J. Robert Ensor

The rapport multimedia conferencing system

This paper describes the design and initial implementation of the Rapport multimedia conferencing system, which supports interactive, real-time, distributed conferences among two or more users. Using computers connected by data and voice networks, this system creates an environment in which many sorts of meetings can take place, including telephone conversations, discussions among colleagues, and lectures. Rapport provides new opportunities for meetings, allowing a user to interact with distant colleagues and to participate in several conferences concurrently at his or her workstation. The system allows many existing computer programs to be used, unmodified, within its conferencing environment. Thus computer-generated data and displays are available to the conferees for manipulation and editing, enhancing the exchange of information during meetings. Although Rapport does not encourage a particular methodology of meeting conduct, such specialized support can be built using Rapport. The paper outlines Rapports conference abstraction and the model on which it is based — the “virtual meeting room.” It then presents an overview of Rapports architecture, discusses the systems environmental requirements, and concludes by mentioning some of our plans for future work with Rapport.

On SIP performance

The performance characteristics of Session