Service Categories (or Entity Categories) help managing the secure release of the right attributes to only appropriate Service Providers – within the eduID.at federation and beyond. The use of the Service Categories documented below constitutes global Best Current Practices for controlled and scalable attribute release policies/decisions. The ultimate responsibility for the release of any personal data to third parties under these terms rests with the institution releasing the data, though.
See also Attribute release in our Shibboleth IDP documentation.
Inspired by prior work by RENATER the international Identity Federations community (lead by InCommon, REFEDS and GÉANT/eduGAIN) has started to create Service Categories (technically called "Entity Categories", since they apply to SAML "entities") in order to ease the management of release of personal data ("attributes") to services within and across large-scale multi-party federations: It has long been apparent that manually writing and maintaining attribute release rules based on individual services does not scale sufficiently, esp. taking into account that e-research today requires international collaboration and does not stop at national or "federation" borders. Dealing with individual requests to access services on a case-by-case basis creates too much work for institutions and their staff, often resulting in students or scholars not being able to access needed (inter-)federated resources because the institiutional SAML Identity Provider did not release the needed attributes to the accessed service.
Entity Categories solve this problem by grouping related services into clearly defined categories (taking into account purpose, risk/benefit and legal requirements) and scaling the management of those categories (i.e., which services are part of which category) across Identity Federations. |
Below you'll find copy&paste-able attribute release filter rules for the widely used Shibboleth Identity Provider software for each of the most important Service Categories. Be sure to carefully read and understand each category's description/definition before use.
Make sure to get the use of these rules within your IT systems approved by management of your institution. Service Categories allow for such approval processes to occur only once per category, covering potentially unlimited numbers of services belonging to that category. |
People managing technical infrastructure (such as servers and Shibboleth configuration files) should not be responsible for decisions resulting in the (non-)release of personal data to third parties. Contact ACOnet with any questions regarding Service Categories!
REFEDS has updated its guidance on the justification for attribute release with regard to the use of Service Categories (or Entity Categories) and the "REFEDS Research & Scholarship" category in particular. This now covers use under GDPR as well. |
Membership in the category REFEDS R&S is reserved to services "that are operated for the purpose of supporting research and scholarship interaction, collaboration or management, at least in part". This globally applicable category takes a risk-based approach to enabling access to high-benefit/low-risk services, releasing only low-risk personal data. Basically only the minimum personal data required for scientific collaboration and attribution of a person's work is released (name, email address and an identifer).
REFEDS R&S is purpose-limited: Only services fitting the purpose requirements may apply, and those services will only recieve a very limited set of low-risk attributes. It is being used gobally. Here's an up-to-date list of all REFEDS R&S services available to eduID.at members. |
The GÉANT (now: REFEDS) Data Protection Code of Conduct for Service Providers finally has been updated for GDPR. We'll be updating our documentation soon. |
As part of the GÉANT Data Protection Code of Conduct's Cookbook you'll find the Recipe for a Home Organisation providing complete instructions on the steps necessary for deployment. This Service Category only applies when the Service Provider (as well as the Identity Provider, which trivially will be case for all eduID.at Identity Providers) is based in the EU/EEA or countries with adequate data protection and uses the EU Data Protection Directive 95/46/EC as common frame for disparate implementations thereof throughout the EU. As such it is mostly meant as a reminder and a reassurance to both service owners and home organizations that the services covered are already subject to (national implementations of) EU data protection law.
GÉANT Data Protection Code of Conduct is legislation-limited: Only services from specific legislations may carry this category and only if they submit themselfs to the GÉANT Data Protection Code of Conduct and all its requirements. Here's an up-to-date list of all GÉANT Code of Conduct services. |
As this Category definition does not specify an attribute bundle (i.e., it doesn't reference one set of attributes which should be released to all category members) the set of attributes to be released may vary from service to service. The data to transmit under this category is limited to attributes "that are necessary for enabling access to the service provided by the Service Provider" (2.b, "purpose limitation"), though. In practice only a limited set of data will be exchanged within/across academic Identity Federations today: That could include a person's name, email address, identifier(s) and role infomation ("affiliation", such as "student" or "staff"), but could also be less than that if the service needs less data to perform appropriate access control.
The confguration below is an example based on the most commonly used attributes in Identity Federations today which most/all eduID.at Identity Providers should be able to generate. I..e, this constitutes the upper limit of what an IDP would release to Service Providers requesting data under the GÉANT Data Protection Code of Conduct category.
Again, be sure to check out the more general attribute release documentation in our Shibboleth IDP documentation, which contains more ready-to-use examples and approaches.
As part of the European Commission's work in the context of the Erasmus+ programme (Student Mobility, Erasmus Without Paper/EWP, etc.) a data structure called the European Student Identifier (ESI) has been defined. Unfortunately it wasn't designed in a way that would allow Service Providers to request it from Identity Providers using standard methods provided by SAML 2.0 Metadata (via the RequestedAttribute element). So in order to enable its scalable use across Europe additional specification work was required. The benefit of that additional specification work is that there now exists a lightweight framework to control the scalable and secure release of the ESI (and related data), governing Service Providers should be eligible to request and consequently receive this data from academic institutions such as those participating eduID.at and eduGAIN.
We're also using this category to control release of the Erasmus Without Paper (EWP) admin entitlement to keep things simple:
Again, be sure to check out the more general attribute release documentation in our Shibboleth IDP documentation, which contains more ready-to-use examples and approaches.