Association for Computing Machinery
Accountability properties, i.e., security and privacy properties for trustworthy data stewardship are becoming increasingly important for Cloud applications. Frequently, they have to be enforced on large-scale service-based applications. In this paper, the authors argue that real-world service infrastructures are best modeled in terms of three abstraction levels and that (partially invasive) adaptations involving all levels are needed to handle accountability properties. They motivate these issues for the case of secure logging, a basic accountability property of Cloud applications. They propose an initial version of a DSL for flexible and expressive control over the execution of service compositions on the three abstraction levels: service orchestrations, interceptors and service implementations.