How to manage complexity in distributed applications
summary
Claims that communication and carefully designed abstractions tame distributed complexity: a two‑year effort produces a working in‑house streaming glue rather than buying Kafka and it proves adaptable. Concrete moves include built a custom streaming glue called film, adding a second, user‑friendly API layer, and inventing IDN message language for portfolios. Surprising metrics recur — six developers is harder to manage than six servers, hundreds of mortgage variants, millions of CPU cores — and practices that work are one‑model‑per‑developer with packaging and a bookkeeper.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 5:04Traditional architecture failures
- 13:56Better abstractions in systems
- 23:32Documentation and ownership
- 29:59Key takeaways
- 31:02Q&A
related talks
23mDesigning Robust and Scalable Distributed Applications
44mCommunication and Transactionality in Distributed Systems
29mFollowing the Data: A Deeper Look into Distributed Applications
01h 03mMicroservices vs The Distributed Monolith
40mDecomposing a live application with confidence
11mScaling systems, scaling influence: Lessons from distributed architecture
01h 26mChaos Development: Removing Software as the Inhibitor of Agile
01h 00mChallenges of distributed systems