You are here:

1. Dept. of Computer Science
2. CS 7
3. People
4. Falko Dressler
5. Autonomic Networking Group

Autonomic Networking Group

Coordination: Dr. Falko Dressler

Self-organization in massively distributed systems...

... and adaptation to changing environments are key elements of future networking infrastructures. Network architectures need to be able to identify available services and capabilities using cognitive approaches. In order to maintain scalability, coordination among participating entities can no longer depend on globally synchronized state information. New techniques are needed to enable efficient management and control functions in complex networked systems. Self-organization is regarded the key paradigm for this new era of control processes. Among others, bio-inspired networking solutions help to address issues of self-organization.

Application Domains

• Wireless Sensor and Actor Networks (SANETs) - SANETs represent a class of networked embedded systems supported by integrated sensing and actuation capabilities. We are working on algorithms and methods for coordination in massively distributed heterogeneous systems - relying on the basis methods of self-organization. Besides other aspects, we are focusing our research on programming techniques for network-centric communication and data management. In particular, our protocol VCP (virtual cord protocol) needs to be named that allows organizing sensor nodes on a virtual cord for optimized data management and routing. Programming is supported for example by RSN (rule-based sensor network), which is a rule language with small footprint and inherent support for heterogeneous node hardware.
• Wireless Networks and Mobile Internet - Besides other issues, the development and analysis of efficient MAC protocols and routing techniques are key instruments to develop mobile applications. We contribute in this domain by developing specialized MAC protocols for industrial application fields and by investigating routing strategies for mobile ad hoc networks (MANETs). We closely collaborate with the industry to adapt existing MAC protocols for low-latency transmissions in industrial automation fields. With our strong background in performance evaluation, we are also contributing to the MANET domain, e.g. in form of reference implementations and simulative analysis of protocols such as DYMO (dynamic MANET on demand).
• Inter-Vehicular Communication (IVC) - IVC has become a major research field in the wireless networking community, driven by the increasing demand to exchange information between vehicles for safety and comfort services. We investigate communication techniques for building vehicular ad hoc networks (VANETs) for decentralized exchange of road traffic information, e.g. to inform other cars about accidents and congestion. Using our simulation environment, Veins (vehicles in network simulation), which bi-directionally couples models for road traffic and network simualtion, we are able to precisely investigate the effects of efficient road traffic planning and effects of communication protocols including the caused CO2 emission. Furthermore, security and privacy issues are addressed, especially focusing on distributed traffic information systems (TIS).
• Bio-inspired Networking - The use of cross-disciplinary approaches for developing efficient communication methods opens new dimensions for methodological approaches in the field of massively distributed communication systems. We are investigating techniques known from molecular biology to find solutions for programming sensor networks, to achieve improved security, and to build adaptive algorithms based on promoter/inhibitor systems.
• Network Monitoring and Attack Detection - Besides the focus on wireless networks, we are investigating issues in the field of network security focusing on high-speed monitoring, traffic identification, and anomaly detection. In this field, we developed the monitoring probe Vermont (versatile monitoring toolkit) for netflow monitoring and packet sampling. Using this tool, we integrate the analysis of network flows for traffic classification (normal, anormal, peer-to-peer, etc.). We are also actively contributing to the IETF standardization in the field of IPFIX (IP flow information export) and PSAMP (packet sampling).

Currently, we are working on the following externally funded research projects.

People