Information Technology Reference
In-Depth Information
time for the cloud was 6 hours (360 minutes). Applying the overhead factor
equation, we have the following result:
360
240
= .
15
which means that the overhead factor to execute the same calculation in the
cloud, compared to the cluster, is 1.5. In other words, the time to execute
the same application with the same data in the cloud takes 50% more time
than the cluster. The weather forecast needs to be executed daily in less than
12 hours; therefore, both solutions present adequate execution time.
The third and final step is related to the economic evaluation of both
solutions. The first input for this calculation is the price of both solutions.
The acquisition cost of the cluster is $1.3 million, and it will be used dur-
ing its lifetime of 10 years. To maintain the cluster, it is necessary to con-
tract a maintenance specialist for $3,000 per month, or $36,000 per year.
Moreover, the energy consumption of this system is $1,000 per month or
$12,000 per year. With all these costs, we can use the yearly cost equation;
the results are
$,
1 300 000
10
,
+
$,
48 000
=
$,
178 000
This result means that the cost per year with the cluster is $178,000; this
value will be used in the break-even point assessment. Another component
of the break-even point is the cost efficiency factor, assuming a cost per hour
of $50.00 for the cloud machine. Using the calculated overhead factor of 1.5,
the resulting cost efficiency factor for the cloud is 75.00 ($/hour). Using both
the yearly cost and the cost efficiency factor, we can determine the break-
even point with the following calculation:
178 000
75
,
(
)
=
98 88
.
days
×
24
but this result is related to full usage for 24 hours a day. The real usage of
the cloud will be 6 hours a day, which is the time required to perform the
weather forecast for our city. Then, we can adjust the break-even point calcu-
lation for 6 hours; the new result is
178 000
75
,
(
)
=
395 55
.
days
×
6
This result is interpreted to mean that the number of days when the cloud
has a better cost-benefit ratio than the cluster is 395 days in a year. We can
conclude that the use of a cloud instead of a cluster is cheaper.
Search WWH ::
Custom Search