Possible application

[Benson Margulies <bi...@gmail.com>]

Hello there,

I'm trying to sort out whether Cassandra is a good pick as the data
store for a problem I've got.

The shape of the thing is a large number of hash tables. On a merely
pretty big scale, it can all run on one pretty big machine. On a
gigantic scale, which is an eventual goal, it will need to spread out
over multiple computers.

My concern is, in essence, whether Cassandra will scale down to
'merely big', or whether I need to code to something else for that
purpose and then be able to switch. Imagine, oh, 49 tables, each
storing a million hashes.

Our current prototype is using redis.

Re: Possible application

["Rock, Paul" <pa...@teamaol.com>]

Well, you can make Cassandra work on a single box (or multiple instances on a single box if need be). My experimental/dev cluster that my team plays with to try things out is 6 nodes running on 6 rather small cloud VM's and it works fine. So I'd say yes, it work in the merely big scale.

On Feb 10, 2011, at 10:00 AM, Benson Margulies wrote:

> Hello there,
> 
> I'm trying to sort out whether Cassandra is a good pick as the data
> store for a problem I've got.
> 
> The shape of the thing is a large number of hash tables. On a merely
> pretty big scale, it can all run on one pretty big machine. On a
> gigantic scale, which is an eventual goal, it will need to spread out
> over multiple computers.
> 
> My concern is, in essence, whether Cassandra will scale down to
> 'merely big', or whether I need to code to something else for that
> purpose and then be able to switch. Imagine, oh, 49 tables, each
> storing a million hashes.
> 
> Our current prototype is using redis.