Effectiveness of Traffic Camouflaging over Computer Networks

Summary form only given. For many Internet applications, the ability to protect the identity of participants and the characteristics of their communication in distributed applications is critical. For such applications, a number of traffic camouflaging systems have been developed over the past several years. The effectiveness of these systems relies greatly on (1) the protocol by which messages are (re-)routed among the participants and (2) the scheme by which links are padded. In this talk, we discuss our recent discoveries on the effectiveness of these camouflaging methods. Our results contradict some of the methods that have been commonly used. For example, we find that using more agents in re-routing may not necessarily increase the probability that a sender can be identified. Furthermore, padding links with a constant-bit rate pattern may result in the worst probability; that an adversary can identify the underlying payload status. We discuss how to develop optimal strategies for these traffic camouflaging systems