NFV-Optimizer: Profiling Network Virtualized Functions and Optimizing their Performance for Telecom Clouds

Due to the massive success of virtualization, more and more services will be run inside virtualized Datacenters to reduce both Operational (OPEX) and Capital Expenditures (CAPEX). Aligned with the general trend of migrating traditional IT architectures to clouds (public or private), next generation of telecommunication (Telecom) networks such as 5G are also envisaged to be run on virtualized environments where network functions are deployed through virtual machines instead of current proprietary boxes. Guaranteeing the performance of such virtualized systems is however challenging, because it requires accurate modelling and efficient optimization of such services in their new environments.

NFV-Optimizer will model the impact of virtualization on Virtual Network Functions (VNFs) and study their behaviour as well as their inter-relation complexities when they are cascaded to build higher-level frameworks for Network Function Virtualizations (NFVs). Together with our industry partners, our ultimate aim is to model, optimize, design and implement NFV specific profiling/optimization frameworks to not only study VNF Service Chains (VNF-SCs), but also efficiently optimize them to close the gap between their performance and those of currently deployed counterparts using proprietary boxes. To this end, NFV-Optimizer will identify Telecom specific NFV deployments, systematically profile them, and provide optimization recommendations to satisfy stringent constrains of 5G networks, as described by both ETSI (European Telecommunications Standards Institute), 3GPP (3rd Generation Partnership Project), and many others. Besides providing scientifically proven techniques, NFV-Optimizer will also strengthen our existing long-lasting collaborations with Ericsson, and establish new connections with VoerEir.

Eleven professionals will work on scientific problems and technical challenges of this project: four (three senior researchers and a PhD student) from Department of Mathematics and Computer Science at Karlstad University (KAU), four from Ericsson AB, and three from VoerEir. Results of the project are envisaged to be integrated into products and solution portfolios of our industry partners to immediately impact industry standards and best practices.