vector code 41 | 3. the integration of Clay codes as a plugin to Ceph 42 | 43 | ## Weakness (Limitations of the paper) 44 | 45 | 46 | ## Future Work 47 | 48 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/Degraded-First-Scheduling-DSN'14.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Degraded-First Scheduling for MapReduce in Erasure-Coded Storage System 5 | @DSN'14 @degraded read schedule 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: There remains an open issue of how to customize the data analytics paradigm (e.g. MapReduce) when they operate in failure mode and need to perform degraded read. 10 | > Based on the observation: while the local tasks are running, the MapReduce job does not fully utilize the available network resources. 11 | 12 | 13 | ***Degraded-first scheduling*** 14 | - **Main idea**: schedule some degraded tasks at the earlier stages of a MapReduce job and allow them to download data first using the unused network resources. 15 | >1. to move the launch of some degraded tasks ahead to take advantage of the unused network resources, so as to relieve the competition for network resources among degraded tasks later. 16 | >2. the default locality-first scheduling launches degraded tasks at the end, thereby making them **compete for network resources**. 17 | 18 | - This paper conducts simple mathematical analysis to compare the default locality-first scheduling and its basic degraded-first scheduling in terms of the runtime of a MapReduce job. 19 | > provide preliminary insights into the potnetial improvement of degraded-first scheduling in failure mode. 20 | 21 | ***Implementation and Evaluation***: 22 | **Implementation**: 23 | This paper implements degraded-first scheduling by modifying the source code of Hadoop 0.22.0. 24 | **Evaluation**: 25 | 1. Comparison of MapReduce Runtime 26 | > Single-job scenario vs Multi-job scenario 27 | 28 | ## Strength (Contributions of the paper) 29 | 1. This paper porposes the degraded-first scheduling, a new MapReduce scheduling scheme that improves MapReduce performance in erasure-coded clustered storage system operating in failure mode. 30 | 2. It also implements a discrete event simulator for MapReduce to explore the performance gain of degraded-first scheduling in a large-scale cluster. 31 | 3. This paper implements degraded-first scheduling on Hadoop and compare the performance of locality-first scheduling and degraded-first scheduling in a 13-node Hadoop cluster. 32 | 33 | ## Weakness (Limitations of the paper) 34 | 1. I think the idea of this paper is not very novel and intutive. 35 | 2. This experiment of this paper is not very enough. 36 | 37 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/Encoding-aware-DSN'15.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Enabling Efficient and Reliable Transition from Replication to Erasure Coding for Clustered File Systems 5 | @DSN'15 @Encoding-aware replication 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: When data blocks are first stored with replication, replica placement plays a critical role in determining both performance and availability of the subsequent encoding operation, 10 | > *random replication (RR)* brings both performance and availability issues to the subsequent encoding operation. 11 | 12 | ***Encoding-Aware Replication*** 13 | - **Main Idea**: For each group of data blocks to be encoded together, EAR keeps one replica of each data block in the same rack, while storing the remaining replicas in other racks by equivalently solving a maximum matching problem. 14 | > It assumes that cross-rack data transfer is the performance bottleneck in a CFS architecture. 15 | 16 | - RR potentially harms performance and availability. 17 | > The primary reason is the replica layout of each data block is independently determined, while the data blocks are actually related when they are encoded together. 18 | > ![1537101426419](paper_figure/1537101426419.png) 19 | 20 | **Eliminate Cross-Rack Downloads** 21 | 1. Formation of a stripe: blocks with at least one replica stored in the same rack.\ 22 | > call this rack the **core rack**. 23 | 24 | **Guarantee the reliability requirements without relocation** 25 | 1. Model the replica layout to a bipartite graph. 26 | 2. Modeling Reliability problem to a **Max matching** $\rightarrow$ a **Max flow problem** in bipartite graph. 27 | 28 | ***Implementation and Evaluation***: 29 | **Implementation**: 30 | ![1537113167500](paper_figure/1537113167500.png) 31 | **Evaluation**: 32 | 1. Testbed Experiment: 13-nodes Hadoop cluster 33 | >a. Encoding Throughput 34 | >b. Write Reponse Time 35 | >c. Impact on MapReduce Jobs 36 | 37 | 2. Discrete-Event Simulations 38 | ## Strength (Contributions of the paper) 39 | 1. This paper presents EAR, a new replica placement algorithm that addresses both performance and availability issues encoding operation. 40 | 2. It implements EAR on HDFS implementation, with only a few modification to the source code. 41 | 3. It conducts testbed experiments on a 13-machine cluster and discrete-event simulations based on CSIM 20. 42 | 4. Examine the replica distribution of EAR, and show that it maintains load balancing in storage and read requests as in RR. 43 | ## Weakness (Limitations of the paper) 44 | 1. This paper considers a very simple scenarios, and ignores the heterogeneous workloads and hardware resources. 45 | ## Future Works 46 | 1. I think it can further study the scenarios with heterogeneous workload and hardware resources. -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/OpenEC-FAST'19.md: -------------------------------------------------------------------------------- 1 | # OpenEC: Toward Unified and Configurable Erasure Coding Management in Distributed Storage Systems 2 | @Erasure code @ongoing work 3 | [TOC] 4 | 5 | ## Summary 6 | ***Motivation of this paper***: 7 | Integrating new erasure coding solutions into existing distributed storage systems is arguably a challenging task that requires non-trivial re-engineering of the underlying storage workflows. 8 | Limitation of EC management 9 | >1. Limited support for adding advanced erasure codes 10 | >2. Limited configurability for workflows of coding operations 11 | >3. Limited configurability for placement of coding operations 12 | 13 | ***OpenEC***: 14 | - Main Idea: 15 | Its main idea is to decouple erasure coding management from the underlying DSS operations 16 | First, this paper design a new erasure coding programming model which is not based on low-level data buffers but based on high-level abstractions, called **ECDAG**. ECDAG is a directed graph to describe the erasure coding process. It has freedom to be enforced with different underlined physical meanings. ECDAG provides the opportunity for control flow and data flow design in erasure coding process. 17 | > Centralized application, distributed application 18 | 19 | For the OpenEC design, it foucs on two types design: **Online EC design** and **Offline EC design**. And it also provides three levels of optimimzation to custom ECDAG designs for offline erasure coding. 20 | >1. Level 0: leverage the **Bind** to avoid the redundant traffic in distributed application of ECDAG. 21 | >2. Level 1: In offline erasure coding, OpenEC leverages the **AddConstraint** takes the locality into consideration and choose location from the locations of all the childs. 