Umeå universitets logga

Zero trust is a paradigm that has emerged from the need for better security in cloud native environments. The paradigm moves away from traditional parameter-based security to a more sophisticated method in which segmentation, user identities, observability, and request studies are utilized. 

 This thesis studies two enterprise solutions offered by AWS and Microsoft Azure which are compared by documentation and in practice. Moreover, the practical comparison focuses on the performance aspects of the solutions and how they change as zero trust principles are applied. In evaluating the solutions, four systems have been implemented. These consist of two solutions with zero trust and two without zero trust principles, where they are compared to show the latency overhead. The purpose of this thesis is to provide guidance as to when one of the two solutions can be preferred in terms of performance and overall zero trust coverage. 

 Comparison by documentation and performance showed a static increase in latency for requests made with zero trust principles applied. However, the increase in latency is hard to define and compare, as the providers do not offer the ability to specify how the services are to be hosted below region and availability zone level. Thus, making the performance obsolete in the choice of the service provider. Instead, the work showed the importance domain knowledge of a service provider as a major factor in the decision.