Perf Test Methodology

Birlasoft Ltd.

Phase I – Planning As part of the planning for performance testing the following steps would be performed: • • • • • • • • • • • •

Review the technical information provided by the customer Set expectations and identify necessary prerequisites for the entire engagement Application Overview (what is the nature and purpose of the application) Architecture (hardware, software, hosting, etc…) Client/User Profile (how do users access the system, where are they located, etc…) User log-in information and test data to be used in the load testing process A description of the Business Processes the customer wishes to load test Create realistic scenarios to generate realistic load. Suggest an appropriate load testing run model Load Testing tools installation Load Testing Standards and naming conventions Prepare a Load Test Plan

Phase II Development Create Load Test Environment The development phase occurs after the planning phase is complete. The Birlasoft Performace Teting will develop all the Scripts required to execute defined Business Processes and scenarios defining the load testing model. Availability of a stable application is a pre-requisite for the development of the load testing scenario. Any changes to the application during this phase of scenario development may result in Scripts needing modification, and as such delay the execution date. Script Development The Birlasoft Performance tester will record and build the load testing Scripts according to the Business Process definition. The completed scripts are stored in the client machines and the version control system of Birlasoft (VSS). < standards and conventions> The client will provide an SPOC to the Birlasoft Performance Testing team, able to provide system and application related information. This is essential for the Script development throughout the development stage.

Smoke Test Smoke test is conducted to verify the scripts prior to the actual load test. Script verification is executed at least one business day prior to the actual load test. The objective of this run is to determine if the Scripts function properly. Script verification utilizes a limited number of virtual users against the system to be load tested. During script verification the customer must monitor the system in order to confirm that the Birlasoft Performance test Scripts are producing the desired impact on the system being load tested. The Birlasoft Performance Testing team will update the Load test Scripts and scenario if any discrepancies with the specified behavior are found during the validation run. These adjustments will be done within one business day assuming that none of the required changes are dependant on the customer, or that no additional information or data is required from the customer.

Classification: Public

Birlasoft Ltd.

Phase III – Execution This phase consists of the actual Load test execution.The Birlasoft Performance Testing team may have to control the load test remotely from various locations (as per the client’s requirement) and coordinate with the client/ application owner via telephone/ DC/ Sametime. Together, they will determine the rate at which virtual user load is applied and/or increased to the system. During the execution phase the Birlasoft Performance Testing Team will: • Conduct one load test (“run”) according to the scenarios agreed with the client prior to the run. • Gradually increase the volume of end-user traffic until one of the following events occurs: o The load test reaches the maximum time limit decided o Transaction response times increase beyond previously agreed upon levels o Transaction error rate increases beyond previously agreed upon levels Phase IV - Reporting Birlasoft Performance Testing team will supply the client with the following deliverables after Performance run completion Run summary • A description of the load model that was implemented on the run • Base performance metrics: o Maximum running Vusers o Total Throughput (bytes) o Throughput (bytes/second) o Total hits o Hits per second A set of graphs presented as static HTML pages including • Running Vusers over time • Throughput over time • Hits per second over time • Transaction per second over time • Transaction response time over time (average) • Transaction response time over time (under load) • Transaction response time (distributed) • Overlay of running vusers and transaction response time over time (average) • Overlay of running vusers and throughput <Smoke Test Results>

Classification: Public

Birlasoft Ltd. For more information please refer our Performance Test Whitepaper.

Classification: Public

Birlasoft Ltd.

Classification: Public