Vitess is built on three major pillars: Query Serving, Cluster Management and VReplication. All are critical to the success of running a distributed SQL database. Vitess is composed of various different components that work in harmony to achieve scalability and reliability.

In this session we will walk through how Vitess Query Serving works and how it goes through the different stages: from the network protocol, through parsing, planning, and down to execution and session management, finally to provide the query result back to the client. The session also discusses what Vitess has to do to provide the appearance of a single monolithic MySQL even as data is sharded. In the end, we will also touch on how we ensure that Cluster Management and Sharding operations have almost zero impact on the query serving uptime.