CSC470 Fall, 2000 Computer System Modelling And Analysis Laboratory #2 October 3/5 & 10/12, 2000 Verification of Authorship: By the signatures below, we certify that we are thoroughly familiar with the contents of this laboratory report: it is our own work, we have referenced all our sources of information, and we are the sole authors. Team Members: Name Signature Date (1). ________________________ _________ (2). ________________________ _________ (3). ________________________ _________ Assignment For Lab # 2 1. Number of Nodes Examine whether, and if so how, the number of nodes in the system influences the relative benefits of the different load sharing policies. Evaluate the cases of N = 2, N = 4, N = 8, for each of the four policies, and provide confidence intervals on your answers. Values for all the parameters are given in the table below: C alpha N D A M I thresh prbcnt 40,000 .02 {2, 4, 8} 10.0 20.0 3.0 1.0 3 1 ***** enter summaries of 12 runs ******* ***** enter means and CI's in table ***** N=2 N=4 N=8 ------------------------------------------------------------ | NLS | [low, mean, high] | ------------------------------------------------------------ | ONEB | | ------------------------------------------------------------ | HOTP | | ------------------------------------------------------------ | RECIN | | ------------------------------------------------------------ For each number of nodes, indicate the preference ordering among the four strategies. (Be careful to assert dominance of one strategy over another only where it can be justified by the confidence intervals.) ***** discuss results briefly noting trends ***** 2. Overall System Load Examine whether, and if so how, the overall system load influences the relative benefits of the different load shar- ing policies. Evaluate the cases of A = 30, A = 15, A = 10, for each of the four policies, and provide confidence inter- vals on your answers. Values for all the parameters are given in the table below: C alpha N D A M I thresh prbcnt 40,000 .05 6 9.0 {30.0, 15.0, 10.0} 18.0 1.0 3 3 ***** enter summaries of 12 runs ******* ***** enter means and CI's in table ***** A=30 A=15 A=10 ------------------------------------------------------------ | NLS | [low, mean, high] | ------------------------------------------------------------ | ONEB | | ------------------------------------------------------------ | HOTP | | ------------------------------------------------------------ | RECIN | | ------------------------------------------------------------ For each interarrival time and resulting overall system load (which is D / A), indicate the preference ordering among the four strategies. (Be careful to assert dominance of one strategy over another only where it can be justified by the confidence intervals.) ***** discuss results briefly noting trends ***** 3. Cost of Moving a Job Examine whether, and if so how, the cost of moving a job from one node to another influences the relative benefits of the different load sharing policies. Evaluate the cases of M = 2, M = 4, M = 8, for each of the four policies, and pro- vide confidence intervals on your answers. Values for all the parameters are given in the table below: C alpha N D A M I thresh prbcnt 40,000 .10 5 30.0 40.0 {10.0, 30.0, 60.0} 1.0 2 3 ***** enter summaries of 12 runs ******* ***** enter means and CI's in table ***** M=10 M=30 M=60 ------------------------------------------------------------ | NLS | [low, mean, high] | ------------------------------------------------------------ | ONEB | | ------------------------------------------------------------ | HOTP | | ------------------------------------------------------------ | RECIN | | ------------------------------------------------------------ For each cost of moving a job, indicate the preference ord- ering among the four strategies. (Be careful to assert dom- inance of one strategy over another only where it can be justified by the confidence intervals.) ***** discuss results briefly noting trends ***** 4. Choosing RECIN Parameters Evaluate the case indicated in the table below for each of the four policies, and using various settings for the two RECIN control parameters, thresh and prbcnt, and provide confidence intervals on your answers. Seek values for the two RECIN parameters that make RECIN perform as well as pos- sible relative to the other three policies on this case (ideally, beating all the others). C alpha N D A M I thresh prbcnt 40,000 .20 8 20.0 30.0 50.0 1.0 ??? ??? ***** enter summaries of runs ******* ***** enter means and CI's in table ***** thresh prbcnt CI ------------------------------------------------------------ | NLS | (n.a.) (n.a.) [low, mean, high] | ------------------------------------------------------------ | ONEB | (n.a.) (n.a.) | ------------------------------------------------------------ | HOTP | (n.a.) (n.a.) | ------------------------------------------------------------ | RECIN | (enter thresh, prbcnt & CI for best RECIN case) | ------------------------------------------------------------ For the best parameter settings you find for the RECIN stra- tegy, indicate the preference ordering among the four stra- tegies. (Be careful to assert dominance of one strategy over another only where it can be justified by the confi- dence intervals.) ***** discuss results briefly noting trends ***** ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Csim Program: ***** insert Csim program (for one of the 18 cases) ***** Csim Output: ***** insert full output of Csim program for one case ***** ***************************************************************************