Performance Analysis for Subgraph Isomorphism Based Embedding Service Function Chains

Abstract

Embedding Service Function Chains (SFCs) into the network represents scheduling network resources to provide user services. Embedding schemes are affected by many factors, such as the topology of Network Service Requests (NSRs), the demands of NSRs on Virtual Network Functions (VNFs) computing power, and network link quality. However, most researches focus on improving the performance and algorithms of embedding SFCs, and the effect of the above factors has yet to be intensively studied. Studying the influence of the above factors can not only guide users to initiate more suitable NSRs but also have guiding significance for scheduling network resources. Due to the subgraph isomorphism that can quickly find all feasible solutions, it is applied to embedding SFCs to explore the influence of the above factors on embedding SFC. However, subgraph isomorphism in large-scale networks is inefficient due to its high computational complexity. We propose an adaptive adjustment strategy to prune the network to reduce the computational complexity of subgraph isomorphism. Simulation results show that the subgraph isomorphism can search all feasible embedding schemes quickly and observe the influence of the above factors on embedding schemes. Those factors significantly impact embedding SFCs, mainly reflected in scheme numbers and latency performance.