Integration of Discovery Service

[Chris Phillips]

Good to see discussion on this.

For us the central cookie and a transition path away from SAML1 is/was
important. Having a HA infrastructure to host the CDS already in place
also helped.

Looking to the future I see an emerging trend to something similar to what
LIGO has done -- discovery by region or arbitrary (research) groupings.
Not just one, but many of them stemming from the same metadata but
operated centrally.

For example, only show me IdPs in Province X or Research consortium Y that
have done something above and beyond the regular federation attribute
release requirements and signalled as such in the metadata via entity
categories.


Use cases to illustrate this are:
- Have a Discovery service of only those IdPs supporting the R&S category
to be used by Services Supporting R&S. (could be any category actually)
- Regional groupings for a specific province/state to only show the IdPs
within said region for services for that region.  This can easily be done
by a Federation assigned and marshalled entity category (e.g. inCommon-NJ
or CAF-Ontario) 

Both cases likely desire to be configured with some co-branding within
that category or region.


Thinking aloud a little bit, one could advocate for a single discovery
code base to support a user experience that can be customized for any
given entity category tag -- a custom filter list of IdPs, list of Sps
that support the tag, an additional logo, tag specific description.

I think one would rather operate a single code base that can be
configurable to observe N entity categories (or lack of one as the base
case for 'display everything') and facilitate such an experience.

Where would could one execute on this?

In the CDS codebase along the lines of Shib v3 velocity style templates
maybe?
In the EDS codebase relying on a sophisticated SP installation v-hosted
per entity category to accomplish filtering?
In the PHP codebase based on SWITCH's approach?
In the DiscoJuice Model?

If others say 'Hey, you can do that already by . . .' let me know.

If this style of Discovery appeals to others, reach out to me to let me
know as well.

I'm also more than happy to submit RFI's to the CDS or EDS along these
lines too. 
That said, I'm sure it would have more 'weight' if more than one group
voiced an interest in this direction and it could be more precisely
defined.


Thoughts welcome as always.

C





On 2015-05-05, 9:08 PM, "Cantor, Scott" <cantor.2 at osu.edu> wrote:

>What specifically about the CDS made it a better choice than running the
>EDS in a shared fashion?
>
>-- Scott