Evaluation:

This page contains information on the IdP test results of the eduGAIN Attribute Release Check Service (EARCS), which allows users from an eduGAIN Identity Provider to check whether it properly releases information in form of attributes is to eduGAIN-enabled services.

The check results are reflected by the following verdicts:

Test Verdicts

IdP sends all necessary information

IdP sends minimal information

IdP sends basic information while some required information is missing

IdP sends eduPersonTargetedId with the wrong (legacy) syntax

IdP sends superfluous personal information

IdP sends some subset of the requested information, but not the basic information (see definition below)

Incorrect value syntax (except for eduPersonTargetedId above)

R&S category support is indicated but its requirements are not satisfied

No attributes received

Bonus points (A-C)

Penalty points (A-C)

Statement for the "No Entity Category Test"

For this test a Service Provider is used that does have no entity category like R&S or the GÉANT Code of Conduct but just declares the attributes eduPersonScopedAffiliation, schacHomeOrganization, email and eduPersonPrincipalName as required attributes in metadata. The results of this tests are the two statements:

"Good data privacy but bad usability":
This means that the IdP was not releasing any attributes to this test SP even though it requested them. Depending on the IdP's iurisdiction, this behaviour is rather restrictive from a usability point of view because users most likely won't get access to eduGAIN services that have no entity categories unless the IdP has configured any specific attribute release rules. Still, IdP administrators in some countries feel safer with this setup from a legal/data privacy perspective.

"Good usability but bad data privacy":
This means that the IdP released some or all required attributes to this test SP just because the SP requested them. This policy is used by few Identity Providers. It is easy to implement and in most cases is beneficial to users because they gain access to more services because their attributes are released by default. From a privacy point of view some argue that IdPs using this approach might be a bit too generous in releasing data about the user, especially in case there is no user consent enforced during the login process and especially for services that are not relevant for the users studies or job. However, so far there are worldwide no cases known in the community where IdPs got into legal issues using this approach.

Terminology