A Quantitative Testing Effort Estimate for Reliability Assessment of Multi Release Open Source Software Systems
Growing software demand in the present virtual world introduces new competitive dynamics for software developers. Recently, Open Source Software (OSS) systems are providing a faster way of software production. To survive in the competitive market, developed OSS system needs enhancement in previous versions. Each enhanced versions are found to be more liable to risks of failures. In the recent software development process, the primary concern of researchers is always to find new ways for assessing the reliability of developed OSS versions. To incorporate modern software development environments and technologies, new failure rate model for reliability estimation of multiple versions of OSS systems has been developed in this paper. Proposed model incorporates a new testing effort factor for integrating varying needs in each release of software development. It comprises imperfect debugging with the possibility of fault introduction. The proposed model has been validated on various releases of Firefox and Genome project failure data set. Parameter estimation for the proposed model has been done using a flower pollination algorithm. Experimental results have shown the enhanced capability of the proposed model in comparison to Goel-Okumotto model, Inflection S-shaped model and PTZ model in simulating real OSS development environment.