David J. Farber

Professor of Computer and Information Scienc

200 South 33rd Street

Philadelphia PA 19104-638

EMAIL: Farber@cis.upenn.edu

To be presented at the The 2nd International Conference on Broadband Islands in Athens Greece 15 June 1993

Abstract

This paper has two aims. First, it proposes an international experiment involving the three major areas of the developed world--Asia/Japan, North America, and the European Community--and the three evolving broadband experimental testbeds currently being developed. In addition, it makes some projections as to where the long term markets will be for very high bandwidth computer communication networks.

The Black Box Network experiment

Each of the developed areas of the world--Asia/Japan, North America, and the European Community--have impending research and advanced development projects whose aim it is to develop very high bandwidth--broadband and beyond--computer communication networks. In all three areas, it is believed that these developments will pave the way for future national infrastructures such as the U.S. National Information Infrastructure (NII).

Each of these experiments is currently an island. There are no bridges or even, to stretch the metaphor, ferry boats. The technology being developed cannot benefit from the effects of collaborations, standards, and applications that could be developed if these islands were connected together. There are many reasons for this isolation. There are some minor technical reasons having to do with obtaining broadband, cross-oceanic communication links, as well as some minor issues having to do with different standards and approaches. In practice, though, the main impediment to international collaboration has been the belief on the part of some of the players that collaboration and cooperation will cause a loss of technological leadership and thus endanger their future commercial competitiveness. It is not profitable to argue as to whether these beliefs have any grounds in reality. It is the author's strong belief that they do not. However, changing such viewpoints will require time and patience, both of which tend to be in short supply

Using this as a model, I propose that each of the geographic areas--Asia, North America, and Europe--interconnect their broadband/gigabit testbeds for a set of exploratory applications in the international use of broadband networks. Each of the areas could keep its internal technology private if it so chose. For example, switch architecture, routers, network management tools, etc. could all be kept private. Each area would, however, be responsible for disclosing to the other areas and standardizing with the other areas the internetwork interfaces which would allow applications to operate between and within the area networks. The results of such agreements would form the basis for future international standardization of broadband applications and lower level protocols, while protecting any perceived future commercial advantage.

What would be the nature of such an experiment? I believe that an initial experiment should be focused on the higher education community. This community is the most amicable to change and is used to being experimented upon. The environment I propose we create is modeled after a proposal generated by the author while at the University of Delaware in 1977. This proposal was called DUNE (Distributed University Environment). The idea behind DUNE was that each classroom in each participating university would be equipped with a device called a DUNEBUGGY. The DUNEBUGGY has a wide-angle television camera and speaker system designed to allow a person to remotely view the proceedings in the classroom and to interact with a lecturer/speaker. The DUNEBUGGY would be interconnected with a local campus high-bandwidth network, and with the national high-bandwidth network. In the Black Box Network experiment, I propose we equip two campuses per geographic area with such a capability. This would allow students to attend (from workstations in their offices or homes) lectures anywhere in the world (at least, the world covered by the Black Box Network). DUNE would allow two-way interaction with faculty and students world-wide. There are many questions raised by such an experiment, some of which are simply answered. For example, time zone differences might pose a problem. But students in particular are well-known to have highly variable clocks. Other difficulties, such as language, are not so simple. The prospect of lecturing at universities in a single language is probably impractical in our world today. However, such a network would provide the incentive for students to learn the languages of other countries if the prize were the ability to attend the lecture of a great mathematician in, say, Japan

[1]

There is much criticism in the USA that centers on the claim that Broadband networks are a toy of the rich scientist. The Luddite fraction says that rather than broadband to the few, it should be kilobits to the masses. Arguments such as industrial competitiveness in the international arena and such get very little response from these people. Perhaps a more challenging question come in the form of where is the market for broadband? Will perhaps non multiplexed broadband networks be the first failure of digital technology to find a true market. When one attempts to describe markets for such networks video tends to play a key role. Unfortunately compression technology has greatly reduced the demand for bandwidth in order to transmit even HDTV. So perhaps we have a technology with no marketplace.

Before we all leave the meeting and return to V32 modem technology, let me state that I believe this is indeed short sighted thinking. We see already applications in the medical imaging area that need such bit rates. But you might say, how can I support the development of such an expensive technology based on such a small market.

