|
CS 6352
Performance of Computer Systems and
Networks
|
|
COURSE DESCRIPTION |
|
Overview of case studies. Quick review of principles of
probability theory. Queuing models and physical origin of random variables
used in queuing models. Various important cases of the M/M/m/N queuing system. Little's law. The M/G/1 queuing system. Simulation of queuing systems. Product
form solutions of open and closed queuing networks. Convolution algorithms and
Mean Value Analysis for closed queuing networks. Stochastic Petri Nets.
Discrete time queuing systems.
|
|
COURSE LEARNING OBJECTIVES |
|
To learn (1) queuing theoretic models and analysis
techniques of computer and communication network systems' performance, (2) to
apply the principles to some practical cases.
|
|
MAJOR
TOPICS |
|
Quick review of probability theory: The Pareto random
variable and its properties; Physical origin of Poisson and Exponential random
variables and their properties; Steady state M/M/1 queuing system analysis;
Performance measures and Little's result; Laplace transform and its use in
functions of random variables; Various special cases of state dependent M/M/1
queuing system; Applications; Steady state M/G/1 queuing system: Derivation of
Pollaczec-Khinchin mean value formula; Application examples; Discrete time
queuing systems; Open Markovian queuing networks; Product form solution;
Performance measures; Closed queuing networks; Product form solution;
Convolution algorithms to solve for product-form state probabilities;
Performance measures; Mean value analysis solution to closed queuing systems;
Performance measures. |