22 | >3. Level 2: The optimization works in **hierarchical data center**. OpenEC applies pipelining technique. And it can sort the children according to their regions to check whether layering can be applied to reduce the cross region network traffic. 23 | 24 | ## Strength (Contributions of the paper) 25 | 1. this paper propose a new programming model for EC implementation and deployment **ECDAG**, which defines the flows if erasure coding operations as a directed acyclic graph. 26 | 2. this paper also designs OpenEC, which translates ECDAGs into erasure coding operations atop a DSS. In particular, OpenEC can self-customize ECDAGs for hierarchical topologies to improve repair performance. 27 | 3. it also prototypes OpenEC on HDFS-RAID and Hadoop 3.0 HDFS. 28 | 4. It conducts extensive experiments on OpenEC in a local cluster and Amazon EC2. 29 | 30 | ## Weakness (Limitations of the paper) 31 | 32 | ## Future Work 33 | 34 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/RotatedReed-SolomonCodes-FAST'12.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | # Rethinking Erasure Codes for Cloud File Systems: Minimizing I/O for Recovery and Degraded Reads 5 | @FAST'12 @ Minimizing Recovery I/O 6 | [TOC] 7 | 8 | ## Summary 9 | ***Motivation of this paper***: Existing erasure codes are not designed with recovery I/O optimization in mind. So it needs 10 | >1. optimize existing codes for these operations 11 | >2. new codes which are intrinsically designed for these operations 12 | 13 | ***Algorithm for minimizing number of symbols***: 14 | The algorithm takes as input a **Generator matrix** whose symbols are single bits and the identity of a failed disk and outputs equations to decode each failed symbol. 15 | This algorithm firstly finds a decoding equation for each failed bit while minimizing the number of total symbols accessed. Given a **code generator matrix** and a list of failed symbols, the algorithm outputs **decoding equations** to recover each failed symbol. 16 | > 1. enumerate all valid decoding equations for each failed symbol 17 | > 2. **directed graph** formulation of problem makes it convenient to solve.*(convert this NP-Hard problem to finding the shortest path through the graph)* 18 | >![1533978994514](paper_figure/1533978994514.png) 19 | 20 | For the expensive computational overhead, solutions to all common failure common failure combinations may be computed offline **a prior** and stored for future use. 21 | 22 | ***Rotated Reed-Solomon Codes***: Rotated Reed-Solomon Codes have the modification to standard Reed-Solomon codes. These codes have been designed to optimize the performance of degraded reads for single disk failures. (**improve the penalty of degraded reads**) 23 | ![1533991845908](paper_figure/1533991845908.png) 24 | ![1533992512318](paper_figure/1533992512318.png) 25 | 26 | ***Analysis and Evaluation***: 27 | 1. Reading from the $P$ drive or using standard Reed-Solomon codes is not a good idea in cloud storage system. 28 | 2. Generally, optimally-sparse and minimum-density codes perform best for disk reconstruction. 29 | **Evaluation**: 30 | 1. Data Recovery Rate v.s. Symbol Size 31 | 2. Data Recovery Rate v.s. Different erasure code 32 | 33 | 34 | ## Strength (Contributions of the paper) 35 | 1. Algorithm minimizes the amount of data needed for recovery and it is applicable to **any XOR** based erasure code. 36 | 2. Its implementation and evaluation of this algorithm demonstrates that minimizing recovery data translates directly into improved I/O performance for cloud file systems. 37 | 3. It develops a new class of codes, called Rotated Reed-Solomon codes. A new class of Reed-Solomon Codes which optimize degraded read performance. Better choice than standard Reed-Solomon codes for the cloud. (**arbitrary numbers of disks and failures**) 38 | 39 | ## Weakness (Limitations of the paper) 40 | 1. This paper mentions that its algorithm is computationally expensive (from seconds to hours of compute-time) 41 | 42 | ## Future Work 43 | 1. This paper can be extended by considering how to decrease the computation overhead instead of computing offline **a prior** and storing it for future use. 44 | 45 | -------------------------------------------------------------------------------- /storage_paper_note/erasure_coding/Techniques for Erasure Coding/Two_Tale_in_HDFS-FAST'15.md: -------------------------------------------------------------------------------- 1 | # A Tale of Two Erasure Codes in HDFS 2 | 3 | @FAST'15 @Techniques for Erasure Codes @Adaptive Coding 4 | 5 | [TOC] 6 | 7 | ## Summary 8 | 9 | ***Motivation of this paper***: The increase in the amount of data to be read and transferred during recovery for an erasure-coded system results in two major problems: **high degraded read** and **longer reconstruction time**. This paper wants to exploit the **data access characteristics** of Hadoop workloads to achieve better recovery cost and storage efficiency than the existing HDFS architecture. 10 | ![1533785674947](paper_figure/1533785674947.png) 11 | ***Hadoop Adaptively-Coded Distributed File System (HACFS)***: 12 | HACFS is the first sustem that uses a combination of two codes to dynamically adapt with workload changes and provide both low recovery cost and storage overhead. 13 | 14 | - Adaptive coding： 15 | It designs an adaptive coding module in HDFS and uses a **state machine** to describe the state transitions in HDFS. To maintain this state machine, HAFCS also accounts for the read accesses to data blocks in a file. 16 | 17 | > Read hot files with a high read count are encoded with a *fast code*. 18 | > Read cold files with a low read count are encoded with a *compact code*. 19 | > ![1533801963837](paper_figure/1533801963837.png) 20 | 21 | ***Implementation and Evaluation***: 22 | 23 | 1. Implementation 24 | This paper has implemented HACFS as an extension to the HDFS-RAID module in the Hadoop Distributed File System. Its key module is the adaptive coding module, which tracks the system states and invokes the different coding interfaces. It includes about 3000 LOCs. 25 | ![1533803970846](paper_figure/1533803970846.png) 26 | 2. Evaluation 27 | Compare with Colossus FS, FB HDFS, Azure 28 | 29 | - Degraded read latency 30 | - Reconstruction time 31 | - Storage overhead 32 | 33 | ## Strength (Contributions of the paper) 34 | 35 | 1. This paper designs the Hadoop Adaptively-Coded Distributed File System (HACFS) that adapts to workload changes by using two different erasure codes. 36 | 37 | > **fast code**: optimize recovery cost of degraded reads and reconstruction 38 | > **compact code**: provide low and bounded storage overhead 39 | 40 | 1. A novel conversion mechanism in HACFS to efficiently up/down-code data blocks between the two codes. 41 | 2. Implement HACFS as an extension to HDFS and demonstrate its efficacy using two cases studies with Product and LRC family codes. 42 | 43 | ## Weakness (Limitations of the paper) 44 | 45 | 1. This paper mentions that the conversion cost is high when the HACFS system aggressively converts blocks to limit the storage overhead by upcoding hot files into compact code. So this can be a issue 46 | 47 | ## Future Work 48 | 49 | 1. The paper can be extended by solving the problem of how to reduce the conversion cost. 50 | 51 | -------------------------------------------------------------------------------- /storage_paper_note/general_storage/Hash/CompareByHash-ATC'06.