How Route enables always-on data for over 1 billion+ orders
summary
Route runs self-hosted CockroachDB on AWS ECS to serve over 1 billion+ orders, scaling from 3x nodes and 12x CPU to ~52 TB storage with no data migrations or downtime. Architecture uses tagged EBS volume groups, multiple stores per node, autoscaling EC2 instances and custom user-data tooling to reattach volumes, letting ranges re-replicate after drive failure. Operational practices: patches without maintenance windows, changefeeds to S3 for analytics separation, covering indexes and SQL Insights, and enterprise support to tune the optimizer.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 0:17CockroachDB vs Traditional SQL DB: What happens during a storage failure?
- 0:42How typical SQL DBs handle storage failure
- 2:03How CockroachDB handles a storage failure
- 3:25Who is Bryan Call?
- 4:30What does Route do?
- 5:09How does CockroachDB help deliver on Route's mission?
- 6:32Scalability in CockroachDB
- 7:49Route's journey with CockroachDB: A Marathon
- 8:55Scaling without data migrations or downtime
- 9:30How a resilient database can increase revenue
- 12:10CockroachDB architecture at Route
- 13:17Running CockroachDB self-hosted in AWS using ECS
- 16:17CockroachDB tips and tricks
- 18:03How database resilience leads to customer satisfaction
- 19:28CockroachDB vs DynamoDB
- 20:08CockroachDB vs Traditional SQL
- 20:29Why should you choose CockroachDB?
- 20:55Query best practices
- 22:43When to use covering indexes
- 23:44Pros and cons of covering indexes
- 24:54Challenges with CockroachDB
- 26:34The biggest benefits
related talks
31mFuture-proofing operational data at scale
48mBuilding mission-critical applications with CockroachDB
23mWhy Squarespace adopted CockroachDB for better global performance at scale
14mHow CockroachDB can enable Multi-Region Deployments and Data Locality
27mBeyond the Roadmap and Into the Future of CockroachDB
39mScaling Shopify's Multi-Tenant Architecture across Multiple Datacenters
33mBuilding an Elastic, Resilient Data Layer to Power New Retail Experiences
32mLow Latency Serving of Offline Data: Efficient, Safe, and Reliable Data...
