RADAR: Data Protection in Cloud-Based Computer Systems at Run Time

Abstract

Research leading to these results received funding from the European Union's Horizon 2020 research and innovation programme under grant agreements no. 731678 (RestAssured) and 871525 (FogProtect). Useful discussions with project partners in RestAssured and FogProtect are gratefully acknowledged.

A Data Protection Focused Adaptation Engine for Distributed Video Analytics Pipelines

“…This is followed by a discussion on the limitations of our approach and its general applicability to other use cases. The runtime behavior of RADAR has been evaluated in previous work [24]. Additional information on the evaluation of the pattern-matching algorithm can be found on the Henshin website 8 .…”

“…They use bigraphs to model the topology and security requirements of the system. Their approach takes a lot of execution time, while our approach builds upon previous research [24], which we could show is well performing even with larger system models. Additionally, both of the above mentioned proposed adaptation mechanisms focus solely on the mitigation of data protection threats, but do not take other parameters into account, such as QoS, performance or energy consumption.…”

“…Our approach is not limited to that kind of data only (i.e., personal data, as explained later), but is rather capable of operating with any kind of sensitive data. Second, we give a brief overview on previous work, i.e., the main components from RADAR [24], which we extended, updated and enhanced for our approach. Third, we explain what kind of adaptations are possible within our approach.…”

“…This work is based on the so called RADAR (Run-time Adaptations for DAta pRotection) approach presented in [24]. RADAR aims at ensuring data protection in dynamically changing cloud-based systems and other related system concepts such as edge or fog computing.…”

“…The search tree is built up during the search, by iteratively adding unexplored child nodes to the nodes already visited. An evaluation of different strategies has shown that a best-first search approach leads to the best results [24]. The model comparison takes the prioritization of the comparison metrics into account.…”