Information Technology Reference
In-Depth Information
Table 5. Cluster table for C
0
Node Number
(P
k
)
Clock Frequency
(f
k
in MHz)
Specialization
(S
n
)
Time to finish
(T
prkn
in μS)
Modules allocated
(M
ij
)
Node Failure
rate (λ
lt
)
P
0
10
J
0
10
00
0.001
P
1
20
J
0
12
10
0.002
P
2
10
J
0
10
31
0.003
P
3
10
J
0
13
33
0.008
P
4
20
J
0
12
43
0.007
Table 6. Matrix D
kl
for Cluster C
0
P
0
P
1
P
2
P
3
P
4
P
0
0
1
2
3
2
P
1
1
0
3
2
3
P
2
2
3
0
1
2
P
3
3
2
1
0
1
P4
2
3
2
1
0
Table 7. E
i0k0
on cluster C
0
for Job J
0
m
00
m
10
m
20
m
30
m
40
P
0
15
20
17.5
10
20
P
1
7.5
10
8.75
5
10
P
2
15
20
17.5
10
20
P
3
15
20
17.5
10
20
P
4
7.5
10
8.75
5
10
Turnaround Time
C
0
for the given job. Final allocation of the job
to the cluster selected considering allocation of
individual modules by TSM is shown in Table 8.
Here, BEC
ik
should be read as the Best Execu-
tion Cost for m
i
module offered by node P
k
and
interpreted as the best turnaround time offered by
a node to a module being considered for alloca-
tion. This becomes the best turnaround time offered
by any node and results in allocation of the mod-
ule to this node, which can then execute it in the
minimum possible time.
The turnaround time for the above allocation
is found to be
= max (NEC
ikn
) for 'k' on which allocation has
been made
= max (38, 29.75, 24, 13, 29.5) = 38
Therefore, the final allocation of the modules
for cluster C
0
is represented in Table 9.
As can be seen from Table 5, the failure rates
of the nodes are determined and the sick (S) and
healthy nodes (H) are marked accordingly by
comparing it with the threshold failure rate λ
th
Search WWH ::
Custom Search