An Evaluation Model for Distributed Computer Systems
During the design phase of a distributed computer system, the designer needs to have access to a set of attributes, the amount of which can be used as common units of measure for both the characteristics of the distributed system and the requirements. It is important that these attributes are chosen independent of any particular candidate system.
The designer needs to make quantitative decisions on trade-offs on amounts of attributes how they affect the final measure of any candidate system relative to the requirements. It is desirable that this decision be based on an evaluation process for selecting an individual distributed system from a set of alternatives on the basis of design decision resulting from different combination of attributes. Measures developed by earlier authors for this type of trade-off analysis are a set of weighted differences of unnormalized surpluses deficits of various attributes selected for the design process. In this paper we have developed normalized measures for the surpluses deficits of the attributes so that the difference of the two is meaningful.
KeywordsWeighted Difference Distribute Computer System Candidate System Common Unit Quantitative Decision
Unable to display preview. Download preview PDF.
- Jensen, E. D., 1987, “The Honeywell Experimental Distributed Processor —.An Overview”, Computer, January 1987, pp. 28–37.Google Scholar
- Gonzalez, M. J., Jr., and Jordan, B. W., Jr., 1980, “A Framework for the Quantitative Evaluation of Distributed Computer Systems”, IEEE Transactions on Computers, C-29, (12), December 1980.Google Scholar