Secure ID Transformation for Robust Pseudonymity against Backflow of Personal Information in SAML Federation
Wataru Oogami, Takaaki Komura, Yasuo Okabe · 2012
Security Assertion Markup language (SAML) is used commonly by Single SIgn On (SSO) system. In conventional frameworks for SAML, a Service Provider (SP) has to log the usage history of users by an pseudonymous ID (including attributes), thus only an Identity Provider (IdP) can analyze this SP's usage logs. If the IdP can look usage logs, it can touch user's privacy information. We define such flow of information as"backflow". We propose a secure countermeasure protocol against the backflow between the IdP and the SP by using a new ID, named "hashedID". In our new approach, the IdP cannot analyze the usage logs by using an Attribute Provider (AP) as an Identifier Transformer (IdT). The IdT can secure transform from an ID that the IdP issued to hashedID. This approach enables the use of more secure SAML communication without sharing personal information that the user doesn't want the IdP to know about. In this paper, we design, implement, and valuate about the IdT and its transformation.