Tabu Search for Efficient Service Function Chain Provisioning in Fog Networks

Fog computing places computation/storage resources at the network edge to overcome delay limitations associated with cloud computing. Namely, fog-based services can provide more responsive relay and caching and alleviate loads at core cloud datacenters. Meanwhile, technologies such as network function virtualization (NFV) are virtualizing atomic network functions and allowing provides to build customized service function chains (SFC). Hence combining NFV with fog computing allows providers to share critical edge resources across clients and achieve much more flexible service designs. However, few efforts have looked at SFC provisioning in fog domains. Hence this paper presents one of the first studies in this area, detailing a Tabu search for virtual function placement and heuristic routing schemes with load balancing for efficient resource utilization. Results show reduced processing and propagation times, and lower energy consumption versus cloud-based methods.