Collaborative computing

	Scalable Conferencing Control Service (SCCS) Due to my work in the area of group communication protocols (especially the T.120 and Internet stuff), I began my research for the design of a scalable conferencing control service (SCCS) for tightly-coupled environments, which uses an efficient resource management scheme and multicast for a better scalability in large scenarios. SCCS provides a service functionality similar to the ITU protocol suite (T.122, T.124). In contrast to existing protocols, SCCS supports complete reconfiguration of the conference tree topology to provide an optimization of the topology for reduction of response times, e.g. for the floor control. Currently, a complete service and protocol specification is available.
	ITU-based Collaborative Conferencing My T.120 research focus was on evaluation of the existing ITU-based conference control protocols for group communication in tightly-coupled environments, i.e., the standard family T.120 that is being used in H.32x conference systems. This standard provides generic functionalities for data conferencing applications. Several weaknesses of the standard were pointed out during this work, especially in the area of scalability in large scenarios. Additionally, a combination of T.120 and the distributed platform CORBA was studied and implemented to extend the T.120 by some useful features from a user’s point of view.

GCDL: Group Communication Description Language
Modeling group communication scenarios is a research topic currently under discussion due to its complexity in terms of interaction between the entities. The main problem is that commonly used models do not take into account the high correlation of the actions in group communication scenarios and the underlying communication using conference control protocols. Thus, the observed results are not valuable due to the unrealistic assumptions of the model. To overcome these drawbacks, the GCDL framework is proposed enabling to model social protocols in group communication scenarios appropriately. With this framework, simulative as well as analytical evaluation of conferencing systems is feasible without implementing and installing the system under consideration in large environments.
The framework is extensively described in my Ph.D thesis Scalable Group Communication in Tightly-Coupled Environments.

Automata-based performance evaluation
An appropriate tool for the evaluation of the underlying protocol mechanisms is the simulation. For the evaluation of large scalable conferencing control protocols, a load model was developed based ondistributed automata using the GCDL framework as described above. For that, the Specification and Description Language (SDL) is used for the description of the interactions, which is then mapped onto a simulation using an own framework for realization. This results in an interaction-based load model which can be used for performance evaluation. In my Ph.D thesis, entitled Scalable Group Communication in Tightly-Coupled Environments, this model was used for the performance evaluation of main features of the underlying conference control protocol, which was actually SCCS.