Why did Facebook prefer to HBase than Cassandra?

[MauMau <ma...@gmail.com>]

Hello, (especially Mr. Jonathan Gray, Facebook folks),

I'm sorry for mentioning particular people in a public ML.

I saw the following note from Facebook that says Facebook chose HBase, not 
Cassandra, as the storage for the next messaging infrastructure.

http://www.facebook.com/notes/facebook-engineering/the-underlying-technology-of-messages/454991608919

I'm glad to see this news because I believe that HBase will be used more 
broadly than Cassandra and recommended HBase to my boss and colleagues. 
(However, I understand Cassandra has unique good features.)

I'd like to know why Facebook, the creator of Cassandra, did not choose 
Cassandra. The above note only describes the reason in one sentence:

"We found Cassandra's eventual consistency model to be a difficult pattern 
to reconcile for our new Messages infrastructure."

What kind of operations/features of the new Message didn't Cassandra work 
well for? Counting message, users or something like that because Cassandra 
needs ZooKeeper to count things correctly? Otherwise, eventual consistency 
leads to the undesirable situation where newer messages could appear in the 
inbox without older ones appearing? I'd appreciate if you could share your 
concrete experience/opinions and let us know when HBase fits better or 
Cassandra is difficult to adopt.

Anyone's opinions or guesses will be appreciated.

Best regards,
- Maumau

Re: Why did Facebook prefer to HBase than Cassandra?

[MauMau <ma...@gmail.com>]

Hello, Jean-Daniel,

Thank you for telling me good pointers. I was afraid I could get my question 
to be noticed by FB people and obtain reliable answers, so I asked here. One 
more reason is that I'm a fan of HBase (I'm sorry I thought this might not 
be the best place to ask this question.)

Anyway, I'm still unsure what usage/operations are not handled well by 
Cassandra's eventual consistency. In other words, I'd very much like to what 
kind of usages HBase is pratical for. Sharing such experience will help us 
understand the goodness of HBase further. I've been wondering if eventual 
consistency is pratically acceptable for many people. FB's experience would 
be influencial.

However, I feel this is not the best place to ask. So, I won't continue this 
question here.

Regards,
- Maumau

----- Original Message ----- 
From: "Jean-Daniel Cryans" <jd...@apache.org>
> This isn't the right forum for that kind of discussion.
>
> I recommend going on Quora which already has a few good threads on the
> subject, answered by FB folks, namely:
>
> http://www.quora.com/Why-did-Facebook-pick-HBase-instead-of-Cassandra-for-the-new-messaging-platform
>
> and
>
> http://www.quora.com/How-does-HBase-write-performance-differ-from-write-performance-in-Cassandra-with-consistency-level-ALL
>
> Thanks,
>
> J-D

Re: Why did Facebook prefer to HBase than Cassandra?

[Jean-Daniel Cryans <jd...@apache.org>]

This isn't the right forum for that kind of discussion.

I recommend going on Quora which already has a few good threads on the
subject, answered by FB folks, namely:

http://www.quora.com/Why-did-Facebook-pick-HBase-instead-of-Cassandra-for-the-new-messaging-platform

and

http://www.quora.com/How-does-HBase-write-performance-differ-from-write-performance-in-Cassandra-with-consistency-level-ALL

Thanks,

J-D

2010/11/19 MauMau <ma...@gmail.com>:
> Hello, (especially Mr. Jonathan Gray, Facebook folks),
>
> I'm sorry for mentioning particular people in a public ML.
>
> I saw the following note from Facebook that says Facebook chose HBase, not
> Cassandra, as the storage for the next messaging infrastructure.
>
> http://www.facebook.com/notes/facebook-engineering/the-underlying-technology-of-messages/454991608919
>
> I'm glad to see this news because I believe that HBase will be used more
> broadly than Cassandra and recommended HBase to my boss and colleagues.
> (However, I understand Cassandra has unique good features.)
>
> I'd like to know why Facebook, the creator of Cassandra, did not choose
> Cassandra. The above note only describes the reason in one sentence:
>
> "We found Cassandra's eventual consistency model to be a difficult pattern
> to reconcile for our new Messages infrastructure."
>
> What kind of operations/features of the new Message didn't Cassandra work
> well for? Counting message, users or something like that because Cassandra
> needs ZooKeeper to count things correctly? Otherwise, eventual consistency
> leads to the undesirable situation where newer messages could appear in the
> inbox without older ones appearing? I'd appreciate if you could share your
> concrete experience/opinions and let us know when HBase fits better or
> Cassandra is difficult to adopt.
>
> Anyone's opinions or guesses will be appreciated.
>
> Best regards,
> - Maumau
>
>