Measuring Performance Testing Effectiveness

on

Measuring Performance Testing Effectiveness.


Performance Testing measures the effectiveness of the computer, network or any software application. Quantitatively the application is tested for the response time upon certain input and qualitatively it's checked for reliability, scalability and interoperability.

Performance Testing is conducted by using various tools like J Meter, Load Runner, Astra Load and WAPT. Performance parameters are defined for the application and results are analyzed on the basis of these parameters like "Response Time", "Throughput", "Transfer Rate" to name a few. A complete performance analysis report is compiled highlighting the performance issues and recommendations as per the analysis. Various types of testing parameters are used to determine the performance of an application.

Stress Testing - (also called Load Testing) is performed to identify the stability of an application under certain level of load. During this test we attempt to stress the application to the point of failure. This reveals all the weak points in the application.

Recoverability Testing - checks if the application is self recoverable or not if treated beyond its threshold. The behavior of application is checked beyond its point of failure and then traversing back. This gives a clear aspect if the application is self recoverable and has been catered for any contingency plans.

Endurance Testing - (also called Longevity Testing) monitors the application behavior on a constant moderate workload for a long period of time. This helps pinpoint bottlenecks and component limitations. Spike Testing - Spike Test use real-world scenarios but under extremely fast ramp up and ramp down times with peaks upto 100% - 150% in a matter of minutes rather than gradual increase.

Performance Tuning - System's ability to adapt to higher load is called scalability. In performance tuning, modifications are made to the system to make it ready for managing higher load. We follow the measure-evaluate-improve-learn cycle for performance tuning which include the following steps :
  • Assess the problem and establish numeric values that categorize acceptable behavior
  • Measure the performance of the system before modification
  • Identify the part of the system that is critical for improving the performance
  • Modify that part of the system to remove the bottleneck
  • Measure the performance of the system after modification
Author: Clara James
Source: Link

Comments

Post a Comment