New techniques for analyzing applied computer security systems

New techniques for the analysis of real-world computer security systems are presented. These methods were developed to allow the analysis of several specific actual systems ranging from secure hardware devices to cryptographic network protocols. Each method is presented in the context of the system it was first used to analyze.; The first system analyzed is a cryptographically enabled radio frequency identification tag that is currently used to protect digital payments and prevent the theft of automobiles. We present new techniques for reverse-engineering a crypto graphic algorithm given only black-box access. That is, uncovering the underlying algorithm given only input/output pairs. These represent the first such techniques published in the open literature.; The second system discussed is a voting machine that is currently in use in several states. By examining the source code that was used in the machine, we were able to discover several serious vulnerabilities. In particular, we located problems with the smartcard protocol and the cryptographic data protection mechanisms.; The third system analyzed is the security protocol in 802.11b wireless networks. Here we show how a cryptanalytic attack on a modification of the underlying cipher can be applied to systems that use initialization vectors and provide structured headers as early plaintext. Moreover, we show how the attack can be amplified when a nonuniform underlying distribution on individual underlying key bytes can be assumed.; The final system analyzed is a collection of candidate digital rights management algorithms. Here we show how the use of unkeyed error correcting codes can simplify cryptanalysis and reverse-engineering of security systems.