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | Compare-by-Hash: A Reasoned Analysis 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | ATC'06 | Hash | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | This paper investigates the digest of a cryptographic hash function is equal on two distinct files. 14 | > instead of comparing them byte-by-byte. 15 | > can quickly check the files for equality by once again comparing only a few bytes rather than reading through them byte-by-byte. 16 | 17 | In rsync and LBFS, the goal of the designers was not to provide **security** via the use of cryptographic hash functions. 18 | 19 | This paper intends to against the work of Henson in HotOS'03, and suggests that it is certainly fine to use a 160-bit hash function like SHA1 with compare-by-hash. 20 | 21 | 22 | ### Method Name 23 | - The basic of hash functions 24 | 1. collision-resistant 25 | 2. inversion-resistant 26 | 3. second-preimage-resistant 27 | 28 | In the context of compare-by-hash, collision resistant is the main concern. 29 | > if it is given a list of $b$-bit independent random strings, it is expected to see the collision after about $2^{\frac{b}{2}}$. (**birthday bound**) 30 | > For example, for MD5, there should begin showing collisions after $2^{64}$ hash values. 31 | 32 | - The central theme in Henson's paper 33 | Her opposition to the notion that digests can be treated as unique ids for blocks of data. 34 | 35 | - Key property 36 | > Most file-system data are in fact **not uniform** over any domain. 37 | > SHA-1 does a very good job at mapping corelated inputs to uncorrelated outputs. (**differential cryptanalysis**). 38 | 39 | 40 | - The probability of collision 41 | 42 | - Attack model 43 | 1. Bad luck attacker: 44 | The chance that two files will have the same hash is about $2^{-160}$ for a 160-bit hash function like SHA-1. 45 | 46 | 2. An intelligent attacker 47 | try to cause collision, may substitute the blocks with same hash value. 48 | 49 | Compare-by-hash holds up well in both attack models. 50 | > In typical compare-by-hash setting there is no adversary 51 | 52 | ### Implementation and Evaluation 53 | 54 | ## 2. Strength (Contributions of the paper) 55 | 56 | ## 3. Weakness (Limitations of the paper) 57 | 58 | ## 4. Future Works 59 | 1. Two practical applications of compare-by-hash 60 | > rsync, Low-Bandwidth File System (LBFS) 61 | 62 | 2. The security of SHA-1 63 | > this paper estimates it would cost 80,000,000 USD and 2 years to find a collision in SHA-1 64 | 65 | 3. Cryptographic hash function in digitial signature 66 | > a document being digitally signed is first hashed with a cryptographic hash function, and the signature is applied to the hash value 67 | > applying a computationlly-expensive digital signature to a large file would be prohibitive. 68 | -------------------------------------------------------------------------------- /storage_paper_note/general_storage/PM/SLMDB-FAST'19.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | SLM-DB: Single Level Merge Key-Value Store with Persistent Memory 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | FAST'19 | PM | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | - LSM-Tree 15 | - Optimized for heavy write application usage 16 | - Designed for slow hard drives 17 | - LSM-Tree limitations 18 | - Large overhead to locate needed data 19 | - scan multiple levels 20 | - High disk write amplification 21 | 22 | ### SLD-DB 23 | 24 | - Persistent Memory + Disk 25 | - B+tree in PM to manage data stored in the disk 26 | - single disk component C1 that does self-merging 27 | - Architecture 28 | - ![image-20211213011550369](../paper_figure/image-20211213011550369.png) 29 | - Persistent MemTable 30 | - No write-ahead-logging (WAL) 31 | - Stronger consistency compared to LevelDB 32 | - Key-index insertion into B+-tree happens during immutable Memtable flush to disk 33 | - Persistent B+-tree Index (FAST-FAIR B+-tree) 34 | - Per-key index for fast search 35 | - No multi-leveled merge structure 36 | - Selective compaction 37 | - why compaction? 38 | - need GC for obsolete KV-pairs 39 | - need to improve **sequentiality** for **RangQuery** 40 | - Selectively pick SSTable files 41 | - Make those files as compaction candidates 42 | - Merge together most overlapping compaction candidates 43 | - Selection schemes for compaction candidates: 44 | - Live-key ratio selection of an SSTable 45 | - Leaf node scans in the B+-tree 46 | - Degree of sequentiality per range query 47 | - Compaction triggered when there are too many compaction candidate files 48 | 49 | ### Implementation and Evaluation 50 | 51 | - Evaluation 52 | - PM simulator: PM write latency 500ns (5x of DRAM write latency) 53 | - PM read latency & bandwidth same same as DRAM’s 54 | - Intel’s PMDK used to control PM pool 55 | - db_bench and YCSB 56 | - microbenchmark 57 | - PM sensitivity 58 | - PM write latency sensitivity 59 | - PM bandwidth sensitivity 60 | 61 | ## 2. Strength (Contributions of the paper) 62 | 63 | - A new design of KV-stores with persistent memory 64 | - High write/read performance compared to LevelDB 65 | - Low write amplification and near-optimal read amplification 66 | 67 | ## 3. Weakness (Limitations of the paper) 68 | 69 | ## 4. Some Insights (Future work) 70 | 71 | - Level-DB 72 | - Memory: 73 | - MemTable, Immutable MemTable 74 | - In-memory skiplist to buffer updates, mark immutable when becoming full 75 | - flush to disk 76 | - update the **write-ahead-log** (**no fsync**) 77 | - sequential write to the disk 78 | - Disk: 79 | - MANIFEST: store file organization and metadata 80 | - Sorted string tables (SST): merge from Level N to N+1 81 | - LSM-tree optimizations 82 | - Improve parallelism: RocksDB, HyperLevelDB 83 | - Reduce write amplification: PebblesDB 84 | - Optimize for hardware (SSD): VT-tree, WiscKey -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1522721106458.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1522721106458.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1522938283913.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1522938283913.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1523241524030.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1523241524030.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1523329925719.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1523329925719.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1526027787768.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1526027787768.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1526305327024.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1526305327024.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1526389148656.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1526389148656.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1531322452717.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1531322452717.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1531501985565.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1531501985565.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533545264630.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533545264630.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533558772424.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533558772424.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533559666806.