"[Paxos] provides a new way of implementing the state-machine approach to the design of distributed systems."
Original Paxos paper (written with somewhat whimsical style):
https://www.microsoft.com/en-us/research/publication/part-ti...
Follow up paper to explain Paxos more simply:
https://www.microsoft.com/en-us/research/publication/paxos-m...
See also: https://en.wikipedia.org/wiki/Paxos_(computer_science)
The basic idea is how to coordinate multiple independent agents over an unreliable network. For example, multiple servers trying to manage a shared database. Lots of HN people know about this because it's a key building block of reliable and scalable cloud computing.
Raft is a consensus algorithm that is touted as being easy to understand. It is well specified, compared to paxos, which leaves many implementation details up to the creator. Raft, however, is fairly explicitly specified on page 4 of this paper: https://raft.github.io/raft.pdf
Consensus algorithms allow us to build reliable castles out of constantly failing servers (sand). Systems like zookeeper, etcd, chubby, consul, and others use these algorithms to achieve high availability AND strong consistency (linearizability, the strongest possible) despite up to a majority of the cluster failing.
This is a great introductory course about it in my opinion: https://www.coursera.org/learn/cloud-computing
Then, it's just reading papers like, which you mentioned, "The part time parliment" (paxos)
1. I am technically pretty strong but I have no idea what this paper is about
2. So many people know this is about that it shot up to #1 on HN
Can someone give a pointer (a link or two) to the lay, interested audience here about what the field IS. Just a sort of intro guide to someone who knows about programming and math, but has never heard the term paxos?
I am curious, and I am sure many others are as well.
edit: spelling