Research On The Feasibility And Evaluation Method Of Cloud-based Software Testing

Software testing is a challenge for several software development companies,particularly for companies involved in large-scale projects.In large-scale projects,the number of test cases can grow up to several thousand,which require a significant amount of time and resource to be executed,therefore,bringing a number of emerging challenges/issues for the testing department.One solution is to use parallel test execution,which still requires expensive network devices,storage,and processing servers.Either these computing resources are underutilized or out of use after testing,thus adding extra expenses to the overall budget.Moreover,some manual testing like regression testing is time-consuming and laborious.Furthermore,some tests like testing of Saa S and Cloud testing can only be done in the Cloud.To cope with challenges of software testing,employing Cloud Computing(CC)technology would be the best choice.However,conducting testing in the CC environment is neither cost-effective,nor it is the best possible solution to all testing problems.Moreover,recent studies confirm the lack of a comprehensive model for assessing the suitability of CC for software testing.This study aims to develop a technology acceptance model for decision making regarding cloud adoption for software testing.The proposed model will provide decision support to software development organization(SDO)in the form of various predictors and determinants that will guide SDO towards cloud adoption for software testing.In order to develop cloud testing adoption assessment model(CTAAM),for data collection,we performed a systematic literature review(SLR)by applying our customized search strings,which were derived from our research questions based on the multi-vocal mapping study.We performed all the SLR steps,such as protocol development,initial selection,final selection,quality assessment,data extraction,and data synthesis.In results of the SLR study,we identify 70 influential factors(IFs)from a sample of 136 papers.To validate the SLR findings and to rank the IFs a questionnaire survey was conducted in the software testing industry in which 95 experts,from 20 different countries,participated.Based on the collected data,we performed two-phase analysis based on MCDA methods(such as Fuzzy MADM and ISM)and machine learning approaches(Hybrid SEM and ANN).We develop CTAAM model based on the two-phase analysis results.Two case studies were conducted for evaluation of the CTAAM.To rank the influential factors,we used our developed fuzzy MCDM based algorithm.For this purpose,we have developed an evaluation scale with 133 items based on the previous researches.Out of 11 inhibitors,only six inhibitors were considered critical.Furthermore,all the 10 predictors were considered critical.To research into inner correlation among the influential factors,we once used ISM approach,based on the results of a major questionnaire survey;we selected a panel of ten experts based on their experience.Only 44 factors,which majority of the experts agreed on and considered distinct,were considered for further analysis through ISM techniques.We found all 44 factors correlated to each other and for further analysis,we used MICMAC analysis.The key 44 factors were distributed into four Quadrant of power cluster matrix.In the second phase of data analysis,two-stage multipleinvestigative techniques were applied by combining the SEM method and ANN exploration method.This stage is further divided into two steps.In the first step,we have tested 10 main and 12 sub proposition or hypothesis.Out of 10 main propositions,eight were accepted while two were rejected based on significant value(p-value).Out of 12,sub-proposition,only one was rejected.The results show the positive effect of five predictors and negative effect of three predictors on Cloud adoption while the effect of two predictors were not considered significant in our SEM based cloud adoption study.This study contributes positively to the existing empirical literature on the CC adoption in the context of software testing and is the foundation of Cloud adoption for future research in this field.This study also helps the organizational managers,policy,and decision makers in articulating the effective policies and strategies for the implementation of Cloud services.The infrastructure needed,the time frame required,the training services and the requirement of knowledge for the effective adoption may be investigated.The application value in industrial field and theoretical contribution of the present fuzzy section is the design and development of a general framework to improve multi-attribute assessment models.The model can be applied to determine organizational Cloud adoption in connection to software testing based on various influential factors and predictors as evaluation criteria.The industrial contribution of the model is that it can be used as an assessment tool for SDO vendors,and will indicate their weakness using a fuzzy version of the Motorola instrument developed by SEI and revised to the fuzzy environment by us.The ISM analysis is of great value to the researchers and practitioners and has significant implications for CC service providers and organizational managers for the formulation of better strategies and policies for the current CC adoption.The developed model helps service providers in understanding the association among various significant factors and their role in the CC adoption in the software development industry for software testing.In addition,the organizations would come to know about the potential benefits of Cloud and the reasons why they should go for CC.The SEM-ANN part of the study shows a causal relationship between the independent variables and the dependent variable in order to predict the willingness to adopt CC services by employing a hybrid approach of linear and nonlinear modeling.The proposed hybrid approach improves the predictive ability of the model.