A Method for Adaptive Resource Adjustment of Dynamic Service Function Chain

Abstract

Service function chain (SFC) is widely exploited to provide heterogeneous network services, which refers to an ordered sequence of network functions that a given flow should traverse. However, its traditional implementation is inflexible and high-cost due to relying on proprietary hardware. Emerging technologies such as network function virtualization and software-defined networking are the two promising paradigms for improving the flexibility to implement SFC. A major challenge in this respect is the SFC resource allocation that deals with the efficient allocation of network resources to network functions and their connections belonging to SFC. Previous researches focused on heuristic algorithms for the SFC resource allocation assuming that the topology and the resource requirement of SFC do not change. In this paper, we remove this assumption and study how to adjust resource allocation scheme optimally when SFC changes, which is referred as dynamic SFC resource adjustment problem (DSFC-RAP). It is a challenging problem for various reasons, such as contradiction between reducing resource consumption and minimizing adjustment operation overhead. We first formulate the DSFC-RAP as an integer linear programming. Then, we propose a heuristic-based method to find the near-optimal solution with low computational complexity. Evaluation results demonstrate that our method allows network to accommodate more SFCs while incurring relative low network resource cost and adjustment operation overhead compared to other heuristics.

Keywords

• Dynamic service function chain
• integer linear programming
• network function virtualization
• resource adjustment
• software defined networking