When storing time-series data, many developers start with some well-trusted system like Postgres, but as their data hits a certain scale, give up its query power and ecosystem by migrating to a NoSQL or other "modern" time-series architecture.
In this talk, I describe why this trade-off is unnecessary, and how we've built TimescaleDB, an efficient, scalable time-series database engineered up from Postgres. The nature of time-series workloads--appending data about recent events--presents different demands than transactional (OLTP) workloads. We've architected our time-series database to take advantage of and embrace these differences.
TimescaleDB improves insert rates by 15X over Postgres, even on a single node. By right-sizing chunks, it avoids the "performance cliff" Postgres experiences once reaching table sizes of 50+ million of rows, while offering compelling complex query performance improvements. TimescaleDB is packaged as a Postgres extension, released under the Apache
Erik is a senior software engineer at Timescale, focusing on both the core database and infrastructure services. Before joining Timescale, he worked at Spotify on their backend service infrastructure. Erik earned his MSc and PhD from Uppsala in Sweden, then worked as a postdoc and research scientist at Princeton. In his (post)doctoral work, he largely focused on networking and distributed systems, including a new end-host network stack for service-centric networking.