Data-Parallel Actors - A Programming Model for Scalable Query Serving Systems
Explore a 14-minute conference talk from NSDI '22 that introduces data-parallel actors (DPA), a novel programming model for constructing distributed query serving systems. Learn how DPA simplifies the development of complex systems like Apache Druid, ElasticSearch, and InfluxDB by allowing developers to build them from single-node components while automatically providing critical properties such as data replication, fault tolerance, and update consistency. Discover how the researchers used DPA to create a new query serving system based on MonetDB and enhance existing ones like Druid, Solr, and MongoDB with features like load balancing and elasticity. Gain insights into how DPA can significantly reduce code complexity while maintaining state-of-the-art performance and adding rich functionality to query serving systems.