I will now explore an application marketplace where if indeed our projections are correct, the market will not only be substantial but key in the success of the fundamental changes in business that have occurred due to the shrinking world. In fact it may be more important to the EEC than to the USA or Japan

The University of Pennsylvania is part of a landmark project -- Aurora, with counterparts at the Massachusetts Institute of Technology and with scientists and technology developers from both Bellcore and IBM Research working at their respective facilities in New Jersey and New York that - if successful - could become a watershed event in the transformation of the way research, education, business, and most coordinated work, is conducted in the coming decades.

As part of the national effort, researchers linked to the four site test-bed will be responsible for both exploring alternative network technologies and new paradigms for a network service/application interface. But, just as important, is the concurrent exploration of the ways scientists and research groups, designers and project teams, individuals and organizations can make use of broadband speed networks. The promises offered by multi-media exchanges enhanced by ready access to multiple sources of transformable, self-tailored data are great. We, at Penn, believe that these next-generation networks could provide the bandwidth needed to create for geographically dispersed collaborators a "virtual" workplace that will not be just a mere substitute for the traditional meeting room, but instead an enhancement providing added "intelligence" for such purposes as filtering information, exception reporting, and meeting management, including services equivalent to a large backup staff, as one might have for a board of directors meeting.

Supporting group work conducted not for the completion of a specific project, but instead for on-going coordinated efforts, group consensus and organizational innovation remains a challenge for developers of communication and information systems. Applications designed to support distributed work have failed, according to recent research, because they:

do not allow for the wide range of exception handling and improvisation that characterizes group activity;

often threaten existing political structures;

are usually so complex that they handicap meaningful, generalizable analysis and evaluation

de-motivate the users who are most crucial to its success, providers of information upstream.

I would at this point like to briefly explore a number of distributed applications to give a feeling for the issues and approaches being considered.

Distributed Software Development

The work necessary to start-up and manage the Broadband efforts will be software oriented and distributed. This setting is well suited to drive our efforts The process of negotiation necessary so that the tasks to be performed by collaborators at each site are understood, compatible and completed will require the kind of unique combination of rich graphics and media that we are envisioning in this research. During "video" conferences from their workstations and/or conference rooms, Designers will be able to discuss the interface and design issues, seeing the Pert charts and design diagrams displayed in immediate response to the changing assumptions and approaches being taken by others. For example, it will be possible to integrate the time and cost budgets into discussions about the distribution of responsibilities, as well as design bugs - the solutions of which often affect budgets. Technically, this is not easy. The current software design process is not easily extended to the type of information sharing and presentation implied by such an environment.

In the business world distributed processing to support decision making is a prime candidate for the innovative thinking we suggest.

There are complex research issues in multi-party interaction in that it represents not only a difficult technical control problem, but a social one. Necessary research will draw upon both human communication theory and past experiences with the design of human computer interfaces. One assumption in classical communication theory is that for human communication to take place, there must be something shared at the start. Sharing usually is taken to mean having access to the same reality, documents, images and physical space. This assumption has received two challenges. One comes from constructivism and suggests that individuals cognitively construct their own world in which to understand their own actions, including communication acts with others. Cognitive sharing, according to this approach, literally is impossible, at best unlikely, and certainly not adequate to explain human communication. The other challenge comes from recent ideas in human computer interface design, enabling users a considerable freedom to configure their own local work station, enabling them to construct virtual realities without impairing the ability to translate data, documents and images from someone else's reality into one's own. The messages sent need not be identical to those received. The commonalty on which communication is based therefore is virtual, not real.

Technology that links individual realities has profound practical implications in all kinds of computer mediated conferencing, distributed problem solving and decision making, especially when communication partners have or wish to retain their own perspective, their own culture and familiar world constructions

Decision Making within the Business Setting:

Much has been written about the probable effects network technology might have on the organization of the future, i.e. the decentralization of responsibilities to middle and lower managers, increased empowering of teams, multidimensional organizational structures that rely on knowledge rather than hierarchical authority, the rise of the "virtual organization" consisting of a minimal core for most business functions. Rarely does the purchase and installation of new technologies to support the sharing and exchanging of information lead to automatic adoption within a large organization. Stories of subordinates not revealing sacred information, even under organizational insistence, are many.

If we are successful, perhaps we can create a market that will be substantial enough to attract the attention of the carriers.