This seems cool, and I sincerely wish them nothing but success. That said, I had a major sense of déjà vu while reading this post -- I worked at FoundationDB prior to the Apple acquisition, when we published a blog post with a very similar feel:<p><a href="http://web.archive.org/web/20150325003241/http://blog.foundationdb.com/databases-at-14.4mhz" rel="nofollow">http://web.archive.org/web/20150325003241/http://blog.founda...</a><p>I'm not trying to make a comparison between a system I used to work on and one that I frankly know little to nothing about; rather, I'd suggest that building a system like this just isn't enough to be compelling on its own.
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.<p>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.<p>Or otherwise - if you continue counting writes in this special/misleading way - you'd have to multiply every other distributed db benchmark's performance numbers with a factor of 3-15x to get an apples-to-apples comparison.<p>The 12k batched writes/sek through what I assume is a paxos variant is still pretty impressive though! Good to get more competition/alternatives for zookeeper & friends!
This description is very misleading.<p>120,000 writes per second is accurate, talking about actual durable storage (disk) writes. But it's only 3,330 transactions, which should be the number that a user cares about.<p>I don't have proper data and I'm a bit rusty, but I feel like Cassandra could blow that away if you set similar consistency requirements on the client side (QUORUM on read, same for write?). Am I understanding this correctly, or does Fauna/Calvin give you something functionally better than what C* can do?
Maybe I am missing some special point but a decent PG box will do 1,000,000+ TPS vs 3,000+ TPS here. When pgXact lands it will do close to 2,000,000 TPS. So reading all the posts about the amazing new db "X" that can do about N times less than PG on a multi-node cluster I get confused why the numbers are being presented as some sort of achievement.
This is a new one to me...the referenced paper is here: <a href="http://cs.yale.edu/homes/thomson/publications/calvin-sigmod12.pdf" rel="nofollow">http://cs.yale.edu/homes/thomson/publications/calvin-sigmod1...</a><p>How does this algorithm compare to whatever Google Spanner does?
"Calvin's primary trade-off is that it doesn't support session transactions, so it's not well suited for SQL. Instead, transactions must be submitted atomically. Session transactions in SQL were designed for analytics, specifically human beings sitting at a workstation. They are pure overhead in a high-throughput operational context."<p>Is this specifically for distributed SQL only? I think there are some scalable SQL systems that don't support sessions either.
2011: Benchmarking Cassandra Scalability on AWS - Over a million writes per second<p><a href="http://techblog.netflix.com/2011/11/benchmarking-cassandra-scalability-on.html" rel="nofollow">http://techblog.netflix.com/2011/11/benchmarking-cassandra-s...</a><p>Also a single SSD from 2015 is rated at 120K writes per second:<p>PM1725: <a href="http://www.samsung.com/semiconductor/global/file/insight/2015/11/pm1725-ProdOverview-2015-0.pdf" rel="nofollow">http://www.samsung.com/semiconductor/global/file/insight/201...</a>