Parameters for system effectiveness evaluation of distributed systems

Author

Kumar, Anup ; Agrawal, Dharma P.

Author_Institution

Dept. of Eng. Math. & Comput. Sci., Louisville Univ., KY, USA

Volume

45

Issue

6

fYear

1996

fDate

6/1/1996 12:00:00 AM

Firstpage

746

Lastpage

752

Abstract

In a Distributed Computing System (DCS), the failure of one or more system components causes the degradation in its effectiveness to complete a given task as opposed to a complete network breakdown. This paper addresses the issue of degraded system effectiveness evaluation by introducing two static measures, namely, Distributed Program Performance Index (DPPI) and Distributed System Performance Index (DSPI). These metrics can be used to compare networks with different features for application execution. It can be used to determine if the network with high reliability and low capacity, or low reliability and high capacity is better for a given program execution. An algorithm is also developed for computing these indices, and it is shown to be not only efficient, but general enough to compute many other existing measures such as computer networks, distributed systems, transaction based systems, etc

Keywords

computer network reliability; directed graphs; distributed processing; fault tolerant computing; performance evaluation; reliability; software metrics; application execution; computer networks; degraded system effectiveness evaluation; directed graphs; distributed program performance index; distributed system performance index; distributed systems; metrics; static measures; system effectiveness evaluation; transaction based systems; undirected graphs; Availability; Computational modeling; Computer network reliability; Computer networks; Degradation; Distributed computing; Distributed control; Performance analysis; Predictive models; System performance;

fLanguage

English

Journal_Title

Computers, IEEE Transactions on

Publisher

ieee

ISSN

0018-9340

Type

jour

DOI

10.1109/12.506430

Filename

506430