ECS 256A: Graduate Course
ECS 256A ANALYTIC METHODS FOR COMPUTER SYSTEMS DESIGN (3) I
Lecture: 3 hours
Prerequisite: Course 20, 154A- 154B or EEC 170, and Statistics 131A or equivalent; course 150 or 151A - 151B recommended
Grading: Letter; based on exams, homework, reports on journal papers.
Catalog Description:
Use of simulation and queueing theory in computer design. Applications
to memory hierarchies; file storage; computer networks; fault-tolerance;
scheduling.
Goals:
This course provides an integrated introduction to the analytical and
simulation-based approaches to performance modeling of computer systems.
I. Introduction to Queueing Theory
-
A. Service and arrival distributions, scheduling policies
B. Little's Rule
C. Continuous-time Markov Chains
II. Development of Simulation Software
-
A. Random number generators
B. Event-oriented approach to simulation; associated data structures
C. Development of the SMPL language through Pascal procedures
III. Introduction to Queueing Networks
-
A. Open networks
B. Jackson's Theorem
C. Closed networks
IV. Statistical Aspects
-
A. Model selection and validation
B. Regenerative approach to analysis of simulation results
V. Applications to the Design of Computer Systems (examples given throughout course)
-
A. Processor scheduling
B. Primary and secondary storage management
C. Measurement of system performance and capacity
Textbook:
K.S. Trinedi, Prpbability and Statistics with Reliability, Queuing
and Computer Science Application Prentice-Hall, 1982.
Instructor: N. Matloff
THIS COURSE DOES NOT DUPLICATE ANY EXISTING COURSE
1/96