Interfederation

[Leif Johansson <>]

On 05/28/2013 10:36 PM, Tom Scavo wrote:
> On Tue, May 28, 2013 at 4:15 PM, Basney, Jim 
> <>
>  wrote:
>> * Effort to add RegistrationInfo elements to InCommon metadata.
> Minimal
>
>> * Effort to publish interfed metadata aggregate to InCommon members.
> I think you mean the effort required to make eduGAIN available to
> InCommon members, which is minimal, but I'm reminded of the old
> saying:
>
> "You can lead a horse to water but you can't make him drink."
>
> I'm pretty confident in saying eduGAIN metadata would NOT be bundled
> in the current InCommon aggregate.
Let me offer a data-point.

In SWAMID we've had the same discussion and decided to offer two
aggregates: one transitive and one non-transitive (with or without
interfederation-provided entities that is). That seems to work well.
>> * Ongoing participation in eduGAIN steering group.
> This might be significant, I don't know.
>
>> * Handling interfederation technical support, complaints, and incidents.
> This, I think, is a steep "hidden" cost. In the InCommon Federation
> today, members work with each other to resolve problems (such as
> attribute resolution) but the InCommon admin staff don't generally get
> involved. If all of a sudden InCommon admin were to start receiving an
> influx of support requests, that would be a problem since we're not
> set up for that.
>
>> * No consistent standard for entity registration.
>>   Need to check each federation's registration practice statement.
> This is akin to some members reading POPs, which we know doesn't work.
> If, OTOH, we decide that bilateral agreements between federations are
> required to bring trust to the next level, well yes, that will be a
> lot of work (and my gut tells me that work wouldn't be justified at
> this point).
>
>> * No consistent level of assurance of identities.
> You mean of individuals who submit metadata? This is part of
> registration practice statement. Are you thinking of something else?
>
>> * SPs need to negotiate attribute release directly with IdPs.
> This is out of scope, but as I've said, it is critically important.
> (FWIW, I was part of a back-channel REFEDs discussion that suggested
> REFEDs will take up the task of standardizing the R&S entity attribute
> value, which would be great.)
>
> Tom