Privacy Preserved Attribute Aggregation to Avoid Correlation of User Activities across Shibboleth SPs

摘要

Privacy is one of the most important issues in Identity Federation, a technology in which local IDs and credentials such as passwords managed at one site may be used to access many online services, including cloud services provided outside of users' organization. Attribute aggregation is an advanced technique that may be employed in identity federation, collecting attributes about a user from multiple distinct identities to provide a complete picture about a user necessary for some services. However, conventional methods of attribute aggregation require a persistent shared unique ID. This may restrict the use of federated identity for some services because these unique ID's could be used by bad actors to correlate user activity or user data. This paper proposes a new method of attribute aggregation that doesn't require a universal unique ID. SAML, a widely used federated identity standard, is used as the basis for this work. This privacy-preserving attribute aggregation technique has been validated with a successful implementation for the open source federated identity software project Shibboleth.