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533559666806.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533632589963.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533632589963.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533635551383.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533635551383.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533713900014.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533713900014.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533716802051.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533716802051.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533785674947.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533785674947.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533801963837.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533801963837.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533803970846.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533803970846.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533832651101.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533832651101.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533909218190.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533909218190.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533978994514.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533978994514.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533991845908.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533991845908.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1533992512318.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1533992512318.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534147681755.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534147681755.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534231184704.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534231184704.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534231209288.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534231209288.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534322692743.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534322692743.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534324237488.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534324237488.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534491739567.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534491739567.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534491996294.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534491996294.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534492052150.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534492052150.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534492144193.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534492144193.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534493095149.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534493095149.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534493102024.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534493102024.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534514760021.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534514760021.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534820931384.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534820931384.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534843008014.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534843008014.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534859524747.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534859524747.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534927153524.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534927153524.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1534941877505.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1534941877505.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535025690952.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535025690952.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535080941866.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535080941866.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535118132918.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535118132918.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535424966773.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535424966773.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535509143693.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535509143693.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535529964676.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535529964676.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535530049894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535530049894.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535630992315.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535630992315.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1535718654925.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1535718654925.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536304812187.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536304812187.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536304821166.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536304821166.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536311837927.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536311837927.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536314355540.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536314355540.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536411514566.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536411514566.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1536422986790.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1536422986790.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537101426419.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537101426419.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537113167500.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537113167500.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537156395887.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537156395887.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537170968620.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537170968620.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537502009449.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537502009449.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537519750596.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537519750596.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537523583894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537523583894.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537620570242.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537620570242.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537856398909.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537856398909.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1537856407481.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1537856407481.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1538214740187.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1538214740187.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1538237284824.