Friday, December 14, 2012

Why response time of a page does not equal the sum of its requests

The response time for a page typically differs from the sum of its requests. This does not mean that your data is incorrect. The difference can be caused by concurrent requests, page connection times, inter-request delays, and custom code within a page.
The most common reason for the sum of the individual request times within a page to exceed the total page response time is that requests are often sent concurrently (in parallel) to a server. Thus some of the individual request response times overlap so the sum of the request response times would exceed the page response time.
Additionally, the page response time can exceed the sum of the individual request response times within the page for the following reasons:
  • The individual request response times do not include time to establish connections but the page response time does include the connection request time.
  • Inter-request delays are not reflected in the individual request response time but are reflected in the page response time.
  • Custom code placed within a page is executed serially (after waiting for all previous individual requests to complete) and thus contributes to the page response time. It does not affect individual request response times. However, we recommend that you place custom code outside of a page, where it will not affect page response time.

Saturday, December 8, 2012

Performance Testing Estimation Preparation

It depends of which estimation techniue you are using...if it is WBS then you will have to break all the performance testing activities in smaller parts then using your prior experience you can estimate no of days for each activities. Also take some time for each activity in spare so that you can get time in case of any environmental or deploymental delay or issue. Incase of WBS following activities can be considered: 

A. Planning 
1. Understanding of application 
2. Identifing of NFR 
3. Finilazing the workload model 
4. setup of test environment and tools& monitors 
5. Preperation of Test plan 

B. Preperation 
1. Creation & validation of Test scripts 
2. Creation of Test Data 
3. Creation of business scenarios 
4. Getting approval 

C. Execution 
1.Run a dummy test 
2. Baseline test 
3. Upgrade or tune the environment (if needed) 
4. baseline test2 
5. final performance run 
6. Analysis 
7. Final performance run2 
8. Benchmarking etc.. 

D. Reporting 
1. Creation of performance test report 
2. Review with seniors or peers 
3. Update the report 
4. Publish the final report. 
5. Getting signoff