Vertical Parallelization of Differential Evolution Heuristic for Network Slicing in 5G Scenarios

Abstract

The 5G mobile network is based on a virtualized infrastructure and offers a virtual network (VN) creation service considering many new scenarios arising from the 5G vision. The diversity of scenarios and the instantiation of VN on-demand induce pressure on the virtual network embedding (VNE). VNE is the mapping of virtual nodes and links to real nodes and links obeying the QoS parameters present in the VN request and available resources. Since this is an optimization and NP-Hard problem, multiple efforts have been made to create VNE algorithms. Considering such efforts, this work presents: (i) a fitness function regarding multiobjective optimization; and (ii) a Parallel Differential Evolution (PDE) approach to face the VNE. We designed the PDE due to the lack of viable parallel solutions in the 5G scenario. We compared our approaches with different versions of Greedy, Stress, and Genetic Algorithms, totaling ten approaches. The results demonstrate that DE and its parallel version obtained a higher number of mapped requisitions. Also, the parallel performance decreases the execution time in certain conditions; in a favorable scenario, the parallel version obtains up 21.04% of runtime reduction.