attribute resolver behavior

[David Bantz]

Thanks for the explanation and reassurances Scott and Chad.

We are not experiencing a performance problem, so I will push my merely
'in principle' concern into the background and remind myself to do a load
test when volume of transactions increases significantly.

David Bantz
UA OIT IAM

On Wed, 29 Aug 2012, at 16:25 , "Cantor, Scott" <cantor.2 at osu.edu> wrote:

> ...filter policies include plugins that can conditionally
> evaluate things based on the data you resolve, so you can't filter first.
> There's no way to know what the set will be.
> 
> If you have issues with efficiency, you can cache the results. Before
> worrying you would be advised to test it and determine whether there's an
> issue at all….
> 

On Thu, 30 Aug 2012, at 07:50 , Chad La Joie <lajoie at itumi.biz> wrote:

> Unless you've changed the default
> scripting language, all those scripts are compiled at IdP startup and
> run at native Java code speed….
> 
> Attribute resolution is *very* rarely the bottleneck for the IdP (it's
> usually crypto).  It can be if your data sources have very poor
> response characteristics (i.e., very high latency or very low
> availability) but for most sites that's not really the case.
> 
> ...
> Chad La Joie
> www.itumi.biz
> trusted identities, delivered