Notes

Welcome to this course titled: 'Network Performance and Evaluation'

Computer networks are an integral component of modern life so much that access to the global Internet has been recently declared as a basic human right by United Nations. Many aspects of our daily life is largely dependent on networks: business, education, entertainment, media. With computer networks playing such a central role in modern life, it is imperative that: 1) we design high-performance reliable networks and 2) we have the ability to analyze and evaluation the performance of computer networks.

This course is about analyzing and evaluating network performance. But, what is network performance exactly and how do we define it, much less calculate it? Is network performance a subjective entity (which can have different people validly measuring it differently according to their idea of performance) or is it an objective entity (which can be measured in an unambiguous way)?

Is there always a precise way to quantify performance, or is it valid, or even preferable, to sometimes approximate?

Can we do network performance evaluation in a scientific way?

This course is about asking these questions and then progressively answering them.

Network performance is both subjective and objective. We would like to make network performance measurements as objective as possible but no further since in certain cases qualitative reasoning, in contrast to quantitative reasoning, would suffice. We would see that it is often preferable to approximate rather than to toil over unnecessary precision. Yes, we can do network performance evaluation in a scientific way but we will see that application of a scientific method invariably involves some art and judgement on our behalf.

The goals of this course are threefold: Firstly, to introduce the students to the scientific method of performing performance evaluation of networks in an objective, reproducible, and rigorous manner using properly designed experiments and sound statistical techniques for data modeling and analysis. Secondly, to facilitate the students in getting skilled in the art and craft of building purposeful abstracted models that fit their analysis/ design requirements since the application of any scientific method invariably involves some art and judgment based on intuition.  Lastly, an important aim is to develop critical thinking skills in the students so that they can perform original work in any of the sub-fields of computer and communication networks while using the performance evaluation skills for methodical analysis and evaluation.

Course Stuff:

Course material:

Course book:

Course guidebook:

Course sites:

Course breakdown:

Part 1: Science and Art of NPE

Part 2: Empirical NPE

Part 3: Modeling and Simulation