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1538237284824.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1538364574274.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1538364574274.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1539263939506.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1539263939506.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1539311866259.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1539311866259.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1546917876953.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1546917876953.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1546918149839.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1546918149839.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1546932031713.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1546932031713.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1546937172649.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1546937172649.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1546937279046.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1546937279046.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547042229954.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547042229954.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547437042911.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547437042911.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547437239206.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547437239206.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547449080194.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547449080194.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547719418272.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547719418272.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1547724806425.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1547724806425.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1548039339341.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1548039339341.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1550132357894.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1550132357894.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1550132390502.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1550132390502.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1550133510605.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1550133510605.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1550145505105.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1550145505105.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1550738798721.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1550738798721.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1551686443341.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1551686443341.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1551708715146.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1551708715146.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1551708732566.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1551708732566.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552535525593.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552535525593.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552545363095.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552545363095.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552878709205.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552878709205.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552893983832.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552893983832.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552911506553.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552911506553.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1552912503121.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1552912503121.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553325398564.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553325398564.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553325419379.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553325419379.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553335463524.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553335463524.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553335487753.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553335487753.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553497729383.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553497729383.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1553516953870.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1553516953870.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1554376060947.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1554376060947.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1554865684235.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1554865684235.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1555763466457.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1555763466457.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1555765487674.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1555765487674.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1555769164180.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1555769164180.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1555770525391.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1555770525391.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1556250344815.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1556250344815.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1556250623415.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1556250623415.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1556280764754.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1556280764754.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1557370885604.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1557370885604.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1557642537365.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1557642537365.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1557673345371.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1557673345371.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1557997523898.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1557997523898.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1557997548672.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1557997548672.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558079145575.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558079145575.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558494722437.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558494722437.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558513853943.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558513853943.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558519202625.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558519202625.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558614526597.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558614526597.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558614554155.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558614554155.