Application-Aware Adaptive Provisioning in Virtualized Networks

Network virtualization is a feasible solution to tackle the so-called Internet ossification by enabling multiple virtual networks (VNs) running simultaneously on top of a shared physical infrastructure. Network management with virtualization support, however, poses challenges that need to be addressed in order to fully achieve an effective and reliable networking environment. One of the main aspects related to the management of network virtualization environments is virtual network provisioning. Unfortunately, current provisioning solutions focus on a single or a limited set of objectives that may not simultaneously match the requirements of an increasing number of applications deployed in networks everyday. In this thesis, we formulate the Application-Aware Virtual Network Provisioning Problem (AVNP) and propose an adaptive provisioning framework for virtualized networks that takes into consideration the characteristics of multiple applications and their distinct performance objectives. The proposed framework is based on the concept of allocation paradigm, which is defined as a set of provisioning policies that guide the resource allocation process. A paradigm translates objectives from both Infrastructure Providers (InPs) and Service Providers (SPs) to individual allocation actions that actually provision VNs. To determine the efficiency of a particular paradigm, we propose a virtual network performance computation model to quantify the performance of allocated VNs and guide paradigm changing decisions. Simulation results show the feasibility of allocation paradigms in helping network providers to select the best provisioning strategy given a set of InP/SP objectives.