Skip to content

MarigWeizhi/db_impl_course

 
 

Repository files navigation

数据库系统实现 db_impl_course

数据库系统原理与实现是华东师范大学数据学院开设一门数据库系统课程。 授课人胡卉芪

课程主要内容是当下数据库内核实现中的关键技术,主要包括存储,查询,事务,高可用四个方面。

课程内容:

Time Content Time Content
W1 课程介绍 Reading: 课程简介、系统研究关注的内容、性能指标、课程要求
Lab 1(W1):配置实验环境、编译并调试代码
W10 事务-异常与隔离级别 Reading:Serializablity(i), Serializablity(ii)Conflict Serializable, Recoverability-EmoryTransaction Anomaly,Isolation Levels, 分布式一致性与隔离级别的关系
W2 存储-传统数据库 Reading: KVS的接口与设计需求B+tree, COW-B+treePage Structure,Database Buffer, Cache Policy (i,ii),Implement LRU—Cache
Lab 2-1:实现数据库Drop Table功能
W11 事务-并发控制(一) Reading: 2PL,S2PL,Basic Timestamp
Lab 5-2(W10-W11):2PL算法实现
W3 存储-Bitcast结构 Reading: Bitcast, Log-structured store
Lab 2-2:向数据库中增加DATE字段
W12 事务-并发控制(二) Reading:OCC, MVCC,Snapshot Isolation
Lab 5-3(W12):OCC算法实现
W4 存储-LSM-tree架构存储 Reading:Skiplist-LevelDB,LSM-tree structure & LevelDB,More about Compaction
Lab 3-1(W4):在缓冲区中实现LRU淘汰算法
Lab 3-2(W5):实现缓存管理器
W13 事务-日志管理 Reading:日志与缓冲区关系,ARIES
Final Project:实现 Order By 功能
W5 存储-并发索引 Reading: Memory Consistency ModelConcurrent Linklist(Lock coupling,Optimistic Read,Lazy Delete), Concurrent Index
Lab 5-1(W8):并发链表实现
W14 高可用-数据库备份,Raft(一) Reading:Raft SummaryRaft Paper
W6 存储-其他 Reading:Memory AllocationSnapshot, BloomfilterSecond IndexCompression, Design Considerations for Database storage W15 高可用-Raft(二) Reading: 见上
Lab 6(W14-W17):实现Raft算法
W7 查询-执行引擎 Reading:SQL执行过程火山模型
Lab 4-1(W6): 实现多表查询功能
Lab 4-2(W7): 聚合函数
W16 高可用-分布式一致性与Basic Paxos(一) Reading: Distributed consensus revised-Heidi Howard
W8 查询-算子实现 Reading:Join Algorithms,Grace Join,External Sort W17 高可用-分布式一致性与Basic Paxos(二) Reading: 见上
W9 查询-查询优化器 Reading: Query Optimization W18 分布式数据库主要技术扩展(MPP、分布式事务等) Reading:

扩展内容: 可选

Time Content Time Content
E1 Distributed Database Architecture Reading: Share-noting, Share-data. Cloud Database E2 Storage Reading: Data Partition, Consistent Hashing
E3 Storage Reading: More On LSM-tree E4 Transaction Reading: Distributed Concurrency Control
E5 Transaction Reading: Atomic Commit E6 Consensus Reading: There is more on consensus
E7 Transaction+Consensus Reading: Paxos Commit E8 Query Reading: Exchange, Map-Reduce, Massive Parallel Execution

Reading List:

PingCap整理的数据库论文:https://github.com/pingcap/awesome-database-learning

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Makefile 84.6%
  • C++ 7.9%
  • C 4.1%
  • CMake 3.0%
  • Yacc 0.3%
  • Lex 0.1%