Automated bug assignment in software maintenance using graph databases

Abstract

Processes involved in maintaining a system play a crucial role in enhancing customer satisfaction and longevity of the system. Maintenance engineers are the most critical resources in Software Maintenance. They play a significant role in fixing bugs and ensuring the normal functioning of systems. Software maintenance is a tedious task for novice engineers who are new to the system domain. The lack of up-to-date documentation makes system comprehension more challenging for inexperienced engineers. Assignment of high priority bugs to novice engineers may lead to inappropriate fixes and delay in the revival of an impacted system. Such issues may degrade customer satisfaction and also poor fixes can have a severe impact on the functioning of the system at a later stage. Our research is focussed on identification of engineers with the right level of experience to fix a given bug. We have used the concept of page ranking and graph databases to compute the importance of bugs and assignees in a graph. A newly reported bug will be scored and matched with bugs that have a similar score in the graph database. Assignees who have fixed a bug that closely maps the score of the reported bug will be assigned the task of fixing the bug. We have implemented this methodology using bug reports from QT framework on neo4j graph database. Our results are promising and will definitely pave way for a new bug assignment strategy in software maintenance