Robust Monitoring of In-Browser Third-Party Content:An Approach to Develop a Self-Adaptive Monitoring Tool

Abstract

Prior research about software robustness has been done in the fields of distributed systems, automation and mechatronics where systems must be able to operate in erroneous environments. However, less research about the subject has been applied in the area of software engineering. This paper aims to develop a robust and modifiable prototype for monitoring third-party content. The study is divided into two iterations to identify and resolve problems when developing an online monitoring tool. In the first iteration, interviews are held and an initial prototype is developed. The second iteration introduces the usage of the self-adaptive framework MAPE-K together with a defined fault taxonomy to support robustness. Also, the modifiability of the architectures in both iterations is evaluated. From running a test-case on six third-party domains, the result shows that the final developed prototype successfully manages to heal three domains and a total of five domains if minor errors are included. The architectural evaluation shows that the final iteration had the highest modifiability. To conclude, the findings indicate that self-adaptive behaviors from the MAPE-K framework can be applied for online monitoring of third-party content to increase software robustness and modifiability.