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558623033915.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558623033915.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558962484711.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558962484711.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1558962790911.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1558962790911.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1559037152377.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1559037152377.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1560441378608.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1560441378608.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1562405070274.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1562405070274.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1564405750463.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1564405750463.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/1564457932100.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/1564457932100.png -------------------------------------------------------------------------------- /storage_paper_note/paper_figure/image-20201029020046029.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yzr95924/storage_paper_list/ac73b26bbf75167ef4e250c4cd5848a3bd36f4d2/storage_paper_note/paper_figure/image-20201029020046029.png -------------------------------------------------------------------------------- /storage_paper_note/security/CompareByHash-ATC'06.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | Compare-by-Hash: A Reasoned Analysis 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | ATC'06 | Hash | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | This paper investigates the digest of a cryptographic hash function is equal on two distinct files. 14 | > instead of comparing them byte-by-byte. 15 | > can quickly check the files for equality by once again comparing only a few bytes rather than reading through them byte-by-byte. 16 | 17 | In rsync and LBFS, the goal of the designers was not to provide **security** via the use of cryptographic hash functions. 18 | 19 | This paper intends to against the work of Henson in HotOS'03, and suggests that it is certainly fine to use a 160-bit hash function like SHA1 with compare-by-hash. 20 | 21 | 22 | ### Method Name 23 | - The basic of hash functions 24 | 1. collision-resistant 25 | 2. inversion-resistant 26 | 3. second-preimage-resistant 27 | 28 | In the context of compare-by-hash, collision resistant is the main concern. 29 | > if it is given a list of $b$-bit independent random strings, it is expected to see the collision after about $2^{\frac{b}{2}}$. (**birthday bound**) 30 | > For example, for MD5, there should begin showing collisions after $2^{64}$ hash values. 31 | 32 | - The central theme in Henson's paper 33 | Her opposition to the notion that digests can be treated as unique ids for blocks of data. 34 | 35 | - Key property 36 | > Most file-system data are in fact **not uniform** over any domain. 37 | > SHA-1 does a very good job at mapping corelated inputs to uncorrelated outputs. (**differential cryptanalysis**). 38 | 39 | 40 | - The probability of collision 41 | 42 | - Attack model 43 | 1. Bad luck attacker: 44 | The chance that two files will have the same hash is about $2^{-160}$ for a 160-bit hash function like SHA-1. 45 | 46 | 2. An intelligent attacker 47 | try to cause collision, may substitute the blocks with same hash value. 48 | 49 | Compare-by-hash holds up well in both attack models. 50 | > In typical compare-by-hash setting there is no adversary 51 | 52 | ### Implementation and Evaluation 53 | 54 | ## 2. Strength (Contributions of the paper) 55 | 56 | ## 3. Weakness (Limitations of the paper) 57 | 58 | ## 4. Future Works 59 | 1. Two practical applications of compare-by-hash 60 | > rsync, Low-Bandwidth File System (LBFS) 61 | 62 | 2. The security of SHA-1 63 | > this paper estimates it would cost 80,000,000 USD and 2 years to find a collision in SHA-1 64 | 65 | 3. Cryptographic hash function in digitial signature 66 | > a document being digitally signed is first hashed with a cryptographic hash function, and the signature is applied to the hash value 67 | > applying a computationlly-expensive digital signature to a large file would be prohibitive. 68 | -------------------------------------------------------------------------------- /storage_paper_note/security/FrequencyHiding-CCS'15.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Frequency-Hiding Order-Preserving Encryption 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----: | :--------------------------------------------------: | 8 | | CCS'15 | Frequency Encryption, Property Preserving Encryption | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | This paper wants to do a new trade-off. It can clearly increase security while preserving the functionality for most queries relying on the order information. 14 | 15 | It needs to increase **client storage size** and introduce **a small error in some queries**. 16 | 17 | - Application of order-preserving encryption 18 | > Order-preserving encryption enables to perform range queries over an encrypted database without any changes to the database management system. 19 | 20 | **Security notion**: indistinguishability under *frequency-analyzing* 21 | 22 | ### Frequency-Hiding Order-preseving Encryption 23 | 1. Main idea: present an order-preserving encryption scheme that is randomized. 24 | > Repeated plaintexts will (or can) become different ciphertexts. 25 | 26 | 2. It also employs a number of data compression techniques to reduce the amount of information stored on the client. 27 | > make this scheme somewhat practical while still improving the security of ciphertexts. 28 | 29 | 30 | 31 | ### Implementation and Evaluation 32 | 33 | ## 2. Strength (Contributions of the paper) 34 | 35 | ## 3. Weakness (Limitations of the paper) 36 | 37 | ## 4. Future Works 38 | 1. This paper mentions that if the ciphertexts can approximate a uniform distribution, it would improve against frequency-analyzing attacks. 39 | 40 | 2. Some related notions in this area 41 | > Searchable encryption, homomorphic encryption, functional encryption -------------------------------------------------------------------------------- /storage_paper_note/security/OverheadConfidentiality-UCC'19.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | The Overhead of Confidentiality and Client-side Encryption in Cloud Storage Systems 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | UCC'19 | Data Encryption | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | Many cloud storage services are fairly blunt regarding the lack of confidentiality they provide. 15 | > a solution to provide confidential cloud storage is to use client-side encryption (CSE). 16 | 17 | - The issue of CSE 18 | CSE complicates file synchronization techniques, such as deduplication and delta encoding, commonly used to reduce the traffic overheads associated with personal cloud storage systems. 19 | 20 | - The goal of this paper 21 | this paper presents empirical experiments and analysis of CSE-related overheads. 22 | > compare and contrast the security and bandwidth saving features implemented by CSE services and non-CSE services. 