A New Sharding Architecture for MongoDB
summary
An overview of MongoDB sharding and replication trade-offs: how replication improves read scalability and reliability by copying data, while sharding improves write and compute scalability by splitting data across machines. The talk discusses how using both together multiplies hardware needs, leading to deployments with hundreds or thousands of servers, and examines the resulting architectural and operational implications.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 3:50Classic replication issues
- 7:44Read versus write differences
- 10:37Server load balancing
- 13:28Fault tolerance and recovery
- 16:16New sharding architecture
- 19:52Ring topology and data distribution
- 22:21Adding and removing nodes
- 24:35Data balancing issues
- 26:01Debugging and monitoring
- 27:11Q&A: Replica sync
- 28:58Q&A: MongoDB future
- 30:04Q&A: Indexes and performance
related talks
34mScaling Applications with MongoDB: Best Practices and Challenges
24mWhen and How Should you Shard?
45mSharding and Things We'd Like To See Improved
50mApproaches to implementing sharding in modern [No]SQL systems
23mSharding for hyperscale and geographically distributed relational databases
25mHybrid Applications with MongoDB and RDBMS
49mSharding: from scratch to Yandex Disk
24mMoving MongoDB to the Cloud: Strategies and points to consider