A User-Centric Approach for Ranking NFV Services

Significant research efforts in virtualizing networks contributed to the formation of Network Function Virtualization (NFV) and led to the existence of a wide range of NFV service providers. Identifying relevant NFV services of interest tailored to end-user needs remains a major challenge. Given that IoT applications (e.g. self-driving cars, smart city, mission-critical applications) cannot afford to compromise the end-to-end Quality of Service (QoS), it is indispensable to consider efficient methods for enhancing the discovery of relevant NFV services. QoS attributes such as response time, bandwidth, among others, play an increasingly crucial role in differentiating between NFV services. In this paper, we present the User-Centric NFV (UCNFV) framework for improving the NFV service provisioning process based on TOPSIS, a multicriteria decision analysis method. By considering end-user and application requirements, UCNFV optimally ranks NFV services that match client requirements. Throughout the paper, we present experimental and validation results which show the usefulness of our proposed UCNFV optimization approach.