Testing of map reduce application with black box approach using induction method
Ashish Kumar Rai, AK Malviya
Big data is now no longer mere buzz word; industries are started realizing benefits with processing large volume of data known as big data. Earlier gold diamond mines are used represent prosperity but in modern word data resembles new ore for currency. Many technologies have evolved around processing heterogenous data to fetch business value. Hadoop Ecosystem and their components are playing vital role to provide accurate business insights, but this would not be possible without having proper and effective software testing for developed application. Testing big data application is certainly different from traditional approach specially applications which involved map reduce approach to split and combine. Current industrial practice mainly focuses on unit testing and component level tools-based automation for integration and regression. Functional testing with large volume data is still considered as a major challenge and business team also face similar issue for user acceptance testing. In this thesis, we researched on functional testing technique which can utilized large data based on mathematical induction principle to validate functionality of developed application. The proposed testing technique is equally effective on heterogenous data structured as well as unstructured data. The true potential of map reduce or big data application would be only realized when testing activity can be accomplished with accuracy and acceptance is provided accordance to business.