Continuous location validation of cloud service components

Continuously, i.e. automatically and repeatedly checking at what geographical locations cloud service components are hosted aims at validating that the cloud service satisfies regulatory and other compliance requirements. Yet continuous validation is challenging since it requires location techniques to adapt to network changes over time. In this paper, we present adaptive location classification, an approach to continuously validate the location of cloud service components. Our approach combines supervised and unsupervised learning techniques and is capable of adapting to network changes over time. We demonstrate the feasibility of our approach by presenting experimental results where we continuously validate the locations of cloud service components hosted at 14 different locations of the AWS Global Infrastructure.