23 | > compression, delta encoding, and deduplication 24 | 25 | ### Data confidentiality overhead 26 | This paper focuses on pure CSEs. 27 | 28 | - Service 29 | 1. CSEs 30 | Mega, Sync.com, SpiderOak, Tresorit 31 | > Mega: AES-128 32 | > SpiderOak: AES-256-CFB 33 | > Password-Based Key Derivation Function 2 (PBKDF2) 34 | 35 | 2. non-CSEs 36 | Dropbox, iCloud, Google Drive, Microsoft OneDrive 37 | > This was selected base on recommendations in online reviews 38 | > https://www.cloudwards.net/comparison/ 39 | 40 | - Baseline methodology 41 | adding files to the cloud services' sync folders and performing targeted system and network measurements during the sync process. 42 | 43 | 1. Network traffic 44 | using Python modules **netifaces** and **pcapy** among others. 45 | 2. CPU memory overhead 46 | using Python modules **psutil** 47 | 48 | 49 | - Bandwidth saving feature 50 | ![1569071462061](../paper_figure/1569071462061.png) 51 | 1. Client-side deduplication 52 | Dropbox, iCloud, Mega, SpiderOak, and Sync.com 53 | 2. Other 54 | Google drive, One Drive and Tresorit 55 | 56 | ![1569076126086](../paper_figure/1569076126086.png) 57 | 58 | ### Implementation and Evaluation 59 | 60 | - Evaluation 61 | Setting: Macbook Air, high-speed university network through 10Gb/s. 62 | 63 | ## 2. Strength (Contributions of the paper) 64 | 65 | 1. This paper presents a comprehensive analysis of state-of-art cloud storage services 66 | 67 | ## 3. Weakness (Limitations of the paper) 68 | 69 | ## 4. Future Works 70 | 1. how to combine delta encoding with CSEs 71 | the development of optimized delta encoding policies for CSEs, which minimize the bandwidth and storage overhead associated with CSE. -------------------------------------------------------------------------------- /storage_paper_note/security/QuantifyingInformationLeakage-CCSW'19.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | Quantifying Information Leakage of Deterministic Encryption 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | CCSW'19 | Deterministic Encryption | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | - Deterministic encryption 14 | ensures that the same plaintext encrypted under the same key will produce the same ciphertext. 15 | > enable clients to make queries on sensitive data. 16 | > the security implications of deterministic encryption are not well understood. 17 | 18 | 19 | - Deterministic encryption is controversial 20 | - CryptoDB claims that: deterministic encryption is safe to use for sensitive field if every value in a column appears only once. 21 | - Deterministic encryption for non-unique fields is described as **allowing some leakage**. (there does not currently exist a clear understanding of the leakage) 22 | 23 | This paper provides a leakage analysis of deterministic encryption through the application of the framework of **quantitative information flow**. 24 | 25 | ### Quantitative Information Flow (QIF) 26 | - Key Insight: 27 | there is no one "correct" way to measure leakage without a given operational scenario 28 | > different operational scenarios require different leakage measures. 29 | 30 | - QIF definition 31 | - prior g-vulnerability 32 | - posterior g-vulnerability 33 | ![image-20191118003954646](../paper_figure/image-20191118003954646.png) 34 | 35 | it is natural to measure the leakage of the channel by comparing the prior g-vulnerability with the posterior g-vulnerability. 36 | 37 | - Model of deterministic encryption 38 | - Bayes vulnerability 39 | the adversary attempts to guess the entire column. consider different distributions(three kinds of values): 40 | > 1. uniform distribution 41 | > 2. an arbitrarily chosen non-uniform distribution 42 | > 3. a distribution in which two values has the same probability 43 | > 4. a distribution in which two values' probabilities are very close but not the same. 44 | 45 | ### Implementation and Evaluation 46 | 47 | ## 2. Strength (Contributions of the paper) 48 | 1. This paper provides a comprehensive information leakage analysis via considering different distribution and different operational scenarios of the adversary. 49 | 50 | ## 3. Weakness (Limitations of the paper) 51 | 52 | ## 4. Some Insights 53 | 1. This paper mentions the way to mitigate inference attack by inserting fake entries prior to uploading the database to the cloud. 54 | > For encrypted deduplication, can we insert fake chunk to the original workload to mitigate the attack? 55 | 56 | 2. This paper considers different distribution and analyzes the information leakage in different cases. 57 | > In our paper, we do not provide a fine-grain information leakage analysis, we just use KLD as a coarse measurement for information leakage. -------------------------------------------------------------------------------- /storage_paper_note/security/Secret-Sharing/SSSS-ACMCommunication'79.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | How to Share a Secret 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----------------------: | :----------------------: | 8 | | Communication of the ACM 1979 | Secret Sharing Algorithm | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | Goal: design a convenient secret sharing algorithm $(k, n)$ *threshold scheme*. 14 | ### Shamir Secret Sharing 15 | Based on **polynimial interpolation**, for a $(k, n)$ scheme, and data $D$ 16 | 1. pick a random $k - 1$ degree polynomial, 17 | $$ 18 | q(x) = a_0 + a_1x + a_2x^2 + a_3x^3 + ...... + a_{k-1}x^{k-1} 19 | $$ 20 | 2. Generate $n​$ shares 21 | $$ 22 | a_0 = D, D_1 = q(1), D_2 = q(2), ......, D_n = q(n) 23 | $$ 24 | 3. Given any subset of $k$ of there $D_i$ values, with their identifying indices, it can calculate the coefficients of $q(x)$ by interpolation, and then evaluate $a_0 = q(0) = D$ 25 | 26 | The efficiency of polynomial evaluation and interpolation 27 | > $O(nlog^2n)$ 28 | 29 | - To make the claim more precise, it can use **modular arithmetic** instead of real arithmetic. 30 | 31 | - If $D$ is long, it is advisable to break it into shorter blocks of bits (can be handled separately.) 32 | 33 | ## 2. Strength (Contributions of the paper) 34 | 1. When $k$ is kept fixed, $D_i$ pieces can be dynamically added or deleted *without affecting the other $D_i$ pieces*. 35 | 36 | 2. It is easy to change the $D_i$ pieces without changing the original data $D$. 37 | 38 | ## 3. Weakness (Limitations of the paper) 39 | 1. The storage overhead (Storage space blow up) is very large. 40 | 41 | ## 4. Future Works 42 | -------------------------------------------------------------------------------- /storage_paper_note/security/Secret-Sharing/SecretSharing-Summary.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Secret Sharing Algorithm Summary 5 | ------------------------------------------ 6 | 7 | [TOC] 8 | 9 | ## Summary 10 | - Comparison of Secret Sharing Algorithms 11 | 12 | | Algorithm | Confidentiality Degree | Storage Blowup | 13 | | --------- | ---------------------- | -------------- | 14 | | SSSS | $r=k-1$ | $n$ | 15 | | IDA | $r=0$ | $\frac{n}{k}$ | 16 | | RSSS | $r \in [0, k-1]$ | $\frac{n}{k-r}$ | 17 | | SSMS | $r = k -1$ | $\frac{n}{k} + n \times \frac{S_{key}}{S_{sec}}$ | 18 | | AONT-RS | $r = k -1$ | $\frac{n}{k} + n \times \frac{S_{key}}{S_{sec}}$ | 19 | 20 | Formally, a secret sharing algorithm is based on three parameters $(n, k, r)$ where $n > k > r \geq 0$, 21 | > $n, k$ define the fault tolerance degree of a secret. 22 | > $r$ defines the confidentiality degree of a secret. 