For people who studied this book for the interview preps to top tech companies and had success, how did you guys study the book? what part did you find useful? general concepts like partitioning, replication, storage structure (how lsm work), concensus? or focus more on real use cases provided in the book? like twitter pre-computes feed and store it on redis on write path (chapter 1) or using stream processor to do bank reconciliation (chapter 12) which chapters did you focus on and how detailed do you need yo know? do you have to know specific implementation details of all 3 replication strategies and its pitfalls? for transactions should i go into the details of implementation of mvcc and optimistic concurrency control ssi? for consistency/consensus how much details on 2pc and actual consensus algorithms? what about all the different levels of liniarizability? do you have to also know all the details of lsm storage implementations? thanks!!!
This is an amazing book but overkill for the interviews - totally agree
It is an amazing book, but overkill just for the interviews. Grokking the system design should be sufficient. I don’t believe you will need to go to the dark corners of the implementation details, but if you do, they can offer more than you expected 🥳 I was asked ACID, and details on how Isolation works in concurrent transactions.
Do new grads get asked these questions as well?
Nope.