Google Cloud Pub/Sub is a scalable cloud infrastructure service that moves many gigabytes per second between data sources and data sinks. Storage is a critical component in providing guaranteed at-least-once delivery for every message that passes through Pub/Sub. In this talk, we will be looking at requirements and tradeoffs for throughput, latency, consistency, availability, durability, and maintainability. We will explore the space of design choices for building storage for Pub/Sub, investigate the choices we made, and understand the features of underlying technology.


Jason Ganetsky

Senior Software Engineer and Tech Lead | Google

In particular, I deal with building scalable shared services with complex requirements for throughput, latency, isolation, durability, fault-tolerance, consistency, and availability.In the past I have worked professionally with JavaScript (client-side, Node.js, and hacking Rhino internals), Java with various frameworks, PHP, and Erlang.I enjoy functional programming and static type systems. I am familiar with type theory, Haskell, ML, and Rust. I like "doing things the right way" when it comes to languages, libraries, and idioms

Jason Ganetsky

Experience talks like this and many more at our upcoming event

Learn More

Data Council, PO Box 2087, Wilson, WY 83014, USA - Phone: +1 (415) 800-4938 - EIN: 46-3540315 - Email: community (at) datacouncil.ai