Driven by new software development processes and testing in clouds, systemand integration testing nowadays tends to produce enormous number of alarms.Such test alarms lay an almost unbearable burden on software testing engineerswho have to manually analyze the causes of these alarms. The causes arecritical because they decide which stakeholders are responsible to fix the bugsdetected during the testing. In this paper, we present a novel approach thataims to relieve the burden by automating the procedure. Our approach, calledCause Analysis Model, exploits information retrieval techniques to efficientlyinfer test alarm causes based on test logs. We have developed a prototype andevaluated our tool on two industrial datasets with more than 14,000 testalarms. Experiments on the two datasets show that our tool achieves an accuracyof 58.3% and 65.8%, respectively, which outperforms the baseline algorithms byup to 13.3%. Our algorithm is also extremely efficient, spending about 0.1s percause analysis. Due to the attractive experimental results, our industrialpartner, a leading information and communication technology company in theworld, has deployed the tool and it achieves an average accuracy of 72% aftertwo months of running, nearly three times more accurate than a previousstrategy based on regular expressions.
展开▼