Query Execution Optimization for Broadcast Join using Replicated-Reads Strategy – George Wang, Ahana

Query Execution Optimization for Broadcast Join using Replicated-Reads Strategy – George Wang, Ahana

Today presto supports broadcast join by having a worker to fetch data from a small data source to build a hash table and then sending the entire data over the network to all other workers for hash lookup probed by large data source. This can be optimized by a new query execution strategy as source data from small tables is pulled directly by all workers which is known as replicated reads from dimension tables. This feature comes with a nice caching property given that all worker nodes N are now participating in scanning the data from remote sources. The table scan operation for dimension tables is cacheable per all worker nodes. In addition, there will be better resource utilization because the presto scheduler can now reduce the number plan fragment to execute as the same workers run tasks in parallel within a single stage to reduce data shuffles.

Disaggregated Coordinator – Swapnil Tailor, Facebook

Disaggregated Coordinator – Swapnil Tailor, Facebook

In the existing Presto architecture, single coordinator has become a bottleneck in a number of ways for cluster scalability. – With an increasing number of workers, the coordinator has the potential of slow down due to a high number of tasks. – In high QPS use cases, we have found workers can become starved of splits by excessive CPU being spend on task updates in coordinator. – Also with single coordinator, we have an upper limit on the worker pool because of above-mentioned reasons. To overcome with this challenges, we are coming up with a new architecture which supports multiple coordinators in a single cluster.