A Framework for Verification-Oriented User-Friendly Network Function Modeling

Abstract

Network virtualization and softwarization will serve as a new way to implement new services, increases network functionality and flexibility. However, the increasing complexity of the services and the management of very large scale environments drastically complicate detecting alerts and configuration errors of the network components. Nowadays, misconfigurations can be identified using formal analysis of network components for compliance with network requirements. Unfortunately, formal specification of network services requires familiarity with discrete mathematical modeling languages of verification tools, which requires extensive training for network engineers to have the essential knowledge. This paper addresses the above-mentioned problem by presenting a framework designed for automatically extracting verification models starting from an abstract representation of a given network function. Using guidelines provided in this paper, vendors can describe the forwarding behavior of their network function in developer-friendly, high-level languages, which can be then translated into formal verification models of different verification tools.

Keywords

• Network function modeling
• model extraction
• NFV