Site: German National Research Center for Information
IPSI for Integrated Publication and Information Systems
Date Visited: 28 April 1999
WTEC Attendees: A. Ephremides (report author), R. Pickholtz, L. Katehi, T. Itoh, D. Friday
Hosts: Professor Dr. Wolfgang Schoenfeld
Mr. Matthias Hollick, GMD-IPSI
Ms. Nicole Berier, GMD-IPSI
Professor Dr. Ralf Steinmetz, Professor, GMD-IPSI and Technical University of Darmstadt
Dr. Lars Wolf, Technical University of Darmstadt
Ms. Ana Pajares, Technical University of Darmstadt
GMD is the principal research organization of Germany that is primarily publicly funded and that focuses on information technology. Although detailed budget figures were not presented, a broad picture of GMD's activities emerged from the discussion. The organization is composed of eight institutes each of which focuses on separate (although interrelated) sectors of research activity. Four of these institutes are located in the Bonn area (along with GMD's headquarters); two are located in Berlin; and two in Darmstadt. IPSI focuses on information distribution. As such it encompasses activities in the following:
The main activity of the institute that formed the basis of the discussion for the purposes of this study was that of mobile interactive media. As it emerged from the discussion, the main thrust of IPSI's efforts in this area are focused on the upper networking layers (i.e., on the effects of mobility and of dynamic environments on network protocols) and not on the wireless medium per se.
There is close interaction between personnel at IPSI and the research staff (faculty and students) of the Technical University of Darmstadt. In fact, at the WTEC meeting, representatives of both organizations participated.
After the WTEC panel made a brief introduction and described the objectives of the study, Professor Dr. Schoenfeld proceeded with an extensive review of IPSI's programs on mobility issues. Later Professor Dr. Steinmetz joined the meeting and complemented, with a brief commentary, the presentation of Dr. Schoenfeld. After lunch, Dr. Wolf described some additional projects that are jointly pursued by IPSI and the Technical University of Darmstadt. The meeting concluded with a broad discussion among all the participants.
The main focus of IPSI's programs in mobility studies is on seamless roaming for interactive multimedia in dynamic environments in which the logical connectivity among the hosts and the routers is changing. As such, it concentrates on the higher layers of the communication process.
As described by Dr. Schoenfeld, the objective of mobility tracking is to maintain consistent information about every node's connectivity throughout the network. The networks of interest include WLANs, cellular networks, fixed wireless systems, and even systems with satellite links. When an end-node in a routing tree changes position, well-developed and well-understood methods of registration and reporting (from cellular networks) can be used. However, when intermediate nodes change position and connectivity, then the method of "tunneling" can be used. This method has formed the basis of mobile IP and represents a short-term solution since the resulting new route is in general non-optimal. The long-term solution would have to be based on rerouting. This is not easily accomplished in IP networks. Ongoing and future research on this issue is addressing this problem.
In a sense, the search for improved solutions in this context suggests that routing in ad-hoc networks is an important area that may have useful fallout in mobile computing. In fact, the DFN-Verein, which operates the network that interconnects universities and research institutes in Germany, maintains an interest in mobile computing and, through a new program, called MIRIAM, will permit IPSI to test the use or the modification of mobile IP as part of the research in adaptive routing.
A related problem that IPSI is investigating involves the continuous registration of a mobile host at a (possibly remote) home agent. The research concentrates on the theoretical modeling of this issue and on the resource management problem that arises when links that support different quality of service must be chosen. An E.U. project called COSMOS, in which IPSI is participating, addresses this issue. The role of IPSI in this project, which assumes the use of satellite links as well, centers on middle-ware development.
Although IPSI recognizes the importance of the effects of the wireless link on the higher-layers (e.g., QoS vis-a-vis mobility), its primary concern is the development of upper-layer protocols that are compatible with, or are optimized for, dynamic topologies. By the same token, these upper-layer issues are crucially important and, yet, are usually ignored by the physical-layer community.
To handle environment changes, a service location protocol is needed and a classification of services is necessary in order to determine which services should be provided at which location. In this regard, it was clarified that location refers to the notion of topological connectivity and not of actual position coordinates (which is an objective of E-911 in the United States for future wireless services).
To assist in these projects, a Mobile Network Emulator is being developed that will actually include a link propagation model.
In response to a question from the panel, it was clarified that multicasting applications are not currently part of the focus in the IPSI projects.
The meeting, as mentioned earlier, included a description by Dr. Wolf, of the projects that are jointly studied by the Technical University of Darmstadt and IPSI. These include a variety of topics that revolve around multimedia networking from the point of view of content and information distribution. Some of these topics are quality of service, heterogeneity, Internet telephony, security, gateways and protocols, content processing, distance education, middle-ware, video servers, and content distribution. As the nature of these topics makes it clear that the emphasis is on general networking and not on wireless networking. Nonetheless, the relation to mobility (from the point of view of long-term mobility, i.e., the prediction of needs at different locations) renders this research relevant to wireless networking. Also, the issue of pricing of services, which in its own right has long-term implications for Internet networking as well as for wireless communication, is on the research agenda of IPSI and the Technical University of Darmstadt.
In the ensuing broadened discussion, in which the panel's hosts were explicitly asked to address the long-term research issues and bottlenecks in wireless and mobile networking, it was mentioned that there is a 2005 Master Plan for IPSI. This Master Plan encompasses the areas of media streams (including entertainment services), ubiquitous computing, education, applications (such as telemedicine), and technology (e.g., security). Furthermore, it was pointed out that the ultimate bottleneck in network communications would be the "usability" of the products and services, i.e., the needs and capabilities of the consumer and user.
The visit at IPSI highlighted issues in wireless networking that are often ignored by the wireless community, namely the need to address the design of upper-layer protocols and their effects on link operation. However, the WTEC panel visit verified that an integrated approach to networking that bridges gaps among layers is needed but not yet in place.