Resource Provisioning for Virtual Network Function Deployment with In-Subnetwork Processing

Network Function Virtualization (NFV) is an efficient
approach to simplify and accelerate the deployment of diverse network services. In this paper, we study the problem of link capacity
and server purchase to provision the network which enables joint
Virtual Network Function (VNF) placement on servers and traffic
engineering in a network spanning multiple subnetworks. Each
subnetwork is owned and controlled by a different administrator.
Given flow demands, we formulate the problem from a sparse
optimization perspective and propose an efficient approach based
on iteratively solving a sequence of group LASSO problems. Using
purchased link capacities and servers, each subnetwork is able to
carry out VNF placement and traffic engineering locally to meet
flow demands. A scalable and decentralized approach based on the
proximal Alternating Direction Method of Multipliers (ADMM)
is proposed for this problem. The distributed optimization can
be locally solved with minimum information shared with other
administrators. Extensive numerical evaluations show the efficiency
of our approach against existing work.