Scaling Microservices and Distributed Databases in a Multi-Cloud World
summary
Starts on Postgres/RDS, moves heavy documents to dynamodb + lambda streams, adds elasticache in front of hot partitions and hits cost and consistency limits (read/write units billed per kilobyte, GSI eventual consistency). Implements per-field encryption and crypto-shredding for GDPR, then migrates ~20 microservices to a three‑node cluster (i3.2xlarge NVMe) running the new datastore, sustaining about 10,000 writes/sec and 4,000 reads/sec with 99th‑percentile reads <2 ms and roughly 6x cost savings versus prior managed cloud setup.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:34What is Narwha
- 2:34Narwhas scale
- 3:29Narwhas data
- 5:06Sharding
- 6:26Cassandra
- 7:31DynamoDB
- 11:41Queries per second
- 12:29Bursty traffic
- 13:00Brexit and GDPR
- 14:28New Data Source
- 15:22Low Variance
- 16:34Migrations
- 17:02Cluster Overview
- 17:57Operational Numbers
- 18:28Summary
related talks
54mMigrating to Cloud Native with Microservices
42mPostgreSQL Distributed: Architectures & Best practices
26mIntroduction to AWS databases: Modernize your data infrastructure
29mPurpose-Built Databases for Modern Applications
21mSave, Grow and Innovate using AWS Databases
59mModernize your applications with purpose-built databases
43mKeep calm and dive deep into newest database innovations
58mFaster, Cheaper, Greener: Pick Three