PostgreSQL the way only Google can deliver it: A deep dive into AlloyDB

PostgreSQL is one of the most popular databases in the world, and for good reason. But what if it could be faster, cheaper to run, and easier to manage with no application changes? Join us for a deep dive on AlloyDB for PostgreSQL, our fully compatible PostgreSQL database with superior performance, availability, and scale. We’ll go over some of the challenges we’ve seen from customers running PostgreSQL - from tuning performance to managing disruptions from vacuuming, and we’ll explore how AlloyDB solves these challenges. We’ll focus on AI and machine learning-based autopilot management systems, intelligent storage, and advanced caching and dynamic workload-aware data organization.
Speakers: Sandy Ghai, Gurmeet Goindi, Kristofer Sikora
Watch more:
All sessions from Google Cloud Next → goo.gle/next23
#GoogleCloudNext

[@MrTravolta24]

Is this really an accurate / like for like comparison to Exadata? AlloyDB columnstore is eventually consistent right?

[@_____case]

The tradeoff is between availability and consistency. Exadata allows you to sacrifice availability for synchronous replication when using read replicas. Not sure if AlloyDB has this feature, but I think most workloads that achieve moderate scale will have to prioritize availability over consistency at some point anyway, unless the database is doing some real wizardry (e.g. Spanner).