Multi-cloud application management can optimize the provisioning of cloud-based applications by exploiting whole variety of services offered by cloud providers and avoiding vendor lock-in. To enable such management, model-driven approaches promise to partially automate the provisioning process. However, such approaches tend to neglect security aspects and focus only on low-level infrastructure details or quality of service aspects. As such, our previous work proposed a security meta-model, bridging the gap between high- and low-level security requirements and capabilities, able to express security models exploited by a planning algorithm to derive an optimal application deployment plan by considering both types of security requirements. This work goes one step further by focusing on runtime adaptation of multi-cloud applications based on security aspects. It advocates using adaptation rules, expressed in the event-condition-action form, which drive application adaptation behaviour and enable assuring a more-or-less stable security level. Firing such rules relies on deploying security metrics and adaptation code in the cloud to continuously monitor rule event conditions and fire adaptation actions for applications when the need arises.
Kritikos, K., & Massonet, P. (2016). Security-based adaptation of multi-cloud applications. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9481, pp. 47–64). Springer Verlag. https://doi.org/10.1007/978-3-319-29883-2_4