- 了解分布式系统面临的基本问题
- 了解分布式系统的模型和假设构建
- 了解数据复制
- 了解数据分区
- Distributed System For fun and Profit 第一、二章 [1.5h]
- 分布式系统谬论/Fallacies of distributed computing [10min]
- 设计数据密集型应用/Designing Data-Intensive Applications 第五、六 [3.5h]
- scalability
- performance
- availability
- performance
- number of nodes
- distance between nodes
- 可理解性
- 思想:divide and conquer
- Partition
- Replication
- System Model
- node
- communication link
- over-all assumption
- guarantee
- Case Study: Consensus Problem, and impossible results
- Partition Of Key-Value Data
- Partition Of Secondary Indexes
- Rebalancing Partitions
- Request Routing
- Single Leader
- Multi Leader
- Leaderless
- Problems of Replication Lag
- 实现一致性Hash算法
- 设计Dynamic Partition, 过程中保证re-balance和写新数据都成功的方法,且保证可以读到新旧数据
Life beyond Distributed Transaction
Design Data Intensive Application Distributed Data与Derive Data,及其中感兴趣章节的引文