Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the implementation of RocketMQ's million queue capability support using RocksDB KV storage in this 29-minute conference talk. Delve into the existing storage architecture challenges faced by RocketMQ, including high CPU pressure from persistent serialization and CRC32 computing, as well as performance bottlenecks in consumeQueue index implementation. Learn how the introduction of RocksDB KV storage addresses these issues by leveraging WAL and memtable for metadata writing, and utilizing LSM tree to optimize random read/write operations for small files. Gain insights from Zhao Fujian, a senior development engineer at Alibaba, as he shares his expertise on enhancing RocketMQ's scalability and performance for million-queue scenarios.
