Constrained Dynamic Virtual Network Embedding

This paper focuses on a dynamic embedding of client-constrained heterogeneous Virtual Network (VN) requests with multiple nodes and links affinity and anti-affinity requirements. For this Virtual Network Embedding (VNE) problem, we formulate an Integer-Linear Programming (ILP)-based model that achieves joint mapping of virtual nodes and links of each VN onto the dynamically updated Substrate Network (SN). This model not only meets the clients expressed isolation constraints but also includes VN request arrivals and departures to update SN information. Numerical experiments illustrate the efficiency of the proposed methods and their ability to find optimal solutions. Performance reports provide cloud service providers with insights into additional investments in nodes and links they should make to serve clients with anti-affinity requirements.