Mabuxi
  • Hello World
  • 12/2017 - 01/2018 Google 电面汇总
  • LeetCode Summary
    • Data Structure
      • Stack
      • Queue
      • 单调栈/单调队列
      • Heap (PriorityQueue)
      • Trie
      • Union Find
      • HashTable
      • Custom DS Design
    • Array
      • Two Pointer
      • Find Duplicate 系列
      • Single Number 系列
      • Sequence 系列
    • String
      • Substring Search
      • String Rotation
    • Sort
      • Topological Sort
      • 数据太大怎么办 External Sort
    • Binary Search
    • Index
    • Interval
    • LinkedList
      • Fast and Slow pointer
      • Intersection
      • Entrance of the Cycle
    • Math
      • Problems
    • Greedy
    • Divide & Conquer
    • Sliding Window
    • Dynamic Programming
      • Bottom Up
      • Top Down
      • Knapsack Problems
    • Bit Manipulation
      • Bit Vector
    • Recursion
    • Tree
      • Binary Search Tree
      • Traversal
      • Top Down
      • Bottom Up
      • Binary Indexed Tree
    • Graph
      • BFS
      • DFS
      • DFS + Backtracking
    • Advanced Algorithm
      • Boyer Moore Voting Algorithm
      • Virtual Indexing
      • Reservoir Sampling
    • Game Play
    • 综合题
      • Design Hit Counter
    • Brainstorm 脑筋急转弯
  • CS Foundamental 计算机基础
    • Database
    • Java
    • Java Spring
    • Network
    • Linux
    • Cache
    • Distributed System
      • 分布式配置中心
    • Big Data
  • OOD 面向对象设计
  • System Design 系统设计
  • Regular Expression 正则表达式
  • Resume Deep Dive 简历技术栈
    • MongoDB
    • Docker
    • Nginx
  • TODO List
Powered by GitBook
On this page
  • 基本概念
  • Scalability 扩展性
  • 两种架构模式
  • 一致性算法
  • 负载均衡解决办法
  • 故障分析
  1. CS Foundamental 计算机基础

Distributed System

PreviousCacheNext分布式配置中心

Last updated 6 years ago

基本概念

Synchronous vs Asynchronous

Blocking vs Nonblocking

  • If the thread will not give out a signal when it's finished, synchronous

  • If the thread will give out a signal when it's finished, asynchronous

  • If the process wait the thread to be finished, blocking

  • If the process doesn't wait the thread to be finished, nonblocking.

Reference:

Scalability 扩展性

两种架构模式

  • Scale out (Horizontally Scale)，即通过添加更多的机器，来满足更大的需求

  • Scale up (Vertically Scale)，即通过对单一机器增添CPU、RAM等，来满足更大的需求

通常情况下，先采取 Scale up 的方式，扩展单个服务器的性能，来满足业务的需求。但一旦遇到服务器性能的上限，那么就需要 Scale out 的方式来进一步满足需求。

一致性算法

如何解决分布式系统的沟通问题。

  • Paxos

  • Raft

负载均衡解决办法

如何解决分布式系统负载均衡问题。

  • Consistent Hashing

故障分析

  • server down

https://www.zhihu.com/question/19732473/answer/23434554
https://segmentfault.com/a/1190000014503967