A big data based dynamic bandwidth allocation strategy with secrecy constraints

This paper investigates a dynamic bandwidth allocation strategy with secrecy constraints, where big data can be viewed as a resource instead of a burden from the traditional perspective. Unlike usual cases, we take into account big data and security issues along with bandwidth allocation. It is reasonable to assume that big data derived from mobile network, by a series of processing, can generate a binary set S consisting of pairs of users. According to S, a metric closeness can be redefined to describe whether the same confidential content can be shared between two users. On this basis, data driven clusters can be formed. Then two bandwidth allocation algorithms, aiming at increasing secrecy sum capacity and individual secrecy capacity by sharing content in clusters, are proposed. The fairness among users and computation complexity are considered in the first algorithm, while the objective of the second algorithm is to maximize the secrecy sum capacity. In order to validate our proposed schemes, a concise case is presented and numerical results show that a significant performance gain over both secrecy sum capacity and individual secrecy capacity is achieved.