An efficient disk-based tool for solving large Markov models

Daniel D. Deavours; William H. Sanders

doi:10.1016/s0166-5316(98)00010-8

An efficient disk-based tool for solving large Markov models

Research output: Contribution to journal › Article › peer-review

Abstract

Very large Markov models often result when modeling realistic computer systems and networks. We describe an efficient tool for solving general, large Markov models on a typical engineering workstation. It uses a disk to hold the state-transition-rate matrix (possibly compressed), a variant of block Gauss-Seidel as the iterative solution method, and an innovative implementation that involves two parallel processes communicating by shared memory. We demonstrate its use on two large, realistic performance models.

Original language	English (US)
Pages (from-to)	67-84
Number of pages	18
Journal	Performance Evaluation
Volume	33
Issue number	1
DOIs	https://doi.org/10.1016/s0166-5316(98)00010-8
State	Published - Jun 1998

Keywords

Block Gauss-Seidel
Markov models
Stochastic Petri nets

ASJC Scopus subject areas

Software
Modeling and Simulation
Hardware and Architecture
Computer Networks and Communications

Access to Document

10.1016/s0166-5316(98)00010-8

Fingerprint Dive into the research topics of 'An efficient disk-based tool for solving large Markov models'. Together they form a unique fingerprint.

Cite this

@article{30debfe2334c401e8931e25f567a53e8,

title = "An efficient disk-based tool for solving large Markov models",

abstract = "Very large Markov models often result when modeling realistic computer systems and networks. We describe an efficient tool for solving general, large Markov models on a typical engineering workstation. It uses a disk to hold the state-transition-rate matrix (possibly compressed), a variant of block Gauss-Seidel as the iterative solution method, and an innovative implementation that involves two parallel processes communicating by shared memory. We demonstrate its use on two large, realistic performance models.",

keywords = "Block Gauss-Seidel, Markov models, Stochastic Petri nets",

author = "Deavours, {Daniel D.} and Sanders, {William H.}",

note = "Funding Information: This work was supportedi,n part, by NASA Grant NAG 1-1782,a nd by DARPA/ITO under Contract No. DABT63-96-C-0069. Any opinions,f indings and conclusionso r recommendationesx pressedin this materiala re thoseo f the authorsa nddo not necessarilyr eflectt he views of DARPA/ITO.",

year = "1998",

month = jun,

doi = "10.1016/s0166-5316(98)00010-8",

language = "English (US)",

volume = "33",

pages = "67--84",

journal = "Performance Evaluation",

issn = "0166-5316",

publisher = "Elsevier",

number = "1",

}

TY - JOUR

T1 - An efficient disk-based tool for solving large Markov models

AU - Deavours, Daniel D.

AU - Sanders, William H.

N1 - Funding Information: This work was supportedi,n part, by NASA Grant NAG 1-1782,a nd by DARPA/ITO under Contract No. DABT63-96-C-0069. Any opinions,f indings and conclusionso r recommendationesx pressedin this materiala re thoseo f the authorsa nddo not necessarilyr eflectt he views of DARPA/ITO.

PY - 1998/6

Y1 - 1998/6

N2 - Very large Markov models often result when modeling realistic computer systems and networks. We describe an efficient tool for solving general, large Markov models on a typical engineering workstation. It uses a disk to hold the state-transition-rate matrix (possibly compressed), a variant of block Gauss-Seidel as the iterative solution method, and an innovative implementation that involves two parallel processes communicating by shared memory. We demonstrate its use on two large, realistic performance models.

AB - Very large Markov models often result when modeling realistic computer systems and networks. We describe an efficient tool for solving general, large Markov models on a typical engineering workstation. It uses a disk to hold the state-transition-rate matrix (possibly compressed), a variant of block Gauss-Seidel as the iterative solution method, and an innovative implementation that involves two parallel processes communicating by shared memory. We demonstrate its use on two large, realistic performance models.

KW - Block Gauss-Seidel

KW - Markov models

KW - Stochastic Petri nets

UR - http://www.scopus.com/inward/record.url?scp=0345914604&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0345914604&partnerID=8YFLogxK

U2 - 10.1016/s0166-5316(98)00010-8

DO - 10.1016/s0166-5316(98)00010-8

M3 - Article

AN - SCOPUS:0345914604

VL - 33

SP - 67

EP - 84

JO - Performance Evaluation

JF - Performance Evaluation

SN - 0166-5316

IS - 1

ER -

An efficient disk-based tool for solving large Markov models

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Cite this