Beyond NoSQL The Case for Postgres in Read Heavy Workloads
summary
Migrated a small, read‑heavy service from Cassandra+cache to Postgres and sees concrete gains: original workload is 90:10 read:write so caches constantly invalidate under heavy writes; Cassandra forces query‑specific tables and data duplication causing inconsistency without transactions and expensive full‑partition reads. Postgres adds secondary indexes, joins, and buffer cache to speed queries; deployment uses node scaling plus proxy endpoints to send writes to a single writer and reads to readers, then moves to Aurora Serverless with custom endpoints for elastic, cost‑based scaling and improved availability.
AI-written summary · about
related talks
36mPostgres at any Scale
37mPostgres at Any Scale
50mScaling Postgres With Some Help from Redis
51mScaling Postgres to the next level at OpenAI
43mJust Use Postgres.
48mThe NoSQL Way in PostgreSQL
51mHow Postgres is Different (read BETTER than) your database
31mUsing Databases to pull your applications weight