Mastering Multi-Region Deployments With YugabyteDB
summary
Workshop demonstrates concrete YugabyteDB deployments: a three-node Docker cluster, a live bookstore demo, and scripted labs that simulate a region outage and recovery. It compares single multi-region (stretched) cluster, read replicas (fast reads but forwarded writes, observable replication lag), and geo-partitioned cluster (local reads and writes using Postgres tablespaces/partitions). Shows measured latencies ~3–5 ms local, ~70 ms cross‑country, ~140–170 ms intercontinental, and how to set up one‑way cluster replication and standalone clusters for data residency.
AI-written summary · about
chapters (18)
- 0:00Start
- 1:15Geodistributed Applications
- 2:30Geo-Distributed Application
- 17:14Api Layer and Data Streaming Layer
- 19:32Database Deployments
- 26:06Latency
- 30:53Types of Multi-Region Deployments
- 31:03Single Multi-Region Cluster
- 37:47Connect to the Cluster
- 39:55Deploy Our Cluster across Multiple Regions
- 44:10Market Order Application
- 48:01Change the the Preferred Region
- 53:50Read Replica
- 55:51Disadvantages
- 56:35Why Do We Want To Deploy with Replica Cluster
- 1:00:32Queries through the Replica
- 1:11:22Shut Down the Existing Multi-Region Cluster
- 1:14:09Replication
related talks
30mSolving Multi-Region Data Read and Write Challenges in Cross-Border Payments
38mHow Geo-Distributed Apps Use Global Cloud Infra for Speed and Compliance
31mToo Far Too Furious: Multi-Region Cloud Applications
26mTip and Tricks to Deploying YugabyteDB in a Multi-Cloud World
32mBuilding the Multi-Cluster Data Layer
30mBuilding Scalable Spring Applications with YugabyteDB
26mA Winning Multi Cloud Data Strategy
26mIntroducing YugabyteDB for PostgreSQL Users