Analyzing the effect of distribution to the simulation on a network of workstations

被引:0
作者
Porras, J [1 ]
Ikonen, J
Harju, J
机构
[1] Lappeenranta Univ Technol, Lappeenranta, Finland
[2] Tampere Univ Technol, FIN-33101 Tampere, Finland
关键词
conservative simulation; distributed simulation; workstation networks; critical path analysis; network delay; processing time; cellular network simulation;
D O I
10.1177/003754979907200505
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
In this paper, the effect of communication networks to the distributed simulation has been studied. A cellular network application is simulated on a network of workstations, where an ordinary 10 Mbps. Ethernet LAN has been used as the communication network. The effect of distribution is analysed by using the well known concept of critical path analysis, which requires event traces from real simulations 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 simulations instead of sequential. By using event traces and execution times, critical paths are calculated for different numbers of workstations. The use of different configurations reveals the effect of distribution, as well as the scalability of the problem in the distributed environment. The critical path analysis reveals that the cellular network application is quite suitable for distributed simulation when conservative synchronization is used. Good approximations for the execution time can be achieved if the problem is suitably partitioned into different workstations.
引用
收藏
页码:313 / 319
页数:7
相关论文
共 13 条
  • [1] Berry O., 1985, P 1985 SCS MULT DIST, P57
  • [2] Stardust: An environment for parallel programming on networks of heterogeneous workstations
    Cabillic, G
    Puaut, I
    [J]. JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1997, 40 (01) : 65 - 80
  • [3] CHANDY K, 1979, IEEE T SOFTWARE ENG, P440
  • [4] Predicting speedup for distributed computing on a token ring network
    Feldman, PM
    Feldman, RE
    Kim, DB
    [J]. JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1997, 45 (01) : 53 - 62
  • [5] Managing multiple communication methods in high-performance networked computing systems
    Foster, I
    Geisler, J
    Kesselman, C
    Tuecke, S
    [J]. JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1997, 40 (01) : 35 - 48
  • [6] GUNTER M, 1994, P PADS 94, P81
  • [7] HARA V, 1996, P 2 WORKSH PERS WIR, P127
  • [8] JEFFERSON D, 1991, PROCEEDINGS OF THE 24TH ANNUAL SIMULATION SYMPOSIUM, P159, DOI 10.1109/SIMSYM.1991.151500
  • [9] JEFFERSON DR, 1985, ACM T PROGR LANG SYS, V7, P404, DOI 10.1145/3916.3988
  • [10] Lin Y.-B., 1992, ACM Transactions on Modeling and Computer Simulation, V2, P239, DOI 10.1145/146382.146401