PostgreSQL as a Big Data Platform
Chris Travers works in Berlin as a database administrator, in a distributed PostgreSQL environment involving hundreds of TB of data in PostgreSQL. He has experience working with PostgreSQL in many areas from scientific computing in the life sciences, to ERP, from high-velocity ad-tech analytics to geo-spacial databases. He is the author of a number of extensions on pgxn.
His current experience involves mobile advertisement analytics over 400TB data sets and more short-terms log management environments handling PBs of data in PostgreSQL.
In this talk for all audiences, I provide a whirlwind survey of three of our big data environments at Adjust all built on PostgreSQL. These environments are all very different in structure and operation.
One takeaway from this is that it is quite possible to apply NoSQL-like strategies to horizontally scaling various kinds of PostgreSQL analytical environments, but since every large system is unique, the strategies to do so will necessarily differ one system to another. A large system which provides Elastic Search-like functionality will have a very different structure than a distributed data warehouse.
The three systems surveyed are structured very differently and thus represent very different answers to how to horizontally scale PostgreSQL.
- 2019 September 19 14:00
- 50 min
- Winchester (2)
- Silicon Valley 2019