Cleveland State University

Department of Electrical and Computer Engineering

 

EEC 685/785 Modeling & Performance Evaluation of Computer Systems

Fall Semester 2005

 

Course Objectives:          This course provides a comprehensive study of analytic, measurement, and simulation methods for computer systems performance analysis. The course prepares students to understand performance terminology, correctly design performance experiments, use simple queueing or simulation models and interpret results.

 

Pre- or co-requisite:        EEC484/584, Computer Networks

 

Textbook:                         The Art of Computer Systems Performance Analysis: Techniques for Experimental Design, Measurement, Simulation, and Modeling, by Raj Jain, John Wiley & Sons, 1991.

 

Reference Book               Computer Systems Evaluation and Prediction, by Paul J. Fortier, and Howard E. Michael, Digital Press, 2003

 

Instructor:                        Dr. Wenbing Zhao     

Office:                SH317

Phone:                (216) 523-7480

Fax:                    (216) 687-5405

E-mail:                wenbing at ieee dot org

Office hours:      T Th 4:00-6:00pm and by appointment

 

Final Grade:

HW1

HW2

HW3

HW4

HW5

HW6

MT1

MT2

Final

Project

Total

Grade

10

10

38

38

38

10

78

107.5

83

18

86.8

A

9

10

38

33

45

10

78.5

102.5

81

18

85.3

A

8.5

10

38

42

38

10

61.5

75

85

21

80.1

A-

9

10

38

42

45

10

80

60

83

20

80.1

A-

10

10

36

32

41

10

64

82.5

78

14

73.8

B+

8.5

10

38

31

39

10

61.5

81.25

80.5

10

68.8

B

4

0

38

33

38

10

0

82.5

78

20

61.5

B

 


 

Week

Topics

Reading

1

(Aug 29 - Sep 2)

Lecture 1 (notes 2up, 4up): Overview of performance evaluation

Lecture 2 (notes 2up, 4up): selection of techniques and metrics

Chapter 1-2

 

Chapter 3

2

(Sep 5 – 9)

Labor Day – September 5 (no class)

Lecture 3 (notes 2up, 4up): types of workloads, the art of workload selection

 

Chapter 4-5

3

(Sep 12 - 16)

Lecture 4 (notes 2up, 4up): Monitors; program execution monitors and accounting logs

Lecture 5 (notes 2up, 4up): capacity planning and benchmarking; Data presentation (part 1)

Chapter 7-8

 

Chapter 9-10

4

(Sep 19 – 23)

Lecture 6 (notes 2up, 4up): Data presentation (part 2); ratio games

Lecture 7 (notes 2up, 4up): Probability theory and statistics – summarizing measured data

Chapter 10-11

 

Chapter 12

5

(Sep 26 – 30)

Lecture 8 (notes 2up, 4up): Comparing systems using sample data

Lecture 9 (notes 2up, 4up): Simple linear regression model

Chapter 13

 

Chapter 14

6

(Oct 3 – 7)

Midterm Test #1 – Monday, October 3

Lecture 10 (notes 2up, 4up): Work load characterization techniques

Ch. 1-5 and 7-14

Chapter 6

 

7

(Oct 10 – 14)

Columbus Day – Monday, October 10 (no class)

Lecture 11 (notes 2up, 4up): Multi-linear regression models; Introduction to experimental design; 2k factorial designs

 

 

 

Chapter 15-18

 

8

(Oct 17 – 21)

Lecture 12 (notes 2up, 4up): 2kr factorial designs

Homework #3 issued, due Oct.24

Lecture 13 (notes 2up, 4up): 2k-p factorial designs; One factor experiments

Chapter 18

Chapter 19, 20

 

9

(Oct 24 – 28)

Lecture 14 (notes 2up, 4up): Introduction to simulation

Lecture 15 (notes 2up, 4up): Analysis of simulation results

Chapter 24
Chapter 25

10

(Oct 31 – Nov 4)

Lecture 16 (notes 2up, 4up): Random-number generation

Lecture 17 (notes 2up, 4up): Testing random-number generators;

Homework #4 issued, due Nov. 9

Chapter 26

Chapter 27-28

11

(Nov 7 – 11)

Lecture 18 (notes 2up, 4up): Random-variate generation; Commonly used distributions

Lecture 19 (notes 2up, 4up): Simulation of a single server queuing system; review for midterm#2

Chapter 29

 

12

(Nov 14 – 18)

Midterm #2 – Monday, November 14

 

Lecture 20 (notes 2up, 4up): Introduction to queueing theory

Chapter

6;16-20; 25-29

Chapter 30

13

(Nov 21 – 25)

Lecture 21 (notes 2up, 4up): Analysis of a single queue (part I)

Homework #5 issued, due Nov 28

Course Project Progress Report, due Nov 28

Lecture 22 (notes 2up, 4up): Analysis of a single queue (part II)

Chapter 31

 

14

(Nov 28 – Dec 2)

Lecture 23 (notes): Queueing network,

Lecture 24 (notes): Operational laws

Chapter 32

Chapter 33

15

(Dec 5 – 9)

Lecture 25 (notes): Review for final exam

Project presentation – December 7

 

16

(Dec 12 – 16)

Final Exam – December 12, Monday 4:00-6:00pm

Chapter 30-33

 

Assignments

Course project: Choose a system for performance study. You need to describe the system, specify the services, performance metrics, system parameters, workload parameters, factors and their ranges, evaluations technique and workload. Ideally, you should develop a model for the system of chosen, perform analysis, simulation and/or measurements.

l         A team of two students for each project

l         A presentation of your study is required during the class

l         A typed comprehensive report of your study is required

Home work: taken from end-of-chapter problems in the textbook (about 5-10 assignments)

Grading

Midterms 40% (20% each)

Final exam 20%

Course Projects 20%

Homework 20%

 

Note

Examines are closed-book and closed-notes. However, you can bring with you one page of formulas and definitions (US Letter size or smaller). There will be no make-up tests. If you have serious reasons for missing a test (accident, medical emergency, death in the family), upon presentation of an official written document validating the circumstance the test will be rescheduled.

 

The date and time for the final examination are specified by the University. No changes are allowed for any reason, including travel plans.