Lesson 231
Advanced Distributed Systems
Time · Gossip · Hashing · CRDTs
1:00The deeper machinery behind distributed systems — logical clocks, vector clocks, consistent hashing, gossip, CRDTs, tail latency, and cross-service transactions.
By the end, you can
- Explain why wall-clock timestamps cannot reliably order events across machines.
- Apply the Lamport clock rules (increment on send, max+1 on receive) and compute correct timestamps.
- Apply the vector clock rules (bump own slot on local event, element-wise max then bump on receive) and determine whether two events are causally related or concurrent.
- Compare vector clocks component-wise and state whether one happened-before the other or they are concurrent.
- Describe how consistent hashing limits key migration to ~1/N when a node joins or leaves.
- Explain why gossip reaches all N nodes in ~log N rounds and why it is resilient to failures.
- Trace a G-Counter CRDT merge (element-wise max) and compute the resulting value.
- Explain strong eventual consistency and the three CRDT merge properties that guarantee it.
- Describe the tail-latency problem in fan-out requests and explain how hedged requests address it.
- Compare two-phase commit and sagas on atomicity, blocking behaviour, and visibility of intermediate state.
Up next in Graphics, DevOps & Cloud Systems
