question 1. (Single center single server platform)
Given a single server platform with APWP ls discipline and two priority levels,calculate and tabulate the Tw values for increasing values of p (p = 0.4,0.6,0.8).Assume the channel capacity be C=105 op.ns/sec,the average number of op.ns job be Z=105 op.ni.job and the level 1 arrival rate be =35%.
solution
Tw = 0.4,0.6,0.8
C= 105 op.ns/sec
Z = 105 op.ni.job
rate = 35% λ
P= 0.4+0.6+0.8
L q = P 0 ( μ λ ) c ρ / c(1 − ρ) 2
where
c−1 X
(cρ) c
(cρ) m
P 0 = 1/
c(1 − ρ) /
m=0 m!
Note that P 0 denotes the probability that there are 0 customers in the system.
ρ = λ /c μ
p= 1.8 x 35/100/105
ans = 0.63×105
question 2 (Single center Multiple server platform)
Given a multiple sever platform with p=0.4,compute and tabulate the Tw(M,y,C) M/M/LIFO values for a increasing number of channels m(m=1,2,3). Assume the global channel capacity of the platform be C=105 op.ns/sec and the average number of op.ns per job be Z=105 op.ni.job.
solution
Tw(M,λ,C)
M= 1,2,3
C= 105 op.ns/sec
L q =
ρ 2 /1 − ρ
Tw(M,λ,C) = P(Z(C(1-P))
Tw(M,λ,C) =BZ(C(1-P))
Tq(I,λ,C) < Tq(m,λ,C) for m= 1,2,3
The values of increasing channels are Tq(I,λ,C) < Tq(m,λ,C)
question 3.(Multiple center Platform)
Given the Internet Platform P in figure,Consisting of N Clients(T1…..TN) that launched internet service requests to the HTTP Server that, in turn ,when needed,iteratively launches access requests to the Application Server or to the Database Server.
After a number of Appl or DB access iterations,the HTTP server returns the internet request answer to the Client.
Assuming the routing parameters have the values below.
uclient=0.067job/sec
uHTTP =200 job/sec
u Appi = 37 job/sec
UDB = 33 job/sec
number of N clients = 2
service time= exponential
service discipline = abstract
obtain the following:
1. Give the solution to P by use of the markovian method
solution
�����=+1,≥
��1, 1≤��<
���,1, ≤��<1
���,2,1≤��<2,
��,≥�
����1(!)20,0≤��������<,1!!−���0,≤����<1,1−��11−�����1+12!!−���0,1≤����<2,1−��112−���12−�����2+1!!−���0,≥�2,
�={{{��(!)20,0≤��������<,1!!−���0,≤����<1,1−��11−�����1+12!!−���0,1≤����<2,1−��112−���12−�����2+1!!−���0,≥�������2,(3)where1=/1,2=/2,and ��� /.The idle probability � can be obtained as ��0=[−1∑�����=01(!)2+1−1∑��������=1!!−��+2−1∑����=11−��11−�����1+12!!−�+∞∑����=21−��112−���12−����2+1!!
2. Obtain the following performance indices:
2.1 The marginal probabilities of the population.
solution
n�, 1≤����<,, ≥�
����(!)20,0≤��������<,!!−���0,≥�
the probability = ��0=[−1∑����=0(!)2+∞∑��������=!!−�]−1
2.2 The average local response time of the Appl for N=2
solution
n=2
u= 0.067
n = u = 0.0134 x u
2.08 mins per client
2.3 The average global responce time of platform P for N =2
solution
global responce time w ��(2−��122−������2/!1−��+∑���=1(−��1+11(−�1)!!−��2−���12−��2+1(−���1)!!−��)+1−1∑����=+1(−�1+11−��2−���12−��2+1(−���1)!!−��)+2−1∑����=1(−�2+12−��−��2+1(−���1)!!−���)2−�12) ⋅��((2−��12−����2(/−��1)!−�−�1+−1∑���=0(−���1+11(+1)!!−��2−���12−�����2+1(+1)!!−��)+1−1∑����=(−�1+11−��2−���12−�����1+1(+1)!!−��)+2−1∑����=1(−�2+12−��−�����2+1(+1)!!−���)2−��12))−
����=((2−��122−������2/!1−��+∑����=1(−��1+11(!)2−��2−���12−��2+1(−���1)!!−��)+1−1∑����=+1(−�1+11−��2−���12−��2+1(−���1)!!−��)+2−1∑����=1(−�2+12−��−��2+1(−���1)!!−���)2−���12)1−�110−�(1−���0)) ((2−��12−����2(/−��1)!−�−1
n=2
u= 0.067 job/sec
global responce time
n/u =p
g responce time = 29.850 mins