A REFEDS working group focused on the software requirements for Discovery is proposed. This will be complementary to other work on Discovery but with a tight focus on requirements for SAML software projects.
As the size of metadata aggregates continues to increase, the pressure to adopt more sensible "just in time" distribution models to endpoints will continue to increase. The community already has draft standards and implementations of at least one such model [1], with at least one major deployment already in use. For at least some Identity Provider implementations, this is largely a solved problem waiting on adoption.
However, some popular Identity Provider Discovery mechanisms currently rely on the use of metadata aggregates to populate lists of Identity Providers with names and logos. Adoption of just-in-time metadata by Service Providers that use such discovery software is contingent on either changing that software, replacing it, or changing the discovery experience itself.
The goal of this working group is to identify feature gaps in currently active software projects, including but not limited to Shibboleth, that need to be addressed in the near term in order to support a transition to just-in-time metadata models, under the assumption that the current aggregates would not be used. The gap analysis would be used to create feature requests for the projects.
It is not a goal of this working group to propose specific discovery solutions, but it will need to identify the most likely solutions to see adoption alongside new metadata distribution models in order to perform a reasonable gap analysis. In a perfect world, there would be a single answer, but this appears unlikely in a time frame useful to the software planning process.
The following terms apply to all REFEDS Working Groups:
TBC
Work to conclude by the end of 2017.
[1] https://datatracker.ietf.org/doc/draft-young-md-query/