Acknowledgements

We are especially grateful to Daniel Bleichenbacher for the design, analysis, and implementation of the encoding scheme in Section 4. Additional analysis of the key reconstruction scheme presented therein will appear in [1]. We also thank Sape Mullender, Peter Bosch and Qiru Zhou for their assistance in analyzing anomalies in the kernel sound modules on the various platforms we have experimented with in this effort. We are thankful to Greg Kochanski for numerous insightful discussions, and Minkyu Lee, for providing us with signal processing code that attempts to correct the discontinuities in the cut-and-paste attacks. We also thank Monica Ullagadi for collecting data.