Parallelizing Analytical Queries in PostgreSQL Webinar Recording
This is the recording of a webinar presented on June 10, 2015 by Marco Slot in which he discussed the need for sharding and parallel query execution to build an analytical dashboard for big data. He discussed different schemes for sharding tables, specifically range-partitioning and hash-partitioning, and their implications. He showed how CitusDB plans and executes distributed queries on sharded tables to correctly and efficiently push down most of the computation to a cluster of worker nodes. Finally, Marco talked about the pros and cons of using different sharding schemes in relation to parallel query performance and different data loading methods.