Browse Prior Art Database

Analytic Modeling of Computer Systems

IP.com Disclosure Number: IPCOM000131247D
Original Publication Date: 1978-Oct-01
Included in the Prior Art Database: 2005-Nov-10
Document File: 21 page(s) / 72K

Publishing Venue

Software Patent Institute

Related People

Kishor S. Trivedi: AUTHOR [+3]

Abstract

Duke University Deterministic and probabilistic models capable of representing more and more system parameters are being developed. One of their primary attractions is low cost.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 5% of the total text.

Page 1 of 21

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1978 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Contact the IEEE Computer Society http://www.computer.org/ (714-821-8380) for copies of the complete work that was the source of this textual material and for all use beyond that as a record from the SPI Database.

Analytic Modeling of Computer Systems

Kishor S. Trivedi

Duke University
Deterministic and probabilistic models capable of representing more and more system parameters are being developed. One of their primary attractions is low cost.

There are two major approaches to evaluating the performance of a computer system: simulation and analytic modeling. Simulation models are more prevalent in practice since they can represent aspects of the modeled system more faithfully than analytic models. However, simulation models are very expensive to use, and the results of a simulation are harder to interpret. Recent advances in modeling techniques are making analytic models increasingly capable of representing more and more aspects of the modeled system. This has prompted an increased interest in the use of analytic modeling in the design, selection, and application of computer systems.

In this brief survey of analytic models, space has not permitted the inclusion of the rich class of graph models; interested readers should consult Peterson's article.'

Three sets of parameters emerge as being of paramount importance in the evaluation of a computer systems The first set characterizes the workload. The second set specifies the system structure, including the types and capacities of the resources, the interactions between them, and the interconnections that determine the flow of jobs through the system. The third set specifies the system's scheduling algorithms.

Modeling a computer system is a two-phase effort.3 In the analysis phase we want to evaluate the performance of the system given the workload, the system structure, and a set of scheduling algorithms. If all of the system parameters are specified deterministically, then a deterministic analysis of the system may be carried out. However, if one or more system parameters are defined by a probability distribution, then a stochastic analysis is required.

The second phase of the modeling effort is the design phase. It consists of determining the optimum system structure and a set of optimal scheduling algorithms given a workload and an up per bound on the system cost. The dual design problem is to minimize the system cost while meeting a desired performance specification. Again, if all of the system parameters are defined deterministically, then a deterministic mathematical programming model can be formulated. Otherwise, a probabilistic programming model is required.

Model parameters

As stated above, three sets of parameters are important in analyzing and evaluating system p...