Live data from Hacker News

120K distributed consistent writes per second with Calvin

fauna.com

71–73 of 73 posts

Re: 120K distributed consistent writes per second with Calvin

#71

Including your 9x write amplification in the number of "consistent writes" doesn't count -- like at all. I'm amazed nobody called you out on this yet. You're doing 3k batches per second with 4 logical writes each, right? So that is at most 3-12k writes per second using the way that every other distributed database benchmark and paper counts. Or otherwise - if you continue counting writes in this special/misleading wa…

Even if we focus on 120k writes and ignore the low transaction number then it's still an average result. They use c3.4xlarge to run 5 shards each with 3 replication factor. It's 1500 writes per core (120k / (5*16)). Etcd easily does 5227 writes (I believe it can do even more) on "Standard DS12 v2" so it's 1306 per core which is very similar.

Re: 120K distributed consistent writes per second with Calvin

#72
post #35
post #32

Earlier quoted context omitted.

With the cloud version, it's impossible to run jepsen-like tests to validate consistency and to observe cluster's behavior when the network is unstable and nodes tend to crush.

We have some correctness results coming that should address your concerns. Watch this space for updates.

Will you release binaries or it's gonna be a pure cloud solution like DocumentDB or DynamoDB?

Re: 120K distributed consistent writes per second with Calvin

#73
post #72
post #35

Earlier quoted context omitted.

We have some correctness results coming that should address your concerns. Watch this space for updates.

Will you release binaries or it's gonna be a pure cloud solution like DocumentDB or DynamoDB?

We already have enterprise customers in production on-premises; there will also be a downloadable developer edition. There are no service or library dependencies other than the JVM.
Post reply on HN