DocumentCode
2368388
Title
Stochastic modeling of multiprocessor reliability
Author
Papadopoulos, Constantinos V.
Author_Institution
Dept. of Comput. Sci., Piraeus Univ., Greece
fYear
1994
fDate
2-6 May 1994
Firstpage
620
Lastpage
630
Abstract
Colored Petri nets have been shown to create greatly simplified models that have many fewer places and transitions and can be applied to families of systems with minimal changes. The paper presents a modeling tool that uses a form of colored stochastic Petri nets to model the performance of fault tolerant multicomputers as resources fail. The model can be viewed as having tokens that collect to form compound tokens in a manner somewhat analogous to the formation of molecules from atoms. Therefore we call it the Polyvalent Stochastic Petri Net (PSPN) modeling system. Using this approach simple and relatively similar models can be used for systems of varying sizes and configuration. Complex actions such as backtracking and routing around failed links in a communications system can be expressed in the transitions. Failed parts are modeled by simply removing a resource token from a place
Keywords
Petri nets; fault tolerant computing; graph colouring; parallel machines; reliability; stochastic processes; Polyvalent Stochastic Petri Net; backtracking; colored stochastic Petri nets; communications system; compound tokens; failed parts; fault tolerant multicomputers; modeling tool; multiprocessor reliability; routing around; stochastic modeling; Computer errors; Computer science; Degradation; Distributed computing; Educational institutions; Fault tolerant systems; Petri nets; Routing; Stochastic processes; Stochastic systems;
fLanguage
English
Publisher
ieee
Conference_Titel
Massively Parallel Computing Systems, 1994., Proceedings of the First International Conference on
Conference_Location
Ischia
Print_ISBN
0-8186-6322-7
Type
conf
DOI
10.1109/MPCS.1994.367025
Filename
367025
Link To Document