Making Databases Dance: hands on tutorial with Postgres, Linux, Docker, Kubernetes and more
Speaker
Adam Sah
Host
Mike Stonebraker
MIT
Abstract:
Modern Postgres has become a "swiss army knife" for data management and query processing. In this hands-on session (bring a laptop!) we explore some Postgres 11 capabilities that make it especially suitable for research environments and running experiments with data, both small and large. For example, using VIEWs, function indexes and index-only queries to fully abstract the physical layout, for example allowing you to dump 'raw' IoT data into Postgres without needing to process it first (no ETL) including JSON data. For example, using Postgres 11, plv8, citus and docker/kubernetes, you can quickly use Postgres to prototype algorithms in JavaScript and then execute them in parallel across multiple cores and servers, right next to the data, while preserving the data organization features of Postgres as a full-featured RDBMS, including ACID updates and concurrency control, JOINs, naming/VIEWs, secondary indexing, etc.
Bio: Adam Sah has been hacking databases, OSs and programming languages since the early 1990s. He has helped launch a number of successful startups and is credited with helping to invent columnar databases. linkedin.com/in/adamsah
Modern Postgres has become a "swiss army knife" for data management and query processing. In this hands-on session (bring a laptop!) we explore some Postgres 11 capabilities that make it especially suitable for research environments and running experiments with data, both small and large. For example, using VIEWs, function indexes and index-only queries to fully abstract the physical layout, for example allowing you to dump 'raw' IoT data into Postgres without needing to process it first (no ETL) including JSON data. For example, using Postgres 11, plv8, citus and docker/kubernetes, you can quickly use Postgres to prototype algorithms in JavaScript and then execute them in parallel across multiple cores and servers, right next to the data, while preserving the data organization features of Postgres as a full-featured RDBMS, including ACID updates and concurrency control, JOINs, naming/VIEWs, secondary indexing, etc.
Bio: Adam Sah has been hacking databases, OSs and programming languages since the early 1990s. He has helped launch a number of successful startups and is credited with helping to invent columnar databases. linkedin.com/in/adamsah