Profile Picture

Dr. George N. Rouskas

Professor and Director of Graduate Programs
IEEE Fellow

Dr. George N. Rouskas

Professor and Director of Graduate Programs
IEEE Fellow

CSC/ECE 579 — Introduction to Computer Performance Modeling

University Community Standards and Personal Safety Requirements Rule related to COVID-19.

Lecture slides, assignments, and solutions are available from the course Moodle space

Lecture videos are available from the EOL site

Zoom session — only available during scheduled class times

Fall 2020 Schedule of Lectures

Date Lecture # Topic Assignment Due
Aug 10 1 Overview, goals, logistics
Aug 12 2 Introduction HW 1
Aug 17 3 Review of probability theory
Text: Appendix II.1-II.3
Aug 19 4 Review of probability theory (cont'd)
Aug 24 5 Review of Laplace and z transforms
Text: Appendix I
Project 1
Aug 26 6 Introduction to simulation HW 2 HW 1 due
Aug 31 Classes cancelled/postponed by the University
Sep 2 7 Introduction to simulation (cont'd)
Random number generation
Sep 7 8 Random number generation (cont'd)
Sep 9 9 Discussion of Project 1
Simulation design
HW 2 due
Sep 14 10 Estimation techniques
Sep 16 11 Estimation techniques (cont'd)
Sep 21 12 Poisson process
Text: pp. 60-71
Sep 23 13 Markov processes
Text: 2.4
Project 2 Project 1 due
Sep 28 14 Birth-death processes, Birth-death queueing systems
Text: 2.5
Discussion of sample exam problems
Sep 30 Midterm Exam HW 3
Oct 5 15 Discussion of midterm exam
Discussion of Project 2
Oct 7 16 Little's law
M/M/1 queueing system
Text: 2.1, 3.1, 3.2
Oct 12 17 M/M/1 queueing system (cont'd)
Oct 14 18 Simple Markovian queueing systems
Text: 3.3-3.6
Project 3 Project 2 due
Oct 19 19 Simple Markovian queueing systems (cont'd)
Discussion of Project 3
HW 4 HW 3 due
Oct 21 20 The method of stages
M/Er/1 queueing system
Text: 4.2, 4.3
Oct 26 21 Bulk arrivals, Coxian distributions
Text: 4.5, 4.7
Oct 28 22 M/G/1 queuing system
Mean queue length
Text: 5.1, 5.3, 5.5
Nov 2 23 Mean queue length (cont'd)
ARQ protocol analysis
M/G/1 queue with vacations
Nov 4 24 M/G/1 queue with vacations (cont'd)
Pollaczek-Kinchin Transform Equations
Text: 5.6, 5.7
HW 4 due
Nov 9 25 Priority queueing Project 3 due
Nov 11 26 M/G/1 conservation laws
Nov 18
8:30-11am
Final exam

 

Syllabus

Prerequisites

Students who wish to take this course must have completed a course on Probability Theory (MA 421 or equivalent) and a course on Computer Organization (CSC 312 or ECE 218 or equivalent).

Students must also have good working knowledge of a high-level programming language such as C, C++, or JAVA. The programming projects can be challenging, hence good programming experience is required.

Objectives

The purpose of this course is to present simulation techniques and queueing theory as tools for modeling and studying the performance of communication networks and computer systems.

At the conclusion of the course you should be able to:

  • apply simulation techniques to develop models of computer and communication systems;
  • appy queueing-based models to characterize computer and communication systems;
  • use appropriate analytic tools to compute performance measure of interest (e.g., response time and throughput) for a given queueing system;
  • select the system characteristics (e.g., storage capacity) to achieve a given level of performance;
  • evaluate the relative merits of alternative system design solutions; and
  • engage in research in the field of performance analysis and evaluation.

I encourage and expect you to participate actively in the learning process. In particular, I welcome your comments and questions as we cover material in class. One-way lectures quickly become boring, both for you and for me. By asking lots of questions your understanding of the material will be deepened significantly, and the course will be much more fun!

Outline

The course is logically divided in three parts.

Part I: Refresher.
At the beginning of the semester we will review important concepts from probability theory and Laplace and z transforms.

Part II: Simulation Techniques.
This part addresses the development of simulation models, including:

  • generation of random numbers and stochastic variates
  • simulation designs
  • estimation techniques for analyzing endogenously created data
  • validation

Part III: Queueing Theory.
This part introduces a number of fundamental concepts and techniques, including:

  • stochastic processes and Markov processes
  • Poisson process
  • birth-death processes
  • the M/M/1 queue and variants
  • Erlang and Coxian distributions as models of service time
  • the M/G/1 queue
  • priority queueing and conservation laws

Textbook

Students are required to purchase the following textbook:

  • L. Kleinrock, Queueing Systems, vol. 1: Theory, Wiley. ISBN: 0-471-49110-1

I also suggest the following two books as reference:

  • L. Kleinrock, Queueing Systems, vol. 2: Computer Applications, Wiley
  • W. Drake, Fundamentals of Applied Probability Theory, McGraw-Hiil (or any other book on probability theory and transforms)

I will also make available an extensive set of lecture slides.

Grading

Students are required to complete all assignments and show all work in order to receive full credit. The final grade will be determined using the following weights:

  • 45% — Three programming projects (15% each)
  • 15% — Homework assignments (of equal weight)
  • 20% — Midterm exam (open book)
  • 20% — Final exam (comprehensive, open book)

Policies

Attendance: Attendance is not mandatory but strongly encouraged. Students are responsible for making up any course material they miss.

Assignments: No hard copies of assignments or solutions will be handed out. New assignments and solutions will be announced in class and/or the course mailing list, and will be available on the course web page.

Submission: Students must submit their assignments as PDF or Word files using the submit facility. The deadline for submission is midnight (Eastern time) on the day due. Any deadline extensions are up to the discretion of the instructor, and will be announced to the whole class. Extensions may be provided to individual students only in advance of the submission deadline and only under extenuating circumstances.

Late Submission: No late assignments will be accepted and no partial credit will be given for late assignments without a valid excuse.

Cheating: Homework and projects are individual assignments and students are required to submit their own solutions. All students are bound by the University's academic integrity policies (refer to the relevant section below).

Teaching Assistant

Xian Yang (xyang45@ncsu.edu) is the TA for this course.

You may contact him to arrange for an online chat or video call at a mutually convenient time.

Feel free to contact the TA for any questions about the course.

Office Hours

My office is in Room 2306 of the EB II building.

Please email me to arrange for a mutually convenient time to have a discussion over the phone or online chat.

Academic Integrity

Students are required to respect the NC State academic integrity policies.