Main research findings
Surviving as a software organization requires both speed and quality: competition is fierce and releasing a product before your competitor is critical. However, favouring speed at the cost of software quality can have catastrophic consequences.
To solve this dilemma, software developers adopt practices like Continuous Integration and Continuous Delivery, which automate and streamline the way software is developed, validated and distributed. While these practices speed up software development and distribution, it can be challenging for quality assurance efforts to keep up with the new pace, especially if the system produces thousands of new build and test results every day.
To maintain software quality, it is often necessary to analyze the computer-generated logs that result from the test or build activities. To this end, this thesis introduces (i) a method for clustering log documents to quickly provide an overview of all issues and enable similar logs to be processed simultaneously, and (ii) a method for identifying which segments of a log are likely relevant for diagnosing the issue at hand. The methods are easy to adopt, requiring almost no data preparation. The thesis presents and rigorously evaluates the methods on real-world datasets from Cisco Systems Norway.
The work has been conducted at Simula Research Laboratory and the University of Oslo.
Prior to the defence,Carl Martin Rosenberg presented his trial lecture"Continuous integration, delivery and deployment: State-of-the-art, challenges and impact on the software industry".
The PhD defence and trial lecture were fully digital.
- Professor Rui Abreu, University of Porto, Portugal
- Professor Barbara Russo, Free University of Bozen-Bolzano, Italy
- Professor Magne Jørgensen, Department of Informatics, University of Oslo, Norway
- Chief research scientist Leon Moonen,Simula Research Laboratory
- Adjunct Professor Are Magnus Bruaset,Simula Research Laboratory and Department of Informatics, University of Oslo
Chair of defence
- Professor Petter Nielsen, Department of Informatics, University of Oslo