Articles

March 31, 2026#article

A small collection of articles and papers I have read or found interesting. Each entry keeps the article link, a rating when I had one, and a short observation.

Tech Articles

The Mechanics of Consistent Hashing & Stateful Load Distribution

  • Rating: Amazing
  • Note: Good insights on why we need consistent hashing, when it is overkill, and why consistent hashing alone does not evenly balance load across servers.

Slack - Real-time Messaging

  • Rating: Amazing
  • Note: Good deep dive into Slack chat architecture and how it handled around 16 million connections per host. Covers channel servers, admin, gateway servers, and the web app.

Google Search Knowledge Graph

  • Rating: Amazing
  • Note: Great introduction to the "things, not strings" idea. Feels like an important step forward for text-based search.
  • Rating: Decent
  • Note: Nice overview of how Google Search evolved.

Graph RAG

  • Rating: Amazing
  • Note: Easy-to-read article showing why GraphRAG can outperform baseline RAG.

Netflix WAL Implementation

  • Rating: Decent
  • Note: Nice write-up on resilience at Netflix scale. Right now it feels somewhat like an abstraction over SQS with DLQ handling, while also preserving message ordering within a namespace.

Twitter Architecture (2013)

  • Rating: Amazing
  • Note: Very good explanation of how Twitter handled massive read-heavy traffic, with write-time fanout and interesting tradeoffs around eventual delivery and out-of-order replies.

Research

Knowledge Graph and RAG

  • Note: Added for deeper reading.

Google's Crawling and PageRank Algorithm

  • Rating: Amazing
  • Note: Impressive paper for 1998. Crawling 24 million web pages and building a search engine better than the then-current market while introducing PageRank and the surfer analogy. Worth rereading.

Others

SLACK

  • Note: Source tag was entrepreneur.

Building the Slack MVP

  • Note: Added from the Slack founder-story set.

Problem Solving

  • Note: Source tag was others. Good reminder that tools are trainable, while intuition comes from practice and learning to recognize patterns.