Distributed Systems Foundations
Advanced · 5 concepts · 3 clips · 7 source videos. Watch the clips in order, then take the quick check to see what did not land.
Why distribute, and why not
Distribution buys capacity, latency and fault tolerance. It sells you partial failure, which is the hardest class of bug there is.
Kleppmann (Cambridge) on the reasons to make a system distributed, then the reasons not to (7:03-14:36)
Did this clip teach it?
The ideas in these clips, on Atlas
- Consensus is agreeing on one value when any machine may crash and the network may lieAsk what a consensus protocol assumes about crashes and the network; its guarantee is only as strong as those assumptions.
- In a distributed system every request may arrive twice, so make the second one harmlessDesign every retryable operation so that doing it twice has the same effect as doing it once; the network will test this for you.