Physical Replication Based on InnoDB
The MySQL Binlog Replication has a lot of advantages. For example, the binary log is more readable and can be used for various scenarios. And we can even build a very complicated replication topology with careful design; But on the other hand, it requires more disk written, two fsync operations to make a transaction durable. The traditional replication mechanism always constitute a serious bottleneck that limit the ability of scaling up the write workload. In this session , I’d like to introduce another replication mechanism: physical replication based on InnoDB. With physical replication ,we can completely disable log-bin. We can achieve much better performance without reducing availability. I will elaborate how we design and implement this exciting feature. You can also learn some kernel knowledges of InnoDB through this talk. The abstract of the session includes : 1. Background Knowledge. Include the life-cycle of a transaction 2. Why we need physical replication: better performance, less disk written, make it easier to implement some other new features 3. The high level architecture of physical replication 4. Replicate changes of server layer 5. How to implement MVCC on slave 6. Replicate change buffer 7. Planned and Unplanned Failover 8. Test Results
Database Developer, Alibaba
I joined Alibaba in June 2011 and work on MySQL since then. My work includes: Improve the performance of MySQL, develop new features according to requests from our DBA team, and also fix bugs we encountered on production environment. Outside working hours , I like to investigate & test new versions of MySQL and file bugs.