Enterprise Scale: How a Distributed Architecture Can Deliver Scale and Resilience
summary
YugabyteDB demonstrates a distributed SQL stack with automatic tablet sharding across nodes, Raft-based replication using RF3 replication, and leader election that fails over in about 3 seconds; demos run on YugabyteDB Managed showing a 3-node cluster scaling to 6 and then 9 nodes with dynamic tablet splits and rebalancing, a simulated zone loss with zero data loss and no application downtime (claims RPO 0, RTO 3s), and an online minor-upgrade flow with brief latency blips.
AI-written summary · about
chapters (13)
- 0:00Start
- 0:21Intro
- 1:22Why Are the Databases Seamless Scale and Resilient Architecture So Very Important for Businesses
- 2:59Tech Stack Modernization
- 5:28Why You Could Buy Db's Distributed Architecture
- 10:00Horizontal Scalability
- 10:30Achieve High Availability and Resilience
- 11:08Raft Consensus Algorithm
- 11:25How Replication Happens in a Three Node Cluster
- 16:11Gigabyte Smart Drive Smart Driver
- 17:24Network Access and Admin Configs
- 18:40In-House Workload Simulator App
- 23:24Horizontal Scaling of the Database
related talks
24mFailure is Not an Option: Highly Available Distributed SQL
30mBuilding Cloud Native Applications on Distributed SQL Platform
27mIntroduction to YugabyteDB – Design and Architecture
27mIntroduction to YugabyteDB: Design and Architecture
25mYugabyteDB Technical Overview
57mNo Downtime, No Limits
26mIntroducing YugabyteDB for PostgreSQL Users
30mTechnical Comparison of YugabyteDB with Oracle and PostgreSQL