Software Performance Measuring Benchmarks

Abstract

Software performance can be assessed in terms of stability, usability, and reliability which in turn can be obtained by testing and put emphasis on testing strategies. Testing and controlled monitoring increase the life period of software by allowing it to be transformed to meet the customers’ desires. Software with modest design is more reliable and stable than the software with composite design. Simple design is easier to understand, which minimizes time that will be spent understanding the system, and it is easier to revise and helps to realize the software performance. This work emphasizes on performance measuring criteria and its release decisions based on quantitative approach. From business perspective a strong relationship between business requirements and software performance exists, so it is necessary to measure the performance of software before it gets released. As business requirements changes, software needs to be able to adapt to these changes without decreasing reliability. Software which is adaptable and is more stable will deliver maximum performance in its operational environment and will be easier for the development team to take proper release decisions. The purpose of this work is to envisage and propose a new paradigm of judging software reliability in its operational environment and address issues related to releasing of software. The results show that our proposed model after incorporating the concept of window gives optimal output in terms of operational reliability by eliminating maximum number of faults.