Date

8-1-2019

Document Type

Thesis

Degree

Master of Science

Department

Computer Science

First Adviser

Liang Cheng

Abstract

In this paper, we present improvements to a network simulation tool that allows us to more accurately determine the maximum end-to-end delay for data flows of a cyber physical system. We discuss the existing solutions for determining end-to-end delays of a network and the flaws in these methods. Then we explain how the development of a simulation tool can enhance the transition from traditional systems, with a requirement for the maximum end-to-end delay, to safe cyber physical systems. We briefly discuss the field of mathematics called network calculus to show the ideas behind the simulation. We then discuss the base simulation tool and its uses in calculating end-to-end delay. We also mention the existing and potential optimizations to the simulation tool concerning the use of advanced data structures. These include the use of Fibonacci heaps and b-trees as a way to improve performance and add new functionality. Our first contribution is a simulation modal that is an improvement on the existing tool and is more accurate and better able to module the traffic flows of a network. We do this by providing a queueing implementation that allows us greater control over the service flows of network switches. The improvement also offers additional support for different network disciplines and a starting point for future network discipline implementations. Our second contribution is a simulation tool that can simulate time sensitive networks using a deficit round robin scheduler. We then present a case study using the simulation tool on a tandem network and compare the results against the existing methods of calculating end-to-end network delay. We show that the results generated are more precise and tighter bounded than other methods.

Share

COinS