Distributed Systems
How independent machines coordinate over unreliable networks — time, consensus, replication, consistency models, and the failure modes that break production systems.
Time & Ordering
- Physical Clocks, NTP & Clock Skew planned target working high
- Lamport & Vector Clocks planned target working mid
Consensus & Replication
-
Replicated State Machines
budding
working of deep
top
How deterministic execution over one agreed command log turns several unreliable machines into one fault-tolerant service, and exactly where consensus stops and application correctness begins.
-
Byzantine Fault Tolerance & PBFT
budding
deep of deep
high
Why arbitrary faults require 3f+1 replicas and 2f+1 certificates, how quorum intersection prevents conflicting decisions, and how PBFT turns those bounds into a practical partially synchronous protocol.
-
Raft
budding
working of deep
top
Leader-based consensus algorithm that keeps a replicated log identical across a cluster, designed explicitly to be understandable. The protocol behind etcd, Consul, TiKV, and CockroachDB.
- Paxos Family planned target overview mid
- Leader-Follower Replication planned target deep high
- Leaderless Replication & Quorums planned target working high
- Chain Replication planned target overview low
Consistency Models
- Linearizability planned target deep top
- Causal & Eventual Consistency planned target working mid
- CAP & PACELC planned target overview mid
Transactions & Coordination
- Two-Phase Commit planned target working high
- Sagas & Compensating Transactions planned target working high
- Leases, Fencing Tokens & Distributed Locks planned target deep top
- Idempotency & Exactly-Once Semantics planned target deep top
Membership & Data Placement
- Failure Detection & Timeouts planned target working high
- Gossip Protocols planned target overview mid
- Partitioning & Consistent Hashing planned target deep top
-
Kademlia Distributed Hash Table
budding
working of working
mid
A structured peer-discovery and key-lookup system built on the XOR distance, k-buckets, and parallel iterative queries, including why its symmetry simplifies routing and where Sybil resistance must come from.