Skip to Content.
Sympa Menu

mace-opensaml-users - Re: [OpenSAML] signing performance

Subject: OpenSAML user discussion

List archive

Re: [OpenSAML] signing performance


Chronological Thread 
  • From: Chad La Joie <>
  • To:
  • Subject: Re: [OpenSAML] signing performance
  • Date: Fri, 07 Jan 2011 08:14:44 -0500
  • Organization: Itumi, LLC

I can say that in testing the IdP a few years ago (it should be faster now after we made some improvements), I was about to get about 70 requests a second per 2GHz Opteron core. That was doing crypto but not external attribute resolution. Today's JVMs, CPUs, and memory should all be more efficient and you have more cores so you can try to roughly extrapolate from there.

On 1/7/11 8:09 AM, Chris Card wrote:

>
> >One last question: what sort of throughput is it reasonable to expect
> >from an opensaml IDP which is signing assertions?
>
> I don't think that's really answerable, it depends what the IdP is doing.
> Shibboleth apparently handles 100+ transactions at a time on good
hardware
> doing authentication and attribute resolution with sustained subsecond
> responses. So...lots?
>
> -- Scott
>
thanks Scott, that was helpful, I was only looking for a rough estimate.

Chris

--
Chad La Joie
http://itumi.biz
trusted identities, delivered



Archive powered by MHonArc 2.6.16.

Top of Page