Software Reliability Prediction Using Duane-CNN for Dynamic Webservices

Abstract

Web services currently are based on integration of many services to provide response to user requests. When single service published there will not be more problem in aspects of utilizing effective response to user requests. We have analyzed there are certain complex arisen when we access recent days web service due to the nature of micro services for single applications and we ought to find the performance of quality parameters such as composability, security, reliability accountability and interoperability. To resolve above conflicts we need to measure mean time between failure(MTBF) or it is also known as reliability growth whenever changes are made in software design to make better web service. We proposed a scheme called D-CNN(Duane – Convolutional Neural Networks) to predict reliability growth  using J.T.DUANE  model to identify multiple failures  during design and development of applications. The results of Duane model were given to CNN after proper training, pooling and other process to generate output. We have simulated above D-CNN and the results were compared with other existing models.