October 31 we hosted a Halloween-themed Amsterdam Open Source Data Infrastructure meetup at the Textkernel offices.
Watch the meetup recording here: https://www.youtube.com/watch?v=dAW5T3Pouog
We had Boriss Mejias, Solution Architect at EDB and accomplished air-guitarist, talk about “Relational and Post-relational Postgres for Data Application Developers”. And Jelte Fennema-Nio, Principal Software Engineer at Microsoft, elaborated on “The future of connection pooling: PGBouncer or something else?”
Boriss told us that as a developer, learning a new language (SQL in this case) breaks you free from the chains of your comfort zone, and that doing compute close to where your data lives, has performance benefits. Staying on theme, he also claimed that “premature optimization (in this case: normalization) is the root of all evil”.
Jelte took us on a tour of the connection poolers in the market today. PgPool-ll has been around a while and doesn’t have the best performance. PgBouncer (of which Jelte is now a maintainer) is small and thus performant, but packed with niche features contributors added over the years. Odyssey has great multi-threading, Supavisor is brandspanking new, pgcat is good when HA is your main concern…
Which one is the best for your use case depends on a lot of factors. “Most folks will go for the solution build in their language, but look at the maintenance strategies for the tools you’re considering to adopt. And maybe help maintainers / maintaining a project!”
Stay tuned for future meetups!
That’s Nils Dijk (SSE at Microsoft), Boriss Mejias, that’s me, and Jelte Fennema in disguise