Besides personal data originating from your institution, SURFconext also passes a specific piece of personal data on to services you login to which is being generated by SURFconext itself. SURFconext does this to better protect your privacy. Please read below how this works in practice.

Pseudonymous identifier

SURFconext protects your privacy by organisational and technical measures. One such technical measure is to pseudonimise your identifier. A service needs a unique identifier for each user logging in to be able to distinct between one user or another. However, most identifiers SURFconext receives from your institution usually contain more personal information than strictly necessary for a service to uniquely recognise you (such as first and last name). So instead of blindly passing on such an identifier, SURFconext generates a pseudonym to prevent more personal information from being sent to a service than strictly necessary. An example of what such an identifier looks like is: 24d66f51ac1c0b140e617af335b9abb4b8d88a5b (instead of john.smith@universityofharderwijk.nl).

There are 2 different types of this attribute:

eduPersonTargetedID (transient)

The transient type ensures the value of the attribute is different every time you login, which means that the service is not able to recognise you as a returning user the next time you login. In other words, each time you login to the (same) service, the identifier changes. This way, the service is unable to track your usage of the service.

eduPersonTargetedID (persistent)

For the persistent type, the value of the attribute is the same every time you login to the same service, but different across different services. This means the service is able to recognise you as a returning user, allowing for certain features (such as saving preferences). However, the value is different for each service, which means that different services are not able to correlate amongst each other which users they see (based on this attribute only - if your email address is also released as an attribute, this correlation is possible).

  • No labels