idea - looking for input

[Brian Reindel]

I think this is a great idea, but a word of caution that I would not
share domain specific details. Maybe I'm just overly concerned about
security, but our network operations would shy away from me providing
environment configuration details for specific URLs.

On Wed, Oct 23, 2013 at 4:54 PM, Nate Klingenstein <ndk at internet2.edu> wrote:
> I think it would be valuable.  You might hijack this page for the work:
>
> https://wiki.shibboleth.net/confluence/display/SHIB2/IdPProdLoadTestResults
>
> On Oct 23, 2013, at 20:42 , Kevin P. Foote wrote:
>
>>
>> Some recent shib-list threads and comments hatched this idea.. so
>> bare with me. (if it sucks so be it :)
>>
>> Seems clustering is a normal topic here - as seen in the archives. We
>> all have varying ideas to its usefulness and what technologies work
>> best if its used etc.
>>
>> It seems that this can be a pain spot or at least a hurdle to those
>> new to the product. Do they need to run a cluster? If I don't have
>> multiple servers how many logins can this service handle?
>>
>> I'm wondering if it would be helpful to have a page on the wiki where
>> those running an active IdP and interested could add a description
>> of their IdP configurations. The main idea being to give a new individual
>> a gauge of where to start HW/service wise.
>>
>>
>> Something to the effect of
>>
>> Name: someu.edu
>> IdPLogins/day: 50000
>> Hardware/VM: vm,
>> OS: rhel5,
>> Memory: 4g,
>> Cluster: yes,
>> IdP-Nodes: 2,
>> Clustering Method: memcache,
>> SSL Offload: yes,
>> Loadbalancer: hardware yes,
>>
>>
>> Again just an idea... I know much of this is live and learn as well
>> as dependent on userbase / environment.
>>
>> ------
>> thanks
>>  kevin.foote
>> --
>> To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net
>
> --
> To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net