Results for Summer 2013 CAP 4800 Project - Poisson case ------------------------------------------------------- Execution output: ============================================================= == *** Model of a cluster of servers *** == ============================================================= = N = 10 nodes = Mu = 5.000000 req/sec = SAMPLE_INTERVAL = 60.000000 sec = LOW_THRESH = 24.000000 % = HIGH_THRESH = 24.500000 % = INCREMENT = 1 = DECREMENT = 1 ============================================================= = Total CPU time = 4.695000 sec = Total sim time = 2576098.959353 sec =------------------------------------------------------------ = >>> Simulation results - =------------------------------------------------------------ = Mean number active = 1.355546 nodes = Mean response time = 0.222116 sec = 50% response time = 0.200000 sec = 98% response time = 0.422877 sec = 99% response time = 0.496990 sec =------------------------------------------------------------ CSIM Simulation Report (Version 19.0 for Borland C) Sat Jul 13 00:33:40 2013 Ending simulation time: 2576098.959 Elapsed simulation time: 2576098.959 CPU time used (seconds): 4.695 FACILITY SUMMARY facility service service through- queue response compl name disc time util. put length time count -------------------------------------------------------------------------------- Server for fcfs 0.20000 0.205 1.02372 0.23189 0.22651 2637205 Server for fcfs 0.20000 0.045 0.22262 0.04495 0.20190 573480 Server for fcfs 0.20000 0.000 0.00004 0.00001 0.20000 103 Server for fcfs 0.20000 0.000 0.00003 0.00001 0.20000 78 Server for fcfs 0.20000 0.000 0.00002 0.00000 0.20000 61 Server for fcfs 0.20000 0.000 0.00002 0.00000 0.20000 49 Server for fcfs 0.20000 0.000 0.00001 0.00000 0.20000 35 Server for fcfs 0.20000 0.000 0.00001 0.00000 0.20000 25 Server for fcfs 0.20000 0.000 0.00001 0.00000 0.20000 14 Server for fcfs 0.20000 0.000 0.00000 0.00000 0.20000 6 TABLE 1: Table for number of active servers minimum 1.000000 mean 1.355546 maximum 9.000000 variance 0.233051 range 8.000000 standard deviation 0.482754 observations 42934 coefficient of var 0.356133 TABLE 2: Table for response time minimum 0.200000 mean 0.222116 maximum 1.425303 variance 0.003872 range 1.225303 standard deviation 0.062223 observations 3211056 coefficient of var 0.280139 Calculation of energy savings: ----------------------------- With 1.355 machines powered-on on average for 30 days the energy savings is as follows... Base case of 10 machines powered-on always: 10 x 200 x 30 x 24 = 1440.0 kWh Policy case of 1.355 machines powered-on on average: (1.355 x 200 x 30 x 24) + ((10 - 1.355) x 5 x 30 x 24) = 226.2 kWh So, the energy savings is: 1440.0 - 226.2 = 1213.8 kWh (or about $121 per month at $0.10 per kWh) ---