Course Details
Subject {L-T-P / C} : CS6122 : Performance Evaluation of Computer Systems { 3-0-0 / 3}
Subject Nature : Theory
Coordinator : Dr. Sanjeev Patel
Syllabus
Introduction to Probability Models and Simulation: Probability space, Random variables, Discrete and Continuous distribution: uniform, geometric, exponential, normal distribution etc, System Modeling, Measurement techniques, Experimental design, workload design, Simulations, Data Analysis and Visualization Basics of Modeling: Performance metrics: Bandwidth utilization, throughput, delays, error rate, network reliability etc. Poisson process, and Markov chain theory. Queuing Theory: Arrival and service processes, Server disciplines, Queuing networks: Open vs closed networks, M/M/1, M/M/1/K, M/M/m, M/M/m/m. Simulation and Analysis of Computing Systems: time averages versus ensemble averages, Asymptotic bounds and limit theorems, confidence intervals, generating random variables for simulation, Inspection Paradox, Empirical Workload Measurements: Heavy-tailed property, Pareto distributions.
Course Objectives
- Define performance goals for models, methods and algorithms in computational systems. Identify system-specific metrics to evaluate the processor, database, network and server workload performance.
- Employ probability theory and computational statistics to analyze system performance. Apply random variables to model the outcome of random experiments using discrete and continuous probability distributions. Analysis of the system using experimental design.
- Apply the Markovian model to analyze continuous & discrete-time queuing processes.
- Design parameters analytically and experimentally to evaluate system performance. Apply simulation methodology to explore dynamism in system behaviour.
Course Outcomes
After successful completion of the course, students would be able to: <br />CO1: Inspect and examine the outcome of experiments using various approaches or techniques <br />CO2: Select and interpret appropriate evaluation techniques, performance metrics and workloads for a system <br />CO3: Apply and build a Markovian model to develop continuous & discrete-time queuing process by discussing various queuing models <br />CO4: Classify and examine various probability distribution models for a given application and compare the performance of various techniques or algorithms.
Essential Reading
- Sheldon M. Ross, Introduction to Probability Models, Academic Press , 7th Edition
- R. Jain, The Art of Computer Systems Performance Analysis: Techniques for Experimental Design, Measurement, Simulation and Modeling, John Wiley & Sons,
Supplementary Reading
- Kishor S. Trivedi, Probability and Statistics with Reliability, Queueing, and Computer Science Applications,, Wiley
- Sanjay K. Bose, An Introduction to Queuing System, Springer
Journal and Conferences
- IEEE Journal
- IEEE Conferences