Abstract
AbstractThe assumptions used to develop operational analysis computer performance measures, such as number of jobs at a device or response times, are stated in terms of the data itself, rather than the underlying system which produces the data. In spite of claims of validity and as an aid in introducing queueing theory in teaching, little has been written about operational analysis in the past ten years. Accuracy of operational analysis performance measures depend on data behavior assumptions which can be validated with data based error measures. Increased soundness of the operational analysis approach may be obtained by determining the limits of assumption errors as the time period of observation increases. Part I of this paper is a review of operational analysis and addresses some of the previous concerns with its approach. Part II develops further understanding of operational analysis assumption errors by examining their limits. Limits are found for the assumption errors of job flow balance, homogeneous arrivals and homogenous services. While the job flow balance assumption error measure is shown to approach zero over time, the homogeneity assumption error measures, in general, do not.
Highlights
OF OPERATIONAL ANALYSISOperation analysis (OA) was introduced as an aid in computer system performance analysis [1,2,3,4]
This study examines the OPERATIONAL ANALYSISOperation analysis (OA) assumption error measures previously defined [15] over extended time horizons in order to see if assumption errors decrease or stabilize, making the OA performance measure relationships more applicable
We can conclude that the differences in device performance between open and closed systems are minor when observed for a sufficient length of time
Summary
Operation analysis (OA) was introduced as an aid in computer system performance analysis [1,2,3,4]. All variables appearing in the performance equations of a real system should be calculable by direct measurement These rules mean that no assumptions are made that can’t be tested for their validity over a given time period of observation. The paradox is that real world computing systems consistently violated all the model assumptions, but the models agreed closely with observed throughput and response times. This had an impact on Jeffery Buzen’s company, BGS Systems, which built computer industry performance prediction and capacity planning tools. For a system of queues, n= (n1,...,nK) are vectors of the number of jobs at each server In stochastic modeling they assign an equilibrium probability p(n) to each state. The operational characteristics of the three assumptions allow us to calculate errors in these assumptions over finite observation periods [15]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
