Scale up or out?

Jeremiah Peschka jeremiah.peschka at
Tue Jun 26 10:37:37 EDT 2012

Scale out - zerg rush your data.

Many smaller systems means that, in theory, you're going to be affected
less by poor performance of any single instance.

When you're building out instances in EC2 always remember that the largest
instance in a class of instances is most likely to be the only VM on that
server. So, an m2.4xlarge will probably be the only instance on that VM
host. The biggest downside to shared VMs is that you're also sharing a
1gigabit ethernet connection.

If you're okay with the possibility of sharing your connectivity with other
people, then don't worry about instance sizes and just keep scaling out.

If you do worry about what other jerks are doing, then try to size your
instances so that you're the only person on that piece of hardware.
Basically, the instances you're looking for are m1.xlarge, m2.4xlarge,
c1.xlarge, cc1.4xlarge, cc2.8xlarge. The upside is that you can also hog
all of the instance storage and take advantage of local storage AND get the
redundancy that Riak brings to the table - EBS volumes are RAIDed behind
the scenes.

This handy chart will help:

Jeremiah Peschka
Managing Director, Brent Ozar PLF, LLC

On Tue, Jun 26, 2012 at 7:24 AM, Eric Anderson <anderson at>wrote:

> Hey all,
> Question about EC2 (or scale in general): i'm building a decent cluster,
> to handle 15-20k inserts/s and 5-10k gets per second.  (for a rough idea of
> what I'm doing).  I've been playing with a 15-node cluster of m2.xlarge
> systems, but I am wondering what is better: more small systems or less
> larger systems?
> Any recommendations/hints/tricks would be a huge help!
> Eric
> _______________________________________________
> riak-users mailing list
> riak-users at
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the riak-users mailing list