background image
<< Functional and Performance Testing | Local and Agent Computers >>
<< Functional and Performance Testing | Local and Agent Computers >>

Virtual Testers in Functional Tests

Virtual Testers
15
For information about functional testing, see Part 2, Functional Testing with Rational
TestManager
.
Performance Testing
In performance testing, you can measure:
s
The client response time the total end-to-end response time as seen by a user. It is
the time it takes for a user to enter a request, the server to respond to the request,
and the user to see the results.
s
The server response time the time it takes for the server to process a request.
Performance testing can be more complex than functional testing because
performance itself is subjective. What one user might perceive as too slow, another
user might perceive as acceptable. Therefore, when planning performance tests, you
should put some thought into what constitutes acceptable performance.
Another complexity of performance testing is that performance varies widely
depending on workload conditions. Querying a database on a system that is primarily
used for CPU-intensive activities yields a different response time than performing the
same query on a system used primarily for generating I/O-intensive database reports.
For information about performance testing, see Part 3, Performance Testing with
Rational TestManager
.
Virtual Testers
When you play back a test script, you emulate a discrete activity performed by an
actual user (tester). A virtual tester is a single instance of a test script running on a
computer.
Virtual Testers in Functional Tests
A virtual tester running a functional test emulates the user's GUI activities, such as
menu choices and control selections. In other words, these virtual testers represent the
application level view of the client/server conversation. Because the virtual tester
emulates a user interacting with the GUI, you run only one virtual tester at a time on a
computer.
When you record a GUI script in Robot, Robot records the user's activities -- such as
keystrokes and mouse clicks. This traffic is the only activity that Robot records. After
recording, Robot generates an appropriate test script. When you run the script in
TestManager, you are running a virtual tester. During the test run, you see the same
GUI actions and displays that you saw when you recorded the script.