Existing single-node RDBMS databases such as PostgreSQL achieve high performance by implementing “pushdowns” aggressively, which move SQL query .. Read more processing as close to the data as possible. Pushdowns can drastically reduce SQL statement processing time by filtering data before transferring it over the network, loading it into memory, pruning out entire files or blocks that do not need to be read. Meanwhile, a distributed SQL database converts a cluster of nodes into a logical RDBMS, thereby offering high availability and horizontal scalability at the same time. Achieving high performance in a distributed SQL database requires a number of optimizations on top of those found in a traditional, single-node RDBMS. In this talk, we will explore the additional pushdown optimizations and internal designs that are needed in a distributed SQL database to make it a high performance system.
- Share your Experience