简介
我常常主张说,研究分布式系统最难的是改变你思考的方式。对于激发这种改变,我找到的一些很实用的阅读材料。如下。
Thought Provokers
一些让你考虑你设计方式的随笔。不是所有事都可以靠大服务器,数据库和事物来解决的。
- Harvest, Yield and Scalable Tolerant Systems CAP原理在现实世界里的应用来自Brewer等人
- On Designing and Deploying Internet Scale Services James Hamilton
- Latency Exists, Cope! 处理延迟及其架构方面影响的说明
- Latency – the new web performance bottleneck 内容不太新了,但是值得关注下
- The Perils of Good Abstractions 构建完美的API/接口很困难
- Chaotic Perspectives 大规模系统有开发人员不喜欢的所有东西——不可预测,无序,并行
- Website Architecture 一些来自各类大型网站的可扩展架构文章
- Data on the Outside versus Data on the Inside Pat helland
- Memories, Guesses and Apologies Pat Helland
- SOA and Newton’s Universe – Pat Helland
- Building on Quicksand – Pat Helland
- Why Distributed Computing – Jim Waldo
- A Note on Distributed Computing – Waldo, Wollrath 等人
- Stevey’s Google Platforms Rant – Yegge的SOA平台经验
Amazon
有些有关的技术,但更有趣的是他们创造的与之配合的文化和结构。
- A Conversation with Werner Vogels 关于亚马逊转型为一个基于服务的架构的采访报道
- Discipline and Focus 关于亚马逊转型为一个基于服务的架构的另一篇采访
- Vogels on Scalability
- SOA creates order out of chaos @ Amazon
当前分布式系统领域的“火箭科学”(形容艰深的学问)
- MapReduce
- Chubby Lock Manager
- Google File System
- BigTable
- Data Management for Internet-Scale Single-Sign-On
- Dremel: Interactive Analysis of Web-Scale Datasets
- Large-scale Incremental Processing Using Distributed Transactions and Notifications
- Megastore: Providing Scalable, Highly Available Storage for Interactive Services – 实现跨数据中心、低延迟的paxos算法的巧妙设计。
- Spanner – Google的可扩展、多版本、全球分布且同步复制的数据库。
- Photon – 连续数据流的容错和扩容。扩容是非常困难的,尤其是在时钟偏移、高可用性和分布式的情况下.
- Mesa: Geo-Replicated, Near Real-Time, Scalable Data Warehousing 用于存储谷歌互联网广告业务相关的关键测量数据的数据仓库系统。
eBay
有趣的是他们抛弃了大多数的J2EE,并使用了大量的数据库分区。同时,看看他们的网站升级工具。
一致性模型
构建能够适应环境的系统的关键是寻求正确权衡一致性和可用性。
- CAP Conjecture – 一致性,可用性,分区容忍性不可能同时满足
- Consistency, Availability, and Convergence – 证明了在一个典型系统中一致性可能的上界。
- CAP Twelve Years Later: How the “Rules” Have Changed – Eric Brewer 在原来权衡描述工作上的扩展
- Consistency and Availability – Vogels
- Eventual Consistency – Vogels
- Avoiding Two-Phase Commit – 两阶段提交的避免方法
- 2PC or not 2PC, Wherefore Art Thou XA – 两阶段提交不是银弹
- Life Beyond Distributed Transactions – Helland
- If you have too much data, then ‘good enough’ is good enough – NoSQL, 数据理论的未来- Pat Helland
- Starbucks doesn’t do two phase commit – 在起作用的异步机制
- You Can’t Sacrifice Partition Tolerance – 另外的 CAP 说明
- Optimistic Replication – 数据主从复制的弱一致性方法
理论
一些描述了分布式系统设计中各种各样的重要因素的论文。
- Distributed Computing Economics – Jim Gray
- Rules of Thumb in Data Engineering – Jim Gray and Prashant Shenoy
- Fallacies of Distributed Computing – Peter Deutsch
- Impossibility of distributed consensus with one faulty process 也称为FLP [访问需要帐号或付费,免费版本在这里: here]
- Unreliable Failure Detectors for Reliable Distributed Systems.一种处理FLP难题的方法
- Lamport Clocks -当每台电脑的时钟都是独立的时候,你如何建立对时间的全局视图。
- The Byzantine Generals Problem
- Lazy Replication: Exploiting the Semantics of Distributed Services
- Scalable Agreement – Towards Ordering as a Service
- Scalable Eventually Consistent Counters over Unreliable Networks 在不可靠的世界,可扩展计数很困难。
语言和工具
使用特定技术构建分布式系统的问题。
- Programming Distributed Erlang Applications: Pitfalls and Recipes 构建可靠的分布式应用并不仅仅是的选择Erlang还是OTP的问题那么简单。
基础设施
- Principles of Robust Timing over the Internet 即便是调试这么基础的事,管理时钟也很重要。
存储
Paxos 一致性算法
理解这种算法是一个挑战。我建议在阅读其他论文之前先读读“Paxos Made Simple”,然后在读完其他论文之后,再读一遍。
- The Part-Time Parliament – Leslie Lamport
- Paxos Made Simple – Leslie Lamport
- Paxos Made Live – An Engineering Perspective – Chandra等人
- Revisiting the Paxos Algorithm – Lynch 等人
- How to build a highly available system with consensus – Butler Lampson
- Reconfiguring a State Machine – Lamport 等人 -改变集群的成员
- Implementing Fault-Tolerant Services Using the State Machine Approach: a Tutorial – Fred Schneider
其他一致性文章
Gossip 协议(传染行为)
- Epidemic Routing Bibliography
- How robust are gossip-based communication protocols
- Astrolabe: A Robust and Scalable Technology For Distributed Systems Monitoring, Management, and Data Mining
- Epidemic Computing at Cornell
- Fighting Fire With Fire: Using Randomized Gossip To Combat Stochastic Scalability Limits
- Bi-Modal Multicast
- ACM SIGOPS Operating Systems Review – Gossip-based computer networking
- SWIM: Scalable Weakly-consistent Infection-style Process Group Membership Protocol