Flip Kromer interviewed at Velocity 2011
summary
Infochimps describes a cloud architecture that lets teams spin up and reconfigure clusters with the push of a button, combines tools like Chef and Flume to enforce hard interfaces, and runs a distributed scraper that performs a billion object inserts a day by chaining tiny, testable components: a Redis-backed queue, a simple fetcher, Flume for reliable transport, midstream parsers, and per-service Redis instances. The claim: decouple hard, inspectable pieces, avoid integration bloat, and scale by assembling proven components rather than rewriting them.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 0:21Industrial components
- 2:28Chef and Flume decoupling
- 5:18Efficiency best practices
- 7:27Cloud ops impact
- 9:57Small and fast shipping
related talks
05mNick Gerner interviewed at Velocity 2010
03mChris Phelps Interview
04mInterview
43mTo Petascale and Beyond: Apache Flink in the Clouds
43mArchitecting for Speed & Scale
06mInterview with Ted Dunning
29mEnhancing Flexibility and Productivity with Access Patterns and Storage-Agnostic Abstractions
04mFeynman Liang Interview