Computer Science

ECS 256 Performance Evaluation


Lecture: 3 hours

Project: 1 hour

Prerequisite: Courses ECS 20; ECS 152A; ECS 154A-B or EEC 170; Statistics 131A; course ECS 150 recommended

Grading: Letter; 4-6 homework sets (20%), midterm (25%), final (35%), project (20%)

Catalog Description:
Use of simulation and queueing theory in computer and communication system design. Applications to processor scheduling, memory hierarchies; I/O systems; packet and circuit switched networks; fault-tolerance; computer networks applications. Not open for credit to students who have completed ECS 256A.

This course provides an integrated introduction to the analytical and simulation-based approaches to performance modeling of computer and communication systems.

Expanded Course Description:

  1. Introduction to Queueing Theory
    1. Service and arrival distributions, scheduling policies
    2. Little’s Law
    3. M/M/1 system
    4. Generalized state dependent arrival and service system
  2. Advanced Queueing System
    1. M/G/1 system
    2. Markov Modulated Poisson process
    3. Models for burty traffic (M/G/inf, Pareto distribution)
    4. Self-similarity
  3. Queueing Networks
    1. Open networks
    2. Jackson’s Theorem
    3. Closed networks
  4. Development of Simulation Software
    1. Random number generators
    2. Discrete event simulation methodology
    3. Model selection and validation
    4. Statistical analysis of simulation results
  5. Applications to the Design of Computer Systems
    1. Processor scheduling
    2. Primary and secondary storage management
    3. Model for multiprogrammed computer system
    4. Measurement of system performance and capacity
  6. Application to the Design of Communication Systems
    1. Delay models in packet switched networks
    2. Internet traffic characteristics
    3. Blocking models in circuit switched networks
    4. Overflow models

The project will include the design and analysis of a computer and/or communication system using the analytical and simulation methodologies developed in this course.

K.S. Trinedi, Probability and Statistics with Reliability, Queuing and Compute Science Application, Prentice Hall, 1982

Selected papers from the recent literature

Instructors: N. Matloff, D. Ghosal, P. Mohapatra, B. Mukherjee

Prepared by: D. Ghosal (September 2002)

Overlap Statement:
There is no significant overlap with other courses.