How Storj uses CockroachDB to do the impossible with global cloud storage
summary
Claims a global object storage system that is twice as performant as hyperscale alternatives while costing 80% less and cutting carbon footprint by 80%. Uses Storj’s Satellite layer backed by CockroachDB as the metadata orchestrator across 9 cloud regions, tracking 22,000 storage nodes, 5.1 billion rows and a 5 TB cluster with ~5k reads/s and 1.2k writes/s. Describes a live geolocation demo and using time travel queries to debug and recover correctness.
AI-written summary · about
chapters (15)
- 0:00Welcome and Introduction
- 0:30Doing the impossible???
- 1:46How does Storj work?
- 3:01Why Storj object storage is fundamentally different
- 4:13The three key elements of Storj (Nodes, Applications, and Satellites)
- 5:23How Storj keeps their distributed data durable and available
- 6:09Storj demo: Seeing where files are stored in real time
- 7:27Why horizontal scalability, performance, and strong consistency were essential for Storj
- 8:28Why Storj chose CockroachDB for their metadata orchestrator
- 9:09CockroachDB vs Postgres, Citus, Cassandra, Spanner, Yugabyte, and TiDB
- 10:50CockroachDB success story: Working with the Cockroach Labs team
- 11:39CockroachDB success story: Using time travel queries
- 13:02Storj and CockroachDB: The numbers
- 14:33Doing the impossible with CockroachDB
- 15:54Why CockroachDB exceeded Storj's goals / final thoughts
related talks
48mBuilding mission-critical applications with CockroachDB
29mHow City Storage Systems designed its Data Infrastructure to minimize toil
31mFuture-proofing operational data at scale
01h 11mFearless global transactions with
23mWhy Squarespace adopted CockroachDB for better global performance at scale
27mBeyond the Roadmap and Into the Future of CockroachDB
52mCockroachDB
14mHow CockroachDB can enable Multi-Region Deployments and Data Locality