Towards Latency Aware Multi-joint Optimization Method for VNF Placement and SFC Routing Via Swarm Intelligence

Abstract

A coherent placement of virtual network functions (VNFs) permits the efficient forwarding of data flows (routing). VNFs' placement and service function chain (SFC) routing (VNF-PSFCR) in an optimized way to minimize latency has been reported as NP-hard problem. The review of related work highlights the limitations of this domain, which revolve around high time complexity, high delays, and the ignorance of utilization of bandwidth and power consumption. The main consideration of this paper is to solve joint optimization of VNF-PSFCR and envisage the latency requirements of 5G networks. The optimal selection of VNFs and a new routing strategy are required to solve this multi-joint optimization problem via swarm intelligence. Swarm intelligence algorithms, as inspired by the collective behaviors of swarms that offer robust and high-quality solutions, have the viability to solve the mentioned problem effectively compared to conventional algorithms. A novel fuzzy heuristic and swarm intelligence-based algorithm named latency aware multi-joint placement and traffic routing-grey wolf optimizer (LAMPTR-GWO) is proposed in this work to solve the latency minimization problem. The proposed algorithm comprises two phases; the first is the efficient placement of VNFs in the graph, and the second phase is SFC routing. The Takagi Sugeno Kang system (TSK) is employed to guide the wolves to potentially explore the search space and to handle the uncertainties inherent in the NFV infrastructure. It is an effective combination that can achieve the required balance between the exploration and the exploitation of GWO. The proposed algorithm is expected to have the ability to surpass the performance of other swarm algorithms via overcoming the limitations of the GWO for solving the VNF-PSFCR problem.