Phase 5 — Distributed Systems (Weeks 17–22). Treat this as a mini-course of its own. The Raft paper is mandatory reading. Lamport's original clock paper is short and rewarding. Take your time on consensus.
Planned topics for this phase. This page is a public study plan, not a finished note; items marked done already have notes in this topic.
- Distributed Systems — Motivation and Failure Models
- Remote Procedure Calls (RPC) — Mechanisms and Semantics
- Distributed File Systems — GFS, HDFS, Ceph
- Naming and Directory Services
- Time, Clocks, and Event Ordering
- Logical Clocks and Vector Clocks
- Consistency Models — Strong, Sequential, Causal, Eventual
- The CAP Theorem
- Consensus — Paxos Explained
- Consensus — Raft Explained
- Replication Strategies — Primary-Backup, Chain, Quorum
- Fault Tolerance and Failure Detection
- Distributed Transactions and Two-Phase Commit
- MapReduce and Distributed Computation
- Gossip Protocols and Epidemic Algorithms