My main area has been that of Distributed Real-Time and Embedded Systems. I have studied the workload description and the allocation and management of resources for such systems, both on the hosts and on the underlying communication infrastructure. This encompasses the study of scheduling algorithms, schedulability analysis, resource access protocols, establishment protocols, the appropriate operating system support and middleware, and - very importantly - the overall design principles necessary to effectively realize and deploy such systems. Recently I have been focussing on the representation and integration of external aspects, such as effects of power and temperature, on system performance, or environmental influences on resource availability, all in the context of distributed real-time and embedded systems. I am investigating representations and design principles that allow to drive the separation of workload from resources typically used in the description of embedded and real-time systems one step further to the separation of workload, resources, and environmental settings; and possibly other concerns as well.

I have also become interested in the area of Privacy-Preserving and generally Confidential Communication in traditional internetworks and in pervasive communication settings. Given my background in real-time systems, I am currently primarily interested in the timing behavior of workload and network traffic and in the analysis of the effectiveness of countermeasures against timing-based attacks. Our contributions have been to demonstrate

• that effective flow-confidentiality measures for realistic traffic are much harder to implement than generally assumed,
• that most currently used mechanisms in anonymous-communication systems are actually detrimental to anonymity and confidentiality for most applications, and
• that privacy-preserving mechanisms on top of contention-based protocols such as 802.11 are inherently ineffective.

The work in both areas has been characterized by a successful combination of empirical, experimental work on one hand, and formulation of metrics and models on the other. As a result, we have been able to develop calculi that very accurately represent and predict the behavior of real systems, both in terms of security and real-time performance.

My experience includes designing scheduling algorithms, communication protocols, and large software systems.