Energy-Efficient Service Chains with Network Function Virtualization

Service Function Chains (SFCs) are an ordered sequence of network functions, such as firewall. Using the new approaches of Software Defined Networks and of Network Function Virtualization (NFV), the network functions can be virtualized and executed on generic hardware. To optimize network management, it is thus crucial to place dynamically the network functions at the right positions in the network according to the network traffic. In this paper, we consider the problem of SFC placement with the goal of minimizing network energy consumption. We model the problem as an Integer Linear Program, which can be used to solve small instances. To solve larger instances, we propose GreenChains, a heuristic algorithm. We exhibit the benefit of dynamic routing and of NFV on the energy savings. We show that between 30 to 55% of energy can be saved for typical ISP networks, while respecting the SFC constraints.