Internal Projects

Internal Funding

Trust and Reliability
Head of
Project
Thomas Engel
Scientific
collaborators:
Eugen Staab, Dagmara Koenig, Volker Fusenig, Thomas Engel
Domain We are working on Trust and Sabotage-Tolerance techniques, with the aim to ensure that information,
which is acquired from possibly unreliable informa- tion sources, is correct. We are applying the
techniques in the area of Grid Computing and Distributed Intrusion Detection. We are also working on a
flexible anonymity technology for mobile ad-hoc networks with the aim to adopt the anonymity preserving
technology to the application area of the mobile ad- hoc networks. We use a trust system in order to
support the selection of the anonymity level for communication

Running Time Until 2009
Partners: Eugen Staab collaborated with the following Researchers:
Guillaume Muller (Escola Politécnica de Universidade de São Paulo, Brazil): Worked on a Meta-Model for
experience-based Trust and Reputation reasoning.
Martin Rehak (Department of Cybernetics, Czech Technical University, Prague):Worked on Automatic
Adaptation of Intrusion Detection Systems [15].
Dagmara Koenig collaborated with the following Researchers:
Joerg Abendroth (Nokia Siemens Networks): Work on the paper: K-anonymous communication in mobile
ad-hoc networks
Results We focus on trust that is based on direct experiences, and not “hard-security” trust which uses
cryptographic mechanisms. Hard-security trust helps to ensure data authenticity and data integrity (which
both is implicitly assumed in our work). However, they cannot help to ensure the correctness of the
transfered data in terms of trueness. Furthermore, we focus on information that is ob jectively either
correct or not, i.e., the trueness of information is binary and there is no sub jectivity involved. This type of

information appears for instance in Grid Computing, where computations are outsourced and the
verification of the received results would be too costly.

Mesh Sequencer

Service Quality Enhancement and Cooperatively Enforced Reliability in Mesh Networks

Recently the usage of Mesh-Networks became popular for difficult or inaccessible terrain, for bigger buildings or to build up private"Freifunk"-Networks (free radio nets) in bigger cities.
Mesh-Networks, or so called Multi-Hop-Networks, are wireless networks which differ from WLANs in the way, that the access to the wired network (LAN) has not necessary to be done with the next access-point
(single-hop), but it is possible that one or more radio stations are in-between (multi-hop). Access-points in Mesh-Networks also have a routing-function to direct neighbours.
The technical possibility to expand radio-cells up to 50km and the strong increase of possible neighbours leads to unresolved challenges and scientific questions in the scalability and the security-infrastructure of Mesh-Networks.
Within the scope of the Mesh-Sequencer Project a systematic analysis and rating of the existing methods with their weakness according to scalability and securing of the network should be done in the first step. In the second step we try to use this list of critics to develop suggestions for solutions and try to confirm their validity by simulations.
In prototypes the new methods should be implemented in existing Mesh-Hardware (e.G. the MeshCube).

RAIP

Integrated prototyping addresses the engineering challenge to coordinate multidisciplinary construction of hard- and software for complex systems under pressure of shortened time to market. The realtime power automation lab will demonstrate advanced methods developed in energy-conversion and control, power- and microelectronics in industry-near context. Implementation into different industrial products and plants is made visible and use of tools for target conversion demonstrated.
Contact : Jürgen Sachau

SECAN-Lab

Interoperability Laboratory for Security in Ad-Hoc Networks

In the SECAN-Lab the research is targeted (as its name indicates) towards the area of interoperability- and security-aspects of spontaneously created and self-organized wireless networks (MANETs =Mobile Ad-Hoc Networks).
Because of the spontaneous character of Ad-Hoc networks without any centralized infrastructure, participants of such networks depend on the friendliness of their neighbors, if a communication to a distant device
is required. Security-mechanisms like encryption or PKIs are not suitable, if the delivery of data needs to be ensured. These technologies cannot guarantee the absence of malicious or unfair participants of the networks.
There are better ways to increase cooperation. Exclusion or motivation are two ways to go, but these techniques may have problems with definite identities. Therefore the SECAN-LAB focuses on trust-relations as a means to increase cooperation. Trust can be rated between direct neighbors as well as between devices in n-hop distance.
Other targets of research of the SECAN-Lab lie in the interoperability of 1:1-communications, the scalability of Ad-Hoc networks compared to its number of participants, the tolerance of the network to a certain amount of malicious and unfair participants and the linking of Ad-Hoc networks with wired LANs.
The lab sees its focus beside theoretical analysis and development of solutions also as a laboratory of practical experiences with newest mobile devices, especially for securing 1:1-communications in n-hop-distance.

R1F103T02 SoNI 2003-2006

The SoNI project aims at performing research at the border between peer-to-peer environments and ad-hoc networks. SoNI shall provide new middleware platforms for enabling the development of collaborative and reactive applications on the top of ad-hoc, self-organizing networks. These networks encompass mobile devices such as notebooks, PDAs , mobile phones, as well as traditional desktop computers. Primary goals include lower-level routing protocols, a sophisticated service architecture framework, as well as the provision of context-aware services.
UL Members: Prof Dr Pascal Bouvry, Prof Dr Thomas Engel, Assistant-Professor Dr Steffen Rothkugel, Luc Hogie
Contact : Pascal Bouvry
Project-Page : SoNI

ABASSMUS 2004-2006

The emerging future generation of mobile devices will allow communicating in two ways: both via GSM/UMTS as well as via wireless network technologies like Wi-Fi or Bluetooth. Hence, in addition to setting up a link to a backbone network like the Internet, pure local ad-hoc communications is another and complementary option. The ABASSMUS project aims at providing an integrative system architecture that combines the features and advantages of both backbone-based and ad-hoc approaches. We aim at ensuring that adequate security services based on immune systems, particularly in terms of intrusion detection, will be embedded into the environment, taking advantage of contextual information that is available in such an integrative system architecture.
Contact : Steffen Rothkugel
Project-Page : ABASSMUS

R1F104T01 Evo-business 2004-2006

The aim of the project is to automate business decisions in e-business. The proposed solution will be based on a near real-time decision support tool and on-line data analysis & pattern search.
Currently, we focus on different research aspects, including:

• Multi-agent co-evolutionary algorithms
• The fusion of different/distributed data structures
• The generation of a communication interfaces

The implementation of algorithms to discover emergent patterns (e.g. hierarchical structures, temporal patterns, periodic patterns).

UL Members: Prof Dr Pascal Bouvry, Assistant-Professor Dr Christoph Schommer, Grégoire Danoy, Ben Schroeder

Contact : Pascal Bouvry

FESM

Distributed monitoring databases together with standardized evaluation tools are the basis for supply quality management, system comparison and improvement of the whole variety of distributed power systems. Energy system informatics plays a key role ranging from embedded control to optimizing supervisory control and techno-economic design. The aquisition and use of operational data allows for continuous improvement taking into account the experience from system operation together with evolution of supply demand.
A core approach in the transition to sustainable energy infrastructures is the conceptional and operational integration of energy management on both supply and demand side. Widespread use of de-central computing and networking with energy management centres allow advanced methods closing the cycle from design and implementation to monitoring in a continuous improvement process.
The results will assist in the planning of national, regional and local frameworks and projects on Europe's way to sustainable energy and transport with emphasis on renewable hydrocarbons. Instead of increasing financial compensations for emissions, investments are attracted in new industries and crafting, which are suited to compensate for the decreases in traditional industries. The systematic approach through all levels from national over regional to local scale, gives a transparent techno-economic framework applicable by most diverse policy making, planning, implementing and operating partners.
Contact : Jürgen Sachau

HyWercs 2005-2007

Hybrid Wireless Networks provide the potential for combining the advantages of both infrastructure-based as well as ad-hoc networks. The HyWercs research project is focusing on middleware support for applications in this realm. In particular, we aim at optimizing the information exchange process between multiple self-organizing groups of users sharing common interest and a shared backbone infrastructure. Concepts introduced and investigated include so-called active information items that are disseminated using an injection-based communication paradigm.
Contact : Steffen Rothkugel
Project-Page :HyWercs

INTRA 03/2004-12/2007

Within the global framework of the information society, fast, reliable and safe data exchange between local and terrestrial wide-area computer networks has become a top priority problem. Complexity is a key issue in the performance of information traffic. Evidence of traffic complexity appears in many forms, such as the heavy tails distributions, long-range correlations, self-similarity found in traffic measurements at different time scales. The complexity revealed from the traffic measurements has led to the suggestion that information traffic cannot be analysed within the framework of available models. At the same time the performance of computer networks crucially depends on the traffic assessment. In this connection, a major challenge for the emerging high-speed integrated-services communication networks is to elaborate a reliable model that can realistically capture the basic features of network traffic. Such a model will serve as a basis for development of methods and tools for quality assessment, providing more efficient control and management of information flows and protection of computer networks against intruders, etc.
Contact : Ulrich Sorger