Secure Network Coding: Bounds and Algorithms for Secret and Reliable Communications

Network coding allows the routers to mix the information content in packets before forwarding them. This mixing has been theoretically proven to maximize network throughput [1, 44, 40, 29]. For multicast communications it can be done in a distributed manner with low complexity, and is robust to packet losses and network failures [24, 46]. Furthermore, recent implementations of network coding for wired and wireless environments demonstrate its practical benefits [32, 20]. But what if the network contains malicious nodes? Nodes that tap the network aim to eavesdrop on ongoing communication. Further, some nodes may pretend to forward packets from source to destination, while in reality they inject corrupted packets into the information flow. Since network coding makes the routers mix packets’ content, a single corrupted packet can end up corrupting all the information reaching a destination. Unless this problem is solved, network coding may perform much worse than pure forwarding in the presence of such malicious adversaries. This chapter addresses the task of multicast communication using network coding in the presence of passive eavesdroppers and active jammers. Rather surprisingly, it will be shown that high-rate private and reliable communication via schemes that are both computationally efficient and distributed is possible in the settings under study. Despite the complexity introduced by distributed network coding, it turns out that many of the classical results for private and reliable communication over point-to-point links have direct analogues in the network setting. This chapter summarizes almost a decade of study in the very dynamic and intriguing field of private and reliable network communication.