Performance Modelling of Computer Systems
News
- [26.06.12] There will be no tutorial this week.
- [27.05.12] There will be neither a lecture, nor a tutorial next week.
- [11.05.12] As 17.05.12 is a holiday, the next tutorial will be on 24.05.12.
- [09.05.12] The tutorials start on 10.05.12.
- [17.04.12] Exceptionally, the first lecture will take place in the room C 003, at 14:15 on the 3th of May.
Inhalt
Performance evaluation allows one to quantify the use of resources in a system, where the notion of system is broad (e.g., a supermarket, a web server, a communication network, an ecosystem). This course offers the basics of this discipline with emphasis on the techniques devoted to the analysis of computer systems. Special attention is given to analytical models, where the system under scrutiny can be conveniently represented as a mathematical problem. Typical indices of performance that can be obtained from the analysis are throughput (how frequently an event happens), utilisation (the percentage of time that a resource is employed), and response time (how long it takes to perform one or a series of events). A central role will be played by stochastic models, which enable the expression of uncertainties using a probabilistic framework.
The course is self-contained and no prior knowledge is required. It is divided into two main parts. In the first part it covers elementary notions of probability theory, culminating with the study of Markov chains. The second part covers some techniques based on Markov chains to conveniently express and analyse the performance of computer systems, specifically queueing networks and stochastic process algebra. The tutorials will consists of exercises accompanying the lectures, as well as practical applications with software tools such as GNU Octave, R, and other research tools.
Termine
Vorlesung (2-stündig)
- Dienstag 10-12 Uhr, Oettingenstr. 67, 151
Übung (2-stündig)
- Donnerstag 14-16 Uhr, Oettingenstr. 67, C 003
Vorlesungsfolien
Übungsfolien
Personen
Dozent
Übungsleiter
