In order to approach our vision of scalable, reliable, secure, self-organizing, and yet simple and manageable distributed systems, we work in a number of specific and contemporary research areas. This research addresses foundational problems of computer science over a longer period of time and often encompasses multiple specific research projects.

Failure-tolerant Distributed Systems

To compensate for the failure of a distributed service, it needs to be provided in a redundant manner (replicated). Ideally, this is transparent for the user. The distributed systems must therefore take care of error detection and correction. Furthermore, consistency of the redundant services and data must be provided. The focus of our work is the development of platforms and algorithms that can provide these capabilities in an automated manner, relieving service developers from their implementation. Mechanisms need to adapt to different load scenarios and be performant and scalable.

Security and Privacy in Cyber-Physical Systems

Cyber-Physical Systems (CPS) are defined as systems that interact with the physical world by means of sensors and actuators. Examples include process monitoring and control in energy creation and distribution and similar critical infrastructures, intelligent transportation systems, or modern smartphones with ubiquitous computing applications. Such systems exhibit special security characteristics and requirements, as errors caused by malicious attackers can have severe consequences. One also needs to ask how correctness and consistency of sensor data can be ensured, a topic that classical IT security did not address in an adequate manner. Finally, CPS often detect and process personal data, especially location data, and privacy and data protection need to be address accordingly.

Communication for Multi-Media Communication

Applications for multi-media communication require a number of standardized protocols. Today, each applications implements a protocol stack of its own. We investigate, how protocol implementations and other support functionality can be integrated into the system software of a distributed system to make applications independent of the used protocols, can share implementations and may even benefit from new protocols that have been added later on. Communication applications often have real-time requirements because data needs to be processed and forwarded within certain deadlines. A further area of work is intelligent resource allocation that can link user preferences with the demand of a previously unknown set of applications and still meet realtime requirements.

Dynamic, Wireless Multi-Hop Networks

Mobile ad-hoc networks, wireless sensor networks, mesh networks or vehicular ad-hoc networks are all examples of a recent trend towards self-organizing mobile networks. Scalability, reliability, security, and self-organization of such networks is another part of our work.