Metaheuristics as enablers for VNF scheduling in the network slice set up process

Abstract

Network slicing refers to the capability of composing mobile networks by chaining a set of virtualised functions on top of shared infrastructures. In the research domain, special attention is paid on the problem of scheduling network slices, i.e., the challenge of managing efficiently computation resources when multiple network slices share the same infrastructure. So far, the rich toolset that has emerged from the studies on the spectrum resource management, as well as the rapid development of cloud computing, have provided the means for scheduling decisions in 5G networks. Capitalizing on the existing studies, we examine the potential of using metaheuristic algorithms for providing scheduling outputs that minimize the slice set up time. Performance evaluation results show that evolution-based approaches (e.g., a genetic algorithm) provide better overall performance than swarm-based ones (e.g., an ant colony optimiser). However, since the slice set-up process is a real-time process, the processing time that is consumed by the scheduler itself is an important evaluation factor, for which, the swarm-based approaches have an advantage.