| Popis: |
Authentication and blacklisting mechanisms have a key role for service providers to deliver the service to correct users through digital channels. On the other hand, a user may reveal private data through the service, and an authentication/blacklisting mechanism that identifies the user may be used to link such private data to the user herself. Thus, there always have been concerns about privacy of the users. While there are previous works in the literature that provide unconditional anonymity to users, these schemes prevent service providers from blacklisting misbehaving users. At this point, the conditional anonymity concept is proposed as a remedy. A recent approach in the literature for conditional anonymity is blacklistable anonymous credentials, which allows service providers to blacklist users for an authentication session without identifying the user. In this thesis, we improve user anonymity in conditionally anonymous schemes using two complementary mechanisms. First, we define a property, whitelisting property, for blacklistable anonymous credentials and give a construction of this scheme. The whitelisting property can be used to unlink an honestly behaved authentication session from the user. Secondly, we propose an extension of this scheme for a more specific use case, sharing economy services. This scheme allows a service provider to blacklist a user only if the user have not returned the shared asset in due time. We benchmark the performance of our schemes by comparing them with the rival schemes. For communication and computation metrics, our experiments show that our first scheme has comparable performance to previous works, and our second scheme is advantageous compared to a rival one. |