23 | 24 | Storage blowup: the ratio of of the total size of $n$ shares to the size of the original size of the original secret. 25 | 26 | > the storage blowup must be at least $\frac{n}{k}$.' 27 | 28 | ## How to Share a Secret (SSSS) 29 | 30 | 31 | 32 | 33 | 34 | 35 | -------------------------------------------------------------------------------- /storage_paper_note/security/SecretSharing/SSSS-ACMCommunication'79.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | How to Share a Secret 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :----------------------: | :----------------------: | 8 | | Communication of the ACM 1979 | Secret Sharing Algorithm | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | Goal: design a convenient secret sharing algorithm $(k, n)$ *threshold scheme*. 14 | ### Shamir Secret Sharing 15 | Based on **polynimial interpolation**, for a $(k, n)$ scheme, and data $D$ 16 | 1. pick a random $k - 1$ degree polynomial, 17 | $$ 18 | q(x) = a_0 + a_1x + a_2x^2 + a_3x^3 + ...... + a_{k-1}x^{k-1} 19 | $$ 20 | 2. Generate $n​$ shares 21 | $$ 22 | a_0 = D, D_1 = q(1), D_2 = q(2), ......, D_n = q(n) 23 | $$ 24 | 3. Given any subset of $k$ of there $D_i$ values, with their identifying indices, it can calculate the coefficients of $q(x)$ by interpolation, and then evaluate $a_0 = q(0) = D$ 25 | 26 | The efficiency of polynomial evaluation and interpolation 27 | > $O(nlog^2n)$ 28 | 29 | - To make the claim more precise, it can use **modular arithmetic** instead of real arithmetic. 30 | 31 | - If $D$ is long, it is advisable to break it into shorter blocks of bits (can be handled separately.) 32 | 33 | ## 2. Strength (Contributions of the paper) 34 | 1. When $k$ is kept fixed, $D_i$ pieces can be dynamically added or deleted *without affecting the other $D_i$ pieces*. 35 | 36 | 2. It is easy to change the $D_i$ pieces without changing the original data $D$. 37 | 38 | ## 3. Weakness (Limitations of the paper) 39 | 1. The storage overhead (Storage space blow up) is very large. 40 | 41 | ## 4. Future Works 42 | -------------------------------------------------------------------------------- /storage_paper_note/security/SecretSharing/SecretSharing-Summary.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | Secret Sharing Algorithm Summary 5 | ------------------------------------------ 6 | 7 | [TOC] 8 | 9 | ## Summary 10 | - Comparison of Secret Sharing Algorithms 11 | 12 | | Algorithm | Confidentiality Degree | Storage Blowup | 13 | | --------- | ---------------------- | -------------- | 14 | | SSSS | $r=k-1$ | $n$ | 15 | | IDA | $r=0$ | $\frac{n}{k}$ | 16 | | RSSS | $r \in [0, k-1]$ | $\frac{n}{k-r}$ | 17 | | SSMS | $r = k -1$ | $\frac{n}{k} + n \times \frac{S_{key}}{S_{sec}}$ | 18 | | AONT-RS | $r = k -1$ | $\frac{n}{k} + n \times \frac{S_{key}}{S_{sec}}$ | 19 | 20 | Formally, a secret sharing algorithm is based on three parameters $(n, k, r)$ where $n > k > r \geq 0$, 21 | > $n, k$ define the fault tolerance degree of a secret. 22 | > $r$ defines the confidentiality degree of a secret. 23 | 24 | Storage blowup: the ratio of of the total size of $n$ shares to the size of the original size of the original secret. 25 | 26 | > the storage blowup must be at least $\frac{n}{k}$.' 27 | 28 | ## How to Share a Secret (SSSS) 29 | 30 | 31 | 32 | 33 | 34 | 35 | -------------------------------------------------------------------------------- /storage_paper_note/sketch/Count-MinSketch.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | An improved data stream summary: the count-min sketch and its applications 5 | ------------------------------------------ 6 | | Venue | Category | 7 | | :------------------------: | :------: | 8 | | Journal of Algorithms 2003 | sketch | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | In data stream scenario, algorithm for computing data stream context needs to satisfy the following requirement: 14 | 1. the space used by the algorithm should be small, at most polylogarithmic in $n​$. 15 | 2. processing an update should be fast and simple, with a usable accuracy guarantees. 16 | 17 | This paper proposes a new sketch construction with 18 | >1. space used is proportional to $\frac{1}{\varepsilon}$ 19 | >2. the update time is sublinear in the size of the sketch 20 | >3. improve the space bounds of previous results from $\frac{1}{{\varepsilon}^2}​$ to $\frac{1}{\varepsilon}​$ 21 | >4. improve the time bounds from $\frac{1}{{\varepsilon}^2}$ to 1. 22 | 23 | ### Count-Min Sketch 24 | Two parameters: 25 | > $\varepsilon$ and $\delta$: the error in answering a query is within a factor of $\varepsilon$ with probability $1 - \delta$ 26 | 27 | The space and update time will consequently depend on $\varepsilon$ and $\delta​$. 28 | 29 | It is named after the two basic operations used to answer point queries, counting first and computing the minimum next. 30 | 31 | - Three types: 32 | >1. a point query 33 | >2. a range query 34 | >3. an inner product query 35 | 36 | - Data structure 37 | Two-dimensional array counts with width $w$ and depth $d$, set $w = \frac{e}{\varepsilon}$, and $d = ln(\frac{1}{\delta})$ 38 | $d$ hash functions: 39 | $$ 40 | h_1,..., h_d \space (\{1,...,n\} \rightarrow \{1,...,w\}) 41 | $$ 42 | 43 | - Update procedure 44 | An update $(i_t, c_t)$ arrives, i.e., item $a_{i_t}$ is updated by a quantity of $c_t$ 45 | > 1. $c_t$ is added to one count in each row, and the counter is determined by hash function $h_j$. 46 | > 2. the array of $w \times d$ counts 47 | $$ 48 | count[j, h_j(i_t)] \leftarrow count[j, h_j(i_t)] + c_t, \forall 1 \leq j \leq d 49 | $$ 50 | ![1554376060947](paper_figure/1554376060947.png) 51 | 52 | - The analysis of point query 53 | 54 | 55 | 56 | ### Implementation and Evaluation 57 | 58 | ## 2. Strength (Contributions of the paper) 59 | 60 | ## 3. Weakness (Limitations of the paper) 61 | 62 | ## 4. Future Works 63 | -------------------------------------------------------------------------------- /storage_paper_note/survey/Deduplcation-Survey'18.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: paper_figure 3 | --- 4 | A Comprehensive Study of the Past, Present, and Future of Data Deduplication 5 | ------------------------------------------ 6 | @Survey 7 | [TOC] 8 | 9 | ## 1. Summary 10 | ### Motivation of this paper: 11 | 12 | 13 | 14 | ### Name of the Work 15 | 16 | 17 | 18 | ### Implementation and Evaluation 19 | 20 | ## 2. Strength (Contributions of the paper) 21 | 22 | ## 3. Weakness (Limitations of the paper) 23 | 24 | ## 4. Future Works 25 | -------------------------------------------------------------------------------- /storage_paper_note/template.md: -------------------------------------------------------------------------------- 1 | --- 2 | typora-copy-images-to: ../paper_figure 3 | --- 4 | # Physical vs. Logical Indexing with IDEA: Inverted Deduplication-Aware Index 5 | 6 | | Venue | Category | 7 | | :------------------------: | :------------------: | 8 | | FAST'24 | Deduplicated System Design, post-deduplication application | 9 | [TOC] 10 | 11 | ## 1. Summary 12 | ### Motivation of this paper 13 | 14 | ### Method Name 15 | 16 | ### Implementation and Evaluation 17 | 18 | ## 2. Strength (Contributions of the paper) 19 | 20 | ## 3. Weakness (Limitations of the paper) 21 | 22 | ## 4. Some Insights (Future work) 23 | --------------------------------------------------------------------------------