Migrating e-Science Applications to the Cloud: Methodology and Evaluation - Cloud Computing with e-Science Applications - page 108

TABLE 5.2
Documentation of an Identified Problem

ID B7

Name Connection failed

Class Tool (operability)

Severity High

Description Although correct users with the required administrative roles existed

in the MySQL database in the cloud, the application could not connect
to the database.

Error handling We were going through all the security (user and privilege) settings

in the MySQL Workbench.

Solution We set max queries, max updates, max connections to a value greater than zero

for each user.

Adaptation The user should obtain information about the limitations for the different

accounts (users).

a quality model for software by subdividing software quality in different
characteristics and subcharacteristics (Jung, Kim, and Chung, 2004). In our
evaluation, we focused on the characteristics functionality and usability of the
examined tool, in particular on the subcharacteristics suitability (for the for-
mer) and understandability and operability (for the latter), which are the possible
values for the Class attribute. The problem identified in Table 5.2, for example,
is classified under the operability subcharacteristic of usability. The attribute
Severity describes the severity of a problem with respect to the impact on the
migration result. The allowed values are low , middle , high , or critical . A detailed
description of a problem is given with the attribute Description . The attribute
Error Handling describes how the user has proceeded to find a solution for
the problem that occurred. Solution describes how the problem was fixed.
To eliminate the cause of the problem, adaptations of the tool may be needed;
these are described by the attribute Adaptation .
In the third step, the actual gathering of data was performed. Using the
Cloud Data Migration Tool, we migrated the database layer used by the
OPAL web services to the cloud. The selected use case can be mapped to
the migration scenario Cloud Bursting (Strauch et al., 2013b), with Amazon
RDS as the migration target. Throughout all phases of the migration, we
recorded any occurring problems, as shown in Table 5.2. In addition, we
measured the time spent per migration phase supported by our step-by-step
methodology (i.e., Assessment; Analysis and Design; and Migration,
Deployment, and Support), as well as the time spent on testing. No optimi-
zation activity was implemented as part of the case study. In the fourth step
of the evaluation, the previously gathered data were processed to organize
and structure for further analysis. As we had already gathered the data in a
structured and uniform manner (as described in step 2), further processing
was not necessary.