Weekend Reading #56

Weekend Reading: A weekly roundup of interesting Software Architecture and Programming articles from tech companies. Find fresh ideas and insights every weekend.

This week in the Weekend Reading Digest: database normalization guides, Hotstar’s real-time emoji system, Uber’s safe monorepo deployments, and Expedia’s Iceberg Write-Audit-Publish workflow.

5 Database Normalization Forms

This article explains the five levels of normalization in SQL databases. It shows how structuring tables correctly prevents data duplication and inconsistencies, and when it’s okay to denormalize for performance.

Capturing a Billion Emojis

Hotstar’s team details how they built a system to handle billions of real-time emoji reactions. They use asynchronous processing with Kafka and Spark to efficiently aggregate data and deliver live mood updates to users.

Uber: Controlling the Rollout of Large-Scale Monorepo Changes

Uber explains how they safely deploy changes that affect thousands of services in their mono repositories. They use staged rollout orchestration, simulate outcomes, and monitor service health to reduce risk while maintaining deployment speed.

Iceberg Write-Audit-Publish (WAP)

Expedia Group introduces a workflow to improve data quality using Apache Iceberg. They write data in isolated branches, audit it, and then publish only clean data to production, reducing duplication and ensuring traceability.