Computational Intelligence based Analysis of Evolve Packet Core Function Realizations in Network Function Virtualization Environment

点击翻译按钮获取中文摘要

From its onset in 2012, the notion of Network Function Virtualization is gaining a substantial interest, as well as significance. Seen as a network architecture concept allowing for virtualization of network functions, Network Function Virtualization permits an effective realization of specialized networking services. In other words, Network Function Virtualization allows for implementing network functions on standard computing platforms without a need for proprietary hardware. With the growth of the mobile devices and increased utilization of mobile networks, the processes of managing essential components of Network Function Virtualization, i.e., Virtual Network Functions are of special importance. In order to ensure reliable and scalable end-to-end services, Virtual Network Functions have to be properly initiated, monitored, and controlled. The proposed project aims at adapting Machine Learning/Data Mining and Computational Intelligence methods for analyzing data representing utilization of off-the-shelf hardware components for implementing specialized Evolved Packet Core (a framework for providing voice and data over a mobile network) functions in the Network Function Virtualization environment. The project will result in methodology for processing data in order to foresee and identify performance degradation, potential failures and self-healing automation. The project is conducted together with an industrial partner, Expeto Wireless Inc.. Expeto provides expertise, as well as data collected during realization of Virtual Network Functions in various conditions.