The objective of this program is to address the fundamental imbalance between the attacker and the communicator in wireless network coding system by (1) detecting the falsely injected packet at the physical layer; and (2) separating and removing only the falsely injected packet from the polluted packet to extract (restore) the true coded packet. The program thus provides two transformative perspectives: (1) exploiting the physical signals that are available for free in detecting the falsely injected packet; and (2) recycling the polluted packet instead of discarding it, thereby dramatically improving the spectrum efficiency.
Intellectual merit: The intellectual merit of this research lies in the new perspective (signal processing) in addressing the security problem facing current network coding technology. The novelty of the proposed approach is the integration of communication and security aspects, which traditionally have been studied separately. The fundamental tradeoff among the security, rate, and reliability is investigated to develop a theoretical foundation for integrating communication and security issues in network coding systems design.
Broader Impacts: The broader impacts of this project are: (1) broadening the participation of minority and women in engineering programs by developing an experimental demonstration program for local high school students; (2) impacting on emerging secure embedded/sensor/ad-hoc networks that will be applicable to a wide range of applications, including surveillance and security, health and environmental monitoring, and public safety and emergency response.
Network coding is a new paradigm of network routing that provides substantial benefits in terms of the throughput and the robustness against packet failures. These benefits can be of great value in many applications, including wireless networks, content distribution, and distributed data storage. While network coding has proven popular, there is a growing concern regarding the security of the coded packet. This is due to the mixing nature of network coding that makes it particularly vulnerable to pollution (Byzantine) attacks, in which the adversary injects false packets into the coded packet to mislead the decoder. The research supported by this award was designed to address the pollution attack at the physical layer by separating the falsely injected packet from the polluted packet to extract the true coded packet. The trustworthiness of coded packet is determined and used at the decoder to improve the reliability of decoding. We derive the theoretical limit on the probability of decoding error and the throughput in multi-source two-hop wireless networks. The result of our study shows that the proposed approach can significantly improve the throughput and robustness against packet failures over the traditional solution based on the cryptography. The proposed technological innovation offers a new solution to the pollution attack problem facing current network coding technology and addresses the limitation of current solution without requiring additional overhead and computational complexity.
