Reading List
(with references)
- Introductory materials: Three techniques for performance evaluation -- analytical
modeling, simulation and direct measurement. Performance metrics -- response time,
throughput, efficiency, utilization, reliability, availability. Workloads -- synthetic,
kernels, benchmarks. [RJ Chap 1-4]
- Review of Probability and Statistics: Probability, random variable, sample space,
conditional probability, independence, continuous and discrete random variables,
probability mass function (pmf), probability distribution function (pdf), cumulative
distribution function (cdf), expected value, variance, std. deviation, median,
quantile. Common discrete and continuous distributions, properties of exponential and normal distributions. [RJ
Chap 12, 29, plus a statistics book]
- Parameter Estimation: Sample and population, goodness of estimation, standard
error, interval estimation, confidence interval estimation for the
mean using t and normal distributions, relative confidence interval,
determination of suitable sample size. Linear
Regression [RJ
Sections 13.1-5, 13.9, 14.1-4 plus a statistics book]
- Random Number and Variate Generation: Pseudo-random numbers, uniform(0,1) generator, linear
congruential (LCG) generator, multiplicative generator, prime modulus generator, period of a
generator. Seed selection. Random variate generation
using inverse transform technique, generating discrete distributions for a given prob. mass
function. [RJ 26.1-2, 26.6-7, 28.1]
- Discrete Event Simulation: Notion of event, state and time, future event list and
operations on it. Simulation of a single server queue, network of queues, central server
systems. Use of the SMPL package [RJ Sections 25.3-5, handouts from MD
book.]. Input and output analysis for simulations. Use of replications and
batch means analysis. Fitting data to probability distributions.
- Queuing Theory: Definitions and notations, single
queue and queuing networks, queueing models of computer systems.
Operational laws for queueing systems -- Little's law, utilization law, interactive response time law. Bottleneck
analysis
[RJ Chap 30, 33, QSP Chap 1-4 (particularly 3)]. Basic notions of stochastic
processes, Markov chains - discrete and continuous time. Analysis of single
Markovian queue. M/M/1, M/M/c, M/M/c/B etc. Queuing networks - open and
closed. Product form networks. Examples from computer systems. Mean value
analysis. [RJ Chap 31,32,34, handouts from KT Book].
- Case Studies: Modeling local area networks (time permitting).