[pkolaczk]

Cassandra had a minor bug with timestamps which was fixed relatively fast and now it passes Jepsen fully. There were also a few implementation issues around LWT (Paxos) but at that time Paxos was very new. There were no architectural / fundamental changes needed to correct these issues. Also, even before fixing, Cassandra was losing only a tiny fraction of data than Mongo was losing. So, yes, this is what I call "far better", even if not initially perfect.

[aphyr]

Cassandra had a minor bug with timestamps which was fixed relatively fast and now it passes Jepsen fully.

Wait, what?

[pkolaczk]

http://www.datastax.com/dev/blog/testing-apache-cassandra-wi...

[aphyr]

Pretty sure Cassandra still uses local timestamps and last-write-wins, which means you can't do safe updates. Also pretty sure they didn't fix the timestamp-collision row-isolation issue I talked about, because it was on HN's front page a few months ago. ;-)

[pkolaczk]

Using timestamps and last write wins is a well known and well documented behavior, so you must assume updates can be sometimes applied in a different order than they were submitted. However, reordering updates is not the same as rolling back writes as was in case of Mongo. Cassandra does not promise linearizability in this mode of operation (if not using LWTs), and this is a tradeoff to get better availability. There is no better way than last write wins if you want high availability and partition tolerance and don't want to pay the performance and availability price for a consensus algorithm like Paxos or Raft. And no, vector clocks do not solve this problem in practice at all.

As for the post on HN from a few months ago, I remember only one guy who mixed LWT and non LWT updates and was surprised with lack of linearizability. Not a Cassandra fault if somebody doesn't know what he's doing.

[aphyr]

> assume updates can be sometimes applied in a different order than they were submitted

I suggest you re-read the analysis. Some databases can offer safe, generalized commutative updates; e.g. Riak. Cassandra can't: updates, in general, can be lost through reordering.

> There is no better way than last write wins if you want high availability and partition tolerance and don't want to pay the performance and availability price for a consensus algorithm like Paxos or Raft.

There is. There's a whole field of research devoted to this problem. http://hal.upmc.fr/inria-00555588/document

[pkolaczk]

You can have safe updates through clustering columns or if you really insist on destructive updates - through LWTs. With clustering columns you can easily achieve whatever is possible with vector clocks.

http://www.datastax.com/dev/blog/why-cassandra-doesnt-need-v...

As for the research you posted, there is no free lunch. Each of these strategies come with their own set of drawbacks. That's why Cassandra offers choice at a query level.