Skip to content

middzwb/ds

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Distributed System

consensus

Time Clocks and the Ordering of Events in a Distributed System 译文

狭义相对论告诉我们时空中的事件并不存在一个始终如一的全序关系;不同的观察者对两个事件谁先发生可能具有不同的看法。当且仅当事件e2是由事件e1引起的时候,事件e1和e2之间才存在一个先后关系。

concurrent的事件在偏序中无法定义顺序;通过定义不同进程之间事件的顺序,将偏序扩展为全序

在外部系统介入后,逻辑时钟会产生“果先于因”的异常,因此引入物理时钟来满足“Strong Clock Condition”

Lamport timestamp -> vector clock

一些参考阅读

关于Paxos的历史

Lamport’s writings page

Consensus Protocols: Paxos

consensus

paxos的工程实践;mulit-paxos:选择master来减少通信和磁盘io。

shadow data structure

vs raft: 线性一致性:不能读到旧数据 raft:只读操作不能直接从leader返回,需要确保leader不是旧的leader master lease:serve read op locally while master held lease snapshot和raft的类似。每个节点独立的快照,同时需要特定信息来维持log的一致(install snapshot);快照的性能:raft论文中使用fork

paxos lecture

vr-revisit

atomic broadcast

难受住了,figure 1看不懂

Primary-backup is also a special case of Vertical Paxos

两篇关于分布式哈希表的介绍,1, 2

lock

The Chubby lock service for loosely-coupled distributed systems

storage

filesystem

Colossus

kv store/object store

DB

分布式id

open-system

  • bitcoin
  • blockstack

design

tiny-lfu

OS

MISC

Distributed systems for fun and profit

papers we love

DDIA

Distributed Systems and the End of the API

Replication, atomicity and order in distributed systems

papers

lock-free

A Distributed Systems Reading List

ReadingList

5-minute rule

About

distributed system

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published