When Should I Use Network Emulation?

TL;DR

This paper introduces fundamental principles and practical approaches for network emulation, emphasizing centralized methods and an architectural model to help newcomers understand when and how to use emulation tools effectively.

Contribution

It provides an architectural framework for network emulation systems and compares various modeling approaches, filling a gap in introductory and pragmatic literature.

Findings

Centralized emulation approaches are cost-effective and easy to implement.

An architectural model with hardware, impairment, and model levels is proposed.

Different modeling techniques are analyzed for their suitability to various needs.

Abstract

The design and development of a complex system requires an adequate methodology and efficient instrumental support in order to early detect and correct anomalies in the functional and non-functional properties of the tested protocols. Among the various tools used to provide experimental support for such developments, network emulation relies on real-time production of impairments on real traffic according to a communication model, either realistically or not. This paper aims at simply presenting to newcomers in network emulation (students, engineers, ...) basic principles and practices illustrated with a few commonly used tools. The motivation behind is to fill a gap in terms of introductory and pragmatic papers in this domain. The study particularly considers centralized approaches, allowing cheap and easy implementation in the context of research labs or industrial developments. In addition, an architectural model for emulation systems is proposed, defining three complementary levels, namely hardware, impairment and model levels. With the help of this architectural framework, various existing tools are situated and described. Various approaches for modeling the emulation actions are studied, such as impairment-based scenarios and virtual architectures, real-time discrete simulation and trace-based systems. Those modeling approaches are described and compared in terms of services and we study their ability to respond to various designer needs to assess when emulation is needed.