Enhancing automated network function onboarding through language extension and code refactoring

Enhancing automated network function onboarding through language extension and code refactoring

연구 분야: Cryptography

학회: Software and Systems Modeling

The management of modern communication networks requires sophisticated, specialized tooling that, akin to an operating system, provides an appropriately abstracted view of the network and its resources together with a broad and extensible range of functionality to observe, analyze, and change different aspects of the network such as software-defined networking. This research explores the process of extending network management platforms, focusing specifically on a management platform in use at TELUS, a leading Canadian communication and IT company. The TELUS Intelligent Network Automation and Analytics (TINAA) is a model-driven control and management ecosystem. TINAA uses MDE techniques and YANG, a data modeling language widely used in the networking domain, to facilitate the implementation of new network functions. More specifically, to add a network function to TINAA, the following process is used: (1) The function is described using YANG. (2) Skeleton code is generated from these models. (3) The skeleton code is modified and completed as appropriate by a developer. In particular, we demonstrate how YANG’s language extension mechanism, combined with code clone detection and removal, can capture essential information at the model level and generate cleaner, more readable skeleton code, significantly simplifying the manual code completion process. We evaluate our techniques using the L3VPN Network Model (L3NM), a widely adopted model for managing and configuring Layer 3 VPN services. We identify general lessons learned of interest to both the networking and the MDE community.