ns3-lab: a Scalable Online NS-3 Lab Platform for Learning Computer Networks

Abstract

The penetration of computer networking into the telecommunication sector has been requiring electrical engineering professionals to learn the domain knowledge of computer networks. It has been shown that it is effective for STEM college students to learn complex networking protocols by reproducing classic networking experiments in our previous pedagogical telecommunication practices. In this paper, we design an online simulation lab platform by integrating productivity-proved open-source modules for learners to practise networking protocols without tedious system and simulator configurations. The front-end of this platform applies a VUE-based framework and the back-end deploys a docker-based architecture to support elastic on-demand capacity expansion for potentially a large number of learners. We conduct an evaluation study to evaluate the system performance of this ns-3 lab prototype. We instruments two typical labs. In simulating a point-to-point communication link, an off-the-shelf typical server can support hundreds of users for parallel compilation within 10 seconds in this illustrative tutorial lab. In simulating a medium-complex WiFi network, this typical server can support 30 users for parallel compilation within approximately 1200 seconds to finish this lab. Our measurement results demonstrate the performance effectiveness of our prototype design to develop a scalable but user-friendly lab platform to learn computer network protocols in a learning-bv-doina approach.