Unraveling Internet identities : accountability & anonymity at the application layer

Abstract

Both anonymity and accountability play crucial roles in sustaining the Internet's functionality, however there is a common misconception that increasing the anonymity of Internet identities necessitates diminishing their accountability, and vice-versa. This thesis argues that by implementing accountability mechanisms and anonymity protections at the application layer of the Internet, rather than the network layer, it is possible to develop a variety of different types of accountable-anonymous virtual identities tailored to meet the needs of the great diversity of online applications. Examples are drawn from case studies of several identity mechanisms used by existing applications, including e-mail, the virtual community Second Life, the Facebook social network, and the review site Yelp. These case studies focus on potential "points of control" for each application, as well as the ways different proposed identity schemes can leverage these control points to help mitigate the problems observed in existing identity frameworks, especially the issue of "discardable identities," or online identities that can be easily and cheaply disposed of and replaced. General design patterns for implementing accountability are discussed, with particular emphasis on the design of application-specific identity investment-privilege trade-offs, conditional anonymity schemes, and aggregated, identity management systems, as well as the role of scoped identities and linked identities in promoting online accountability.