A scatter search algorithm to configure service function chaining

Network functions virtualisation (NFV) emerges to deal with the challenges of reducing both the capital expenses (CAPEX) and operational expenses (OPEX) of cloud providers. The NFV is done by implementing network functions and providing them as software commodities. Network services are provided by chaining a set of network functions together. The key problem will be how and where network functions should be placed in the network and how traffic is routed through them. This is the well-known service function chaining problem and it is known to be NP-hard so exhaustive search algorithms have no significant benefit in large-scale context. This paper proposes a scatter search-based solution to configure service function chain. The aim is to produce the optimal number and placement of the required virtual network functions with the dynamic traffic steering over them. This is done while respecting computing and network resources constraints.