Abstract:
Secure handling and sharing of Identity credentials across multiple domains
is considered to be the gravest issue faced by Cloud service consumers and
Cloud service providers. Identity credentials are generally considered to be
the most sensitive information since its unauthorized disclosure may lead
to many serious consequences. As a result, many e orts from academia (research
community) and IT industry are underway to circumvent the misuse of
identity information. Despite hundreds of thousands of scienti c reports and
books published on this domain, the issues related to identity management
systems continue to grow in severity and sophistication. Apart from this,
identity management systems have actively followed the evolution of technology
which in turn has made their security and functionality requirements
even more diverse and dynamic. The immediate result of this continuous
arms race is that the domain of identity management has become quite complex.
This complexity is two-fold. Firstly, the massive literature on identity
management remains largely unstructured. Secondly, a cross-domain identity
management system capable enough to ensure secure management of
user credentials across multiple domains is sorely missing.
In this thesis, we explore and address the abovementioned issues by rst
structuring the knowledge in the domain of identity management in the form
of a well-organized taxonomy, and then by implementing a cross-domain
identity management system for Cloud. Comprehensive list of attacks that
are targeted towards identity or identity management systems along with a
taxonomy covering most eminent features and corresponding mechanisms to
avoid those attacks are presented. It is asserted that the proposed taxonomy
helps in making informed decisions while selecting or implementing a cross
domain identity management system for Cloud environment. Further, System
for Cross-domain Identity Management (SCIM) - an open source, extensible
and light-weight protocol for the exchange of identity credentials among
disparate identity management systems is explored and implemented. In addition
to this, we have enhanced SCIM protocol by adding an encryption
module that help ensures the con dentiality of identity credentials during
2 Abstract
transmission across multiple service providers.
We have rigorously evaluated our work from two perspectives, functionality
of the proposed system is certi ed through user de ned test-cases and for
security analysis we have chosen Scyther, a security protocol evaluation tool.
The results of our evaluation con rm that there is signi cant enhancement
in the functionality and security of SCIM protocol.