Characterizing the impact of rigidity on the security of cellular telecommunications networks

Cellular telecommunications networks are the largest, most widely used digital systems ever constructed. Such systems traditionally offer one of the most reliable infrastructure for personal, business and emergency communications. To support as many users as possible, the resources of these networks have been tightly engineered based upon historically expected usage patterns. However, the introduction of a vast range of new services and connections with the larger Internet violate the basic assumptions upon which these networks are founded. Accordingly, the evolution from largely closed, specialized systems to more open, common carrier networks significantly imperils the security and stability of these networks.;In this dissertation, we explore vulnerabilities resulting from rigidity, or the inflexible treatment of all traffic according to telephony-specific optimizations. We present several demonstrative attacks capable of preventing legitimate users from receiving any traffic (voice, text messages or data) in major metropolitan areas such as Manhattan with the bandwidth available to a single home cable modem. We then more formally characterize the impact of such attacks and measure the efficacy of countermeasures using queuing techniques and simulation, and explore the underlying cause of such attacks using traditional network security methods. In so doing, this dissertation highlights the fact that imposing deeply-rooted assumptions about the characteristics of circuit-switched flows on packet-switched traffic creates an exploitable point of amplification through which an adversary can disable these systems. We conclude with a discussion of open problems and future work in securing telecommunications networks.