|
Over the years performance testing has advanced to a level where the user can get reports by navigating through the various combinatorial business workflows of the web application concerned. Earlier performance engineering was done using mathematical equations and formulas but now currently there are sophisticated tools available which gives the statistics summary of the stress undergone by the web application when a number of concurrent users are using the web application undergoing the test.
The new age tools uses the concept of designing the framework of automation using containers , loops , SLA’s (Service level Agreement) and finally report generation in PDF, HTML, XML file. Containers are used to systematically record your script on the module or option used in the web application. For example in a web application when you click on login screen you can have a container defined as login so the actions recorded in login are stored in this container. SLA’s are values which we can define in the SLA’ s Option of the tool and apply it to automation framework. SLA’s are ideally given by the customer or a benchmark of values has to be defined by the testing department in conjunction with the system administrator.
Performance tests can be used in all stages of web service development and deployment process , the key to performance testing is to stimulate the production environment and workload scenarios so as to obtain the most accurate real world results possible.
Defining the Testing Objectives
The purpose of this step is to define the goals of the performance tests.
These goals have a strong influence on cost and effort involved in the testing
project. Examples of testing goals are as follows:
• Determine the Web server capacity.
• Find out the maximum number of concurrent users a Web service supports within the limits of the SLAs.
• Determine the capacity of the application layer.
• Identify the bottlenecks in the Web service infrastructure.
• Identify network impact on the end-user perceived response time. Find out the capacity of the DB server.
• Identify the most expensive Web functions.
Analyzing the test Results:
Analyzing the test results is very important in Performance Testing ideally three main criteria’s are against the number of virtual users what is the average response time of page load in seconds, % CPU utilization, average request response time of the page and server throughput. We can analyze this results against a standard benchmark set by the customer or a benchmark as defined by the testing department in conjunction with the system administrator.
Finally to conclude tools used in performance testing are available in the market in the form of open source which are free or licensed versions which costs you. Open source tools don’t come with support so the testing team has to be trained on it, so it depends on the organization process assets on which tool to evaluate and choose.
The author(Suresh Menon) is a Subject Matter expert on ERP and has worked in various capacities as principal consultant , ERP consultant, Implementation Manager , Test Manager and can be contacted at suresh@digitalstream.in for queries and comments.
