Concurrency and Coupling as Complexity Enemies

#100 - Modern Software Engineering - Dave Farley

Topic 9 of 12

Published Aug 8, 2022 · Clip 43:53–49:33 (5:40)

Dave calls concurrency and coupling the hardest parts of computer science—complexity is moments away as soon as you branch or spawn a thread. He uses the Google Docs collaboration analogy, describes building possibly the world's highest performance financial exchange at LMAX with deep concurrency management, notes naive parallelism can be 300 times slower, discusses microservices in separate repos that can't release independently, and shares the advice of world-class concurrency experts: don't do concurrency unless you must.