3 ms·
I'd throw in the "Virtual synchrony" set of protocols in the vein of Horus, Isis, VSync, and Derecho that are similar to Paxos, but older and take a slightly di
by monocasa 6y ago
I'd throw in the "Virtual synchrony" set of protocols in the vein of Horus, Isis, VSync, and Derecho that are similar to Paxos, but older and take a slightly different view. I prefer them a bit from an ergonomic perspective. These formed the basis of the New York stock exchange, the Aegis Combat System, and the French air traffic control system, among other extremely highly available systems.
It appears the fantastic Virtual Synchrony wiki page has been unfortunately deleted, and simply replaced with a redirect to a couple sentences on the reliable multicast page (despite not relying on reliable multicast any more than Paxos does) so here's an archive.org capture of the wiki page. https://web.archive.org/web/20181231000827/https://en.wikipedia.org/wiki/Virtual_synchrony https://web.archive.org/web/20181231000827/https://en.wikipe...
- jasonwatkinspdx 6y agoVS and Paxos operate at different layers. For VS to provide its guarantees it needs to use true consensus for group membership. You could see it as a more feature expanded version of leases used to optimize away one of the phases of paxos in the common case.
- monocasa 6y agoThat's for sure one way to look at it, but I prefer to view VS as looking at the concept of consensus as a spectrum and providing tools in many interesting points on that spectrum. Yes, the question of "who gets to participate in the consensus protocol itself" is itself solved through total consensus in VS (and a total consensus protocol older than Paxos). But the way that is then used to bootstrap looser and more performant views of consensus as a first class concept I believe makes it still interesting to bring up in a discussion of consensus in general.
- jasonwatkinspdx 6y agoNot really. Atomic broadcast is equivalent to the consensus problem. VS itself can be turned into Paxos and vice versa via shuffling of details. All of the performance optimizations the VS sales literature pushes so hard are just as possible atop Paxos, and in fact are done now in practice. It's just a divide of concepts/terminology for the most part.
- monocasa 6y agoMaybe this is an example my ignorance on the subject. Can you point to off the shelf Paxos libraries that both have a focus on mixing strengths of consensus as a first class concept and have been used in extremely high availability circumstances?
- jasonwatkinspdx 6y agoBasically all of them support some sort of lease mechanism to avoid one of the protocol phases in the common case. Virtually all of them provide snapshot reads as a performance / convenience thing. It's common for systems to use paxos as the configuration/topology layer while fine grained updates flow through something akin to viewstamped replication. Examples would be Kafka, Amazon Aurora.
- fiddlerwoaroof 6y agoThe deletion of this page really reflects poorly on Wikipedia’s editorial policies.
- monocasa 6y agoYeah, just one more example that the deletionists won years ago.
- Groxx 6y agoafaict they've been pretty consistent on "no original content" since the beginning. Wikipedia "wants" to be an aggregator of knowledge, not a primary source, and that page seems very clearly on the opposite side of that divide. whether or not that's a useful place to draw the line: I dunno. wikipedia is what it is with that line, it's hard to say if it's a cause or in spite of it without competition.
- monocasa 6y agoIt's not really original content though, but instead a pretty accurate summary of it's citations. And now it's been distilled down to the point of being both inaccurate and difficult to find.
- Groxx 6y agoA substantial original summary is original content, in the same way as it would be if it were in a blog / magazine / etc. They've also been consistent on that. It becomes something you can't factually check / has no true references.
- monocasa 6y agoA) You can absolutely fact check. There's veritable tomes of VS documentation. Treat it like any other wiki page of a complex topic. B) It was much smaller than the Paxos page. Should we go delete what exists there too?