Migrating e-Science Applications to the Cloud: Methodology and Evaluation - Cloud Computing with e-Science Applications - page 107

Wisdom Data

Step 1
Identify the strategy
for improvement

Step 2
Define what you will
measure

Step 7
Implement
improvement

Step 3
Gather the data

Plan

Act Do

Step 6
Present and use the
information

Check

Step 5
Analyze the
information and data

Step 4
Process the data

Knowledge Information

FIGURE 5.5
CSI seven-step process used for the evaluation. (Adapted from Case, G., and G. Spalding (2011).
ITIL Continual Service Improvement . London: TSO, The Stationery Office.)

scenario and investigate whether it supported the scenario in an effective
and efficient manner. In the second step, which data will be collected needed
to be defined. These data were the basis for the subsequent process steps.
In our evaluation, we collected both qualitative and quantitative data. With
respect to the former, we recorded the user-identified problems that occurred
during the execution of the SimTech SWfMS migration as the means to eval-
uate the software quality of the Cloud Data Migration Tool. Such problems
are gathered only in a qualitative manner (i.e., we were not interested in the
number of problems that occurred but in a comprehensive description and
classification of these problems). This approach increased the effort to gather
the data but in turn enabled a more detailed and potentially more meaning-
ful analysis. In terms of quantitative data, we recorded the time required for
executing the various migration phases. To be able to compare our proposal
with the one by Laszewski and Nauduri (2011), we chose to use their phases
as the metric of the efficiency of our proposed approach. In this manner, we
could attribute time elapsed to higher-level activities in addition to evaluat-
ing the impact of not incorporating the testing and optimization phases in
our proposal.
To enable structured gathering and recording of problems that occurred,
we defined a set of attributes related to them. Table 5.2 shows an example of
such a problem that was identified during our evaluation and the information
we collected for it. Every problem has a unique identifier ( ID ) and a descrip-
tive Name . The attribute Class is used to classify the problem in predefined
categories. We derived these categories from ISO/IEC 9126-1, which defines