TLDR: Paper Notes - WTF - The Who to Follow Service at Twitter
Date: 2024-08-21 Source: https://arpitbhayani.me/blogs/wtf-the-who-to-follow-service-at-twitter
Overview
Explore Twitter's recommendation system - in-memory graph processing with Cassovary, random walks, and SALSA for user suggestions. These are my notes based on the paper Paper Notes - WTF - The Who to Follow Service at Twitter The paper covers Twitter’s recommendation system designed to suggest new users for others to follow.
Key Points
- in-memory graph processing on a single server
- the design of Cassovary - their in-memory graph processing engine
- combining random walks and SALSA for better recommendations
- In-Memory Graph Processing on a Single Server: The decision to process the entire Twitter graph in memory on a single server is an interesting design decision.
- Cassovary: In-Memory Graph Processing Engine: Cassovary was not only a custom-built solution, but it was also open-sourced by Twitter.