Advanced Search

Journal Navigation

Journal Home

Subscriptions

Archive

Contact Us

Table of Contents

Sign In to gain access to subscriptions and/or personal tools.
SIMULATION
This Article
Right arrow Full Text (PDF)
Right arrow References
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to Saved Citations
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Request Reprints
Right arrow Add to My Marked Citations
Citing Articles
Right arrow Citing Articles via Google Scholar
Right arrow Citing Articles via Scopus
Google Scholar
Right arrow Articles by Porras, J.
Right arrow Articles by Harju, J.
Right arrow Search for Related Content
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati   Add to Twitter  
What's this?

Other

Analyzing the Effect of Distribution to the Simulation on a Network of Workstations

Jari Porras

Lappeenranta University of Technology Lappeenranta, Finland

Jouni Ikonen

Lappeenranta University of Technology Lappeenranta, Finland

Jarmo Harju

Tampere University of Technology Tampere, Finland

In this paper, the effect of communication networks to the distributed simulation has been studied. A cellular network application is simulated on a net work of workstations, where an ordinary 10 Mbps Ethernet LAN has been used as the communication network. The effect of distribution is analyzed by using the well known concept of critical path analy sis, which requires event traces from real simula tions and information about event processing times as well as about their mutual relationships. In order to capture the effect of conservative synchronization, the event traces are collected from parallel simula tions instead of sequential. By using event traces and execution times, critical paths are calculated for different numbers of workstations. The use of differ ent configurations reveals the effect of distribution, as well as the scalability of the problem in the dis tributed environment. The critical path analysis re veals that the cellular network application is quite suitable for distributed simulation when conserva tive synchronization is used. Good approximations for the execution time can be achieved if the problem is suitably partitioned into different workstations.

Key Words: Conservative simulation • distrib uted simulation • workstation networks • critical path analysis • network delay • processing time • cellular network simulation

SIMULATION, Vol. 72, No. 5, 313-319 (1999)
DOI: 10.1177/003754979907200505


Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati   Add to Twitter Twitter    What's this?