Whitepapers
Not every paper here solves the same problem. Some help you understand Paxos itself, some help once you start thinking about real systems, and some explain why consensus is such a pain in the first place.
Suggested order
If you read these in the wrong order, Paxos gets harder than it needs to be.
If this is your first real pass through Paxos, start here instead of jumping straight into Lamport and wondering why everyone pretends the original papers are easy reads.
Get the roles, message flow, and the shape of the system.
Come back once the moving parts already make sense.
Read one for history, the other for the theory underneath it.
Lamport's own papers. Important, but not where I would tell most people to start.
The original Paxos paper. Historically important, full of the parliament metaphor, and much harder to learn from than its reputation would suggest.
Read it for: the source material, the metaphor, and the history.
The shorter retelling. Cleaner than the original, but it still leaves a lot of the intuition up to the reader.
Read it for: the stripped down statement of the algorithm once you already know the cast.
The useful stuff once you stop asking "what is Paxos?" and start asking "how would this actually run?"
The paper I would hand most people first. It turns Paxos from an abstract proof object into something you can actually picture running.
Read it for: the implementation view.
A pragmatic look at what starts breaking once Paxos stops being a paper exercise and turns into an actual deployed system.
Read it for: operations, failure modes, and the annoying bits the cleaner papers skip.
Good comparison papers if you want to see what is special about Paxos and what is just distributed systems showing up again.
The obvious contrast case. It was written to be understandable on purpose, which makes it useful even if you are here to learn Paxos, not switch sides.
Read it for: a comparison point built around clarity.
Another foundational replication line. Useful for separating what is specifically Paxos from what is common to this whole family of systems.
Read it for: broader context around replication.
A focused tutorial on one of the harder practical problems: changing cluster membership without breaking safety.
Read it for: reconfiguration.
The theory paper here because consensus did not become painful by accident.
The impossibility result under the whole field. It explains why consensus papers keep making timing assumptions and awkward tradeoffs instead of just "solving" the problem cleanly.
Read it for: the theoretical background.