Computer System Performance
|Flow||Y - Computer Systems|
|Category||Obligatory by selection|
|Class Hours - Lab Hours||3 - 0|
|Lecturers||Andreas-Georgios Stafylopatis, Georgios Siolas (T & R Associates)|
Main approaches for modelling and analysis of computer system performance: analytical models, simulation and empirical techniques. Introduction to queueing theory, queueing networks (general product-form networks, convolution algorithms, mean value analysis), techniques based on queueing networks (operational laws, bounds, hierarchical decomposition, load-independent and load-dependent service models, non-product-form networks, exact and approximate solution algorithms, analysis of specific systems), discrete event simulation (random number generation, program construction, transient removal, statistical analysis of simulation results), measurement techniques and tools (workload, benchmarking, monitoring and capacity planning, design and analysis of experiments). Applications to the performance analysis of various computer systems with emphasison currently established paradigms (client/server, intranets και Internet, Web services etc.).