1. Paper Notes
  2. 1. benchmarks
    1. 1.1. tpch-chokepoints
    2. 1.2. ssb
  3. 2. bigdata
    1. 2.1. mapreduce
    2. 2.2. nephele
    3. 2.3. dataflow model
    4. 2.4. flink
    5. 2.5. flink state management
  4. 3. compiler
  5. 4. databases
    1. 4.1. optimizer
      1. 4.1.1. overview
      2. 4.1.2. system r access path selection
      3. 4.1.3. volcano
      4. 4.1.4. cascades
    2. 4.2. executor
      1. 4.2.1. volcano
    3. 4.3. concurrency control
      1. 4.3.1. evaluation of in-memory mvcc
    4. 4.4. oltp
      1. 4.4.1. through the looking glass
      2. 4.4.2. staring into the abyss
      3. 4.4.3. system r
    5. 4.5. olap
      1. 4.5.1. lakehouse
      2. 4.5.2. delta lake
      3. 4.5.3. vertica
      4. 4.5.4. duckdb
    6. 4.6. htap
      1. 4.6.1. greenplum
    7. 4.7. cloudnative
      1. 4.7.1. aurora
      2. 4.7.2. taurus
    8. 4.8. columnstores vs rowstores
    9. 4.9. kv
      1. 4.9.1. rocksdb cidr17
      2. 4.9.2. wisckey
    10. 4.10. mmdb
      1. 4.10.1. mmdb overview
    11. 4.11. vector db
      1. 4.11.1. hnsw
      2. 4.11.2. ivf-hnsw
      3. 4.11.3. diskann
      4. 4.11.4. product quantization
    12. 4.12. graph db
      1. 4.12.1. kuzu
    13. 4.13. citus
    14. 4.14. cdc
      1. 4.14.1. dblog
    15. 4.15. rum conjecture
  6. 5. datalayout
    1. 5.1. cstore
    2. 5.2. cstore compression
    3. 5.3. dremel
    4. 5.4. rcfile
    5. 5.5. orc
    6. 5.6. table placement methods
  7. 6. data structures
    1. 6.1. btree family
      1. 6.1.1. bw-tree
    2. 6.2. hash table
      1. 6.2.1. linear hashing
    3. 6.3. trie family
      1. 6.3.1. art
      2. 6.3.2. hot
    4. 6.4. bitmaps
      1. 6.4.1. roaring bitmaps
    5. 6.5. skip list
    6. 6.6. bloom filter
  8. 7. distributed system
    1. 7.1. consensus
      1. 7.1.1. flp
      2. 7.1.2. paxos made simple
      3. 7.1.3. paxos made live
      4. 7.1.4. viewstamped replication
      5. 7.1.5. zab
      6. 7.1.6. paxos vs. vr vs. zab
      7. 7.1.7. raft
      8. 7.1.8. paxos vs raft
    2. 7.2. scheduler
      1. 7.2.1. borg
    3. 7.3. primary backup
    4. 7.4. chain replication
    5. 7.5. bolosky
    6. 7.6. holy grail
    7. 7.7. chandy lamport
    8. 7.8. asynchronous barrier snapshotting
    9. 7.9. zookeeper
  9. 8. filesystem
    1. 8.1. gfs
    2. 8.2. polarfs
  10. 9. llm
  11. 10. storage
    1. 10.1. kv store
      1. 10.1.1. dynamo
    2. 10.2. kudu
    3. 10.3. bluestore

论文阅读笔记

kv databases

  • Optimizing Space Amplification in RocksDB, 2017
  • Evolution of Development Priorities in Key-value Stores Serving Large-scale Applications: The RocksDB Experience, 2021 a slightly more detailed version
  • WiscKey: Separating Keys from Values in SSD-Conscious Storage, 2016

Further readings

  • Pipelined Compaction for the LSM-tree, 2014
  • HashKV: Enabling Efficient Updates in KV Storage via Hashing, 2018
  • Dostoevsky: Better Space-Time Trade-Offs for LSM-Tree Based Key-Value Stores via Adaptive Removal of Superfluous Merging, 2018, video
  • Accordion: better memory organization for LSM key-value stores, 2018, HBase Memory Improvement
  • LDC: A Lower-Level Driven Compaction Method to Optimize SSD-Oriented Key-Value Stores, 2019
  • Characterizing, Modeling, and Benchmarking RocksDB Key-Value Workloads at Facebook, 2020