Management Digitaler Identitäten

Hartmut Bauer (Herausgeber), Christiane Büchner (Herausgeber), Lydia Hajasch (Herausgeber)

To prevent the increasing number of identity thefts, more than a douzend technologies are already existing. They have, especially then authentication with passwords, specific disadvantages or advantages, respectively. The authors of this survey analyzed how to combine these communication standards and protocols to provide more security. They recommend new kinds of identity management systems that are flexible for different user roles and are more convenient to use as the existing systems. As a first step to build such an identity management platform the authors describe how to analyze and use the individual behavior of users or objects. As a result sensor data of mobile devices are analyzed. Such devices are internetready mobiles, fitness tracker and smart watches. Therefore devices that users often carry with them. The researchers describe how these little computers can continously analyze movement patterns, data of location and connected networks and compute a trust level from the data. With this trust level, a device can indicate the probability that the current user is the actual owner, because it knows the behavioral patterns of the owner. If the current trust level value (not single biometric data) is send to an external entity like an identity provider, this provider can provide the trust level to all services used by the user. Each service is able to decide which trust level value is necessary for user authentication. If the trust level drops under a this specific threshold the identity provider can deny the access to itself and all other services. The particular advantages of this identity management approach is that no special and expensive hardware is needed but instead smartphone and wearables to evaluate the specific data. Even objects like machines that send data of their own behavior to the internet can be used. The data is continously collected in the background so users do not need to care about it. The data is only used for computing the trust level and never leaves the device. If a user logs into an internet service he does not need to remember a secret anymore, e.g. a password, instead he just needs to give an OK to pass on the trust level. If the user behavior is changing, for example by different movement patterns or unknown or new locations when trying to log into a web services, it can be immediately detected and the access to the smartphone or internet services an be locked for the unauthorized person. In future the evaluation can be extended for example with detecting routines on working days, on weekands or on vacations. The comparisons of learned routines with live data will show if the behavior fits into the usual patterns. This survey gives a comprehensive overview of techniques in digital identity management and the related challenges. First, it describes different kinds of attack methods which attacker uses to steal digital identities. Then possible authentication methods are presented. Eventually a summary of the 15 most important protocols and technical standards for communication between the three involved players: service provider, identity provider and user. Finally, it introduces the current research of the Hasso-Plattner Institute.