Joint VNF placement and SFC scheduling in cloud-Edge system

Network Function Virtualization (NFV) transitions network functions from hardware devices to software-defined Virtual Network Function (VNF) instances, creating a new network paradigm. VNF instances must be sequenced into Service Function Chains (SFCs) to deliver network services. Therefore, joint decision-making on VNF deployment (VNF-D) and SFC scheduling (SFC-S) is crucial. However, the complexity of this issue is amplified by the limited resources available in cloud-edge systems and the dependencies among multiple VNF instances. This paper presents a dual-time framework that separates our approach into frame and slot times to effectively address these challenges. In frame time, VNF-D is modeled as a Binary Integer Linear Programming (BILP) problem, solved by a cost-minimizing Resource-Aware Deployment Algorithm (RADA). In slot time, we introduce a Delay-Aware SFC Scheduling Algorithm (SRSA) that combines greedy methods with random pit-jumping. Our algorithms demonstrate significant improvements in resource efficiency and latency reduction, surpassing existing methods. Specifically, we achieved a 20 % reduction in resource consumption, an 18 % decrease in latency, and a 15 % increase in acceptance rates.