Content-centric Information Protection in Cloud Computing

Abstract

Information security has become increasingly important as organizations migrate systems to third-party infrastructure providers. Once migrated, however, previously transparent network topologies, information paths, and systems infrastructure became more opaque. This loss of control when coupled with storage of corporate and personally sensitive information lead to significant increases in potential vulnerability. In this paper, we present experimental evidence demonstrating the feasibility of using content centric networks with integrated policy-based usage management. We describe a nation-spanning content network developed within Amazon and Rackspace infrastructures and collect performance statistics to show the suitability of various confidentiality strategies in these kinds of large heterogeneous systems. In doing this, we first consider the current state of the art in network information security as well as some of the shortcomings of current designs, and propose a taxonomy of network-enabled usage control architectures that can solve sensitive information transmission problems. We then close with a description of our content-centric network, a discussion of our experience using this system to manage real-time sensitive information flow over commercial cloud systems, and experimental evidence demonstrating the feasibility of the approach.