5G网络切片中基于离散粒子群和Kruskal算法的跨域虚拟网络映射

5G mobile communication network hire more data centers of infrastructure provider to cooperate to construct network section, Aiming the efficiently cross domain virtual network mapping which is a key problem in network lifecycle management in network slice, proposeda two stage cross domain mapping strategy, DPSO-K. Firstly, based on the resource bidding, considered the resource of nodes and the bandwidth between domains, and proposed a optimized discrete particle swarm optimization (PSO) algorithm for inter domain virtual network mapping, which can effectively improve the searching ability; then, proposed a fast algorithm based on Kruise Karle minimum spanning tree to reduce the cost of intra domain mapping, which reduces the time of fragment instantiation and shortens the on-line speed of service. The comparing to the traditional method of the first virtual network mapping request division unified mapping link, the strategy considered the bandwidth overhead in the request division, and payed attention to key link in mapping mapping link, mode of centralized management, distributed control, to realize the effective use of physical and cyber source. Experimental results show that this algorithm can achieve higher acceptance rate with less overhead and less partitioning time.