S. Performance Evaluation of Operating Systems Using Approximate Analytical Methods.pdfVIP

Performance Evaluation of Op erating Systems Using Approximate Analytical Metho ds Gunter Bolch and Stefan Greiner Institute for Mathematical Machines and Data Pro cessing IV University Erlangen Nurn b erg email b olchinformatikunierlangende Novemb er Abstract The motivation for the following pap er were three mo dels for a variant of a UNIX based mul tipro cessor op erating system that was develop ed at the Institute for mathematical machines and datapro cessing IV Jung These mo dels were formerly analyzed by Markovian metho ds to get p erformance measures and the measurements of the real system were also available This mo dels had the following problems  Class switching is allowed There are more job classes in the system available which can switch their class The problem then is that the numb er of jobs in each class is not longer constant but dep ends on the time when you examine the system  Jobs have priorities with a mixed priority strategy That means that some job classes can b e preempted but others not To reduce the time to analyze the system a new approximate technique is develop ed The tech nique is based on the well known Mean Value Analysis MVA This new technique can b e applied to op en and closed queueing networks Description of the System In this section three mo dels of a UNIXbased multipro cessor op erating systems Jung are presented For these mo dels we develop ed a new technique to compute the p erformance measures as these mo dels can not b e anal