Optimal Flow Allocation and Linear Network Coding Design for Multiple Multicasts under the Requirements of Information Theoretical Security.

Weiwei Ding,Jin Wang,Xiumin Wang,Ruimin Zhao,Yanqin Zhu

DOI: https://doi.org/10.1109/trustcom.2016.0112

2016-01-01

Abstract:Linear network coding (LNC) has great advantages in optimizing throughput, balancing network load, increasing robustness, etc. Since the coded data packets transmitted in the network are linear combinations of original data packets, the attacker cannot decode to obtain the original data when it cannot acquire enough coded data packets. Therefore, LNC can also provide efficient information security against various attacks in a communication network when not using the secret key to encrypt and decrypt. In this paper, we will study the flow allocation and LNC design for multiple multicasts to maximize the secure transmission rate (STR) and minimize the bandwidth cost (BC) under the requirements of information theoretical security (ITS), which is referred to as the Information Theoretically Secure Multiple Multicasts (ITSMM) problem. Firstly, we analyze the ITSMM problem theoretically and prove that it can be transformed to a constrained maximum network flow problem. Then, based on the theoretical analysis, we model the ITSMM problem as two linear programming (LP) formulations, which can firstly obtain the maximum STR, and then find the optimal flow allocation to minimize total BC under the maximum STR guarantee. Finally, we conduct extensive simulations and show the effectiveness of the proposed algorithm.