Thanks, I also found a lot of writing on the web that suggests that a
medium is the best value. However, since I'm using SQS to buffer
writes to S3, I find that my app is mostly IO bound and I'm wondering
if it is better to use a medium or four smalls. I couldn't find
anything that specific. I guess I should make my own test and see!

Thanks again,

--fran


On Jun 6, 1:32 am, Arie Fishler <[email protected]> wrote:
> it's all about your own perf. The smallest instance on Aamzon is a real
> lousy configuration. Would do for configuring and setting up your site. From
> reading around the web and testing it myself using some simple load tests we
> found that the next level, the medium size seems to be the best value for
> money.
> Just take into account that the 32bit and 64bit have different roles so
> transition is not automatic and you need to separately configure a 64bit
> role in order to launch one. So - moving from small to medium requires
> nothing but going to large and extra large needs its own roles.
>
>
>
> On Fri, Jun 5, 2009 at 9:57 PM, Fran_K <[email protected]> wrote:
>
> > OK, guys, thanks for the advice. I decided that MySQL just wasn't
> > going to work for me and that I'd prefer to use a file system. So, I'm
> > writing my state files to S3 and will promote some of them to
> > CloudFront if needed. The idea of using SQS and servers who stuff data
> > into S3 seems better to me. Time will tell.
>
> > The one thing I'm stuck on now is if I should use small or large
> > instances. Everything should be IO bound, so I'm not sure if large
> > instances buy me anything. Any ideas on that?
>
> > Thanks again,
>
> > --fran

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"scalr-discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/scalr-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to