Editing Estimating Performance Test Effort

From PeformIQ Upgrade
Revision as of 13:27, 25 January 2008 by PeterHarding (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Designing Performance Tests

Here is a starting point for estimating the amount of effort required in implementing an interactive script based performance test which a tool such as Mercury's LoadRunner.

A Simple Estimation Model

The following breakdown of activities provides a starting point to estimate the amount of effort required to implement a load test, consisting of a number of tasks in a single scenario, from scratch:

0)  Creation of estimate - 1 day

1)  Setup of test environment - 5 days
	Involves:
		Familiarisation with the environment/application
		Identification of business/application tasks to be scripted
		Setting up access to the environment (user IDs and profiles)
		Designing the workload scenario (determining mix and frequency
			of tasks to be run)

2)  Per business/application task - 2 days
	Involves:
		Specifying the steps in the task
		Identifying and creating any specific data required
		Creating basic script (minimal error handling)
		Testing of script

3)  Setup of data (depends on complexity of application) (but say 1 day per task)
	Will involve mapping out the data flows and dependencies for all the tasks to
	be included in  the scenario workload.

4)  Test execution and preliminary analysis (2 days per pass - allow for three (3) passes) - 6 days

5)  Analysis and write up of final report - 3 days

Another likely activity for which an allowance should be made is the identification and rectification of faults in the system. While these may not increase the effort involved in implementing and executing the tests significantly, they often result in an increase in elapsed time for the overall testing process. If measurement and analysis of faults, severity, frequency of occurence and effort of rectification is being done as part of a development project's management then some risk analysis based on these metrics should provide a basis for estimating this.

As an example based on the above, implementation of a scripted performance test requiring 10 scripts (tasks) allow 1 + 5 + (10 * 2) + (10 * 1) + 6 + 3 = 45 days

See also Performance

Categories