System design algorithms

Algorithm you should know for preparing system design interview. For example, Learn Geohash / S2 Geometry algorithm for "How to design Uber like system?"

Sponsor

海外兔小班教学帮助你高效准备面试，学员大厂率超过 65%，入职不成功不收费。

How to contribute

The algorithm and its resources should:

1. Can answer a system design question. For instance, Building a complete Tweet index can answer "How to implement Twitter search" or "How to implement hashtag in Twitter".
2. Free to read or watch.
3. Text would be better than videos.

Requirements

• Know when to use ☑️
• Know how it works ✅

Table of contents

• Bloom filter ✅
• Consistent hashing ✅
• Geohash / S2 Geometry ✅
• Two-Phase Commit (2PC) ✅
• Exponential Backoff ✅
• SSTable ✅
• Leaky bucket / Token bucket ✅
• Inverted index ✅
• Distributed Consensus Algorithms (e.g., Paxos, Raft) ✅
• Cache Eviction Policies (e.g., LRU, LFU) ✅
• Rsync algorithm ✅
• HyperLogLog ✅
• Trie algorithm ✅
• Lossy Counting ☑️
• Frugal Streaming ☑️
• Operational transformation ☑️
• Quadtree / Rtree ☑️
• Ray casting ☑️

Bloom filter

A Bloom filter is a data structure designed to tell you, rapidly and memory-efficiently, whether an element is present in a set.

• Supercharging the Git Commit Graph IV: Bloom Filters

Consistent hashing

Consistent hashing is an algorithm designed to distribute data across a cluster in a way that minimizes re-distribution when nodes are added or removed. It is particularly useful in distributed systems, such as distributed caches, distributed storage systems, and load balancing.

• A Guide to Consistent Hashing

Geohash / S2 Geometry

Geohash can used by 1) dating apps to find romantic matches within a particular cell, and to create chat apps.2) Find nearby locations, and identify places of interest, restaurants, shops and accommodation establishments in an area. 3) Geohashers go on global expeditions to meet people and explore new places.

• Location-based search results with DynamoDB and Geohash

Two-Phase Commit (2PC)

Two-Phase Commit is a distributed algorithm that ensures all participating nodes in a distributed transaction either all commit or all abort the transaction, maintaining ACID properties across distributed systems.

• Difference Between Two-Phase Commit and Saga Pattern

Exponential Backoff

Exponential backoff is a retry strategy that increases the delay between retry attempts exponentially, helping to handle transient failures and prevent system overload in distributed systems.

• Retry with backoff pattern

SSTable

SSTable (Sorted String Table) is an immutable data structure that stores key-value pairs in sorted order, used in modern databases like Cassandra and BigTable to achieve high write throughput and efficient read operations.

• SSTable Definition
• SSTables : The secret sauce that behind Cassandra’s write performance..

Leaky bucket / Token bucket

A mechanism to control the amount and the rate of the traffic sent to the network

• Everything You Need To Know About API Rate Limiting
• How to Design a Scalable Rate Limiting Algorithm

Ieversed index

An inverted index is a data structure used primarily in text search engines.

• How search engines work: Crawling, Indexing, And Ranking
• Building a complete Tweet index

Distributed Consensus Algorithms

Distributed consensus algorithm that enables nodes to agree on a value despite failures, requiring a majority for safety and progress.

Cache Eviction Policies

Cache eviction policies determine which items to remove from a cache when it reaches its capacity, with common strategies including Least Recently Used (LRU), First In First Out (FIFO), and Least Frequently Used (LFU).

• Caching in theory and practice

Rsync algorithm

The rsync algorithm is a technique for reducing the cost of a file transfer by avoiding the transfer of blocks that are already at the destination.

• Streaming File Synchronization

HyperLogLog

HyperLogLog is an algorithm for the count-distinct problem, approximating the number of distinct elements in a multiset.

• Redis HyperLogLog Explained

Trie algorithm

Trie is an efficient information reTrieval data structure. Using Trie, search complexities can be brought to optimal limit (key length)

• How to Design an Autocomplete System
• Spell Checkers
• prefix matching words (IP Addresses, Phone Numbers)
• Auto-complete feature using Trie

Lossy Counting

The lossy count algorithm is an algorithm to identify elements in a data stream whose frequency count exceed a user-given threshold.

• Fast and Reliable Ranking in Datastore
• Frequency Counts over Data Streams
• How we built rate limiting capable of scaling to millions of domains
• Rate-limiting strategies and techniques

Frugal Streaming

Frugal Streaming uses only one unit of memory per group to compute a quantile for each group

• Find the nth percentile of the data stream

Operational transformation

Operational transformation (OT) is a technology for supporting a range of collaboration functionalities in advanced collaborative software systems.

• How Google docs handle editing collisions
• Collaborative editing

Quadtree / Rtree

• Spatial Indexing with Quadtrees
• Find nearby interest points

Ray casting

Ray casting is the most basic of many computer graphics rendering algorithms that use the geometric algorithm of ray tracing. Given a point with longitude and latitude, return the Country of the point.

• Ray Casting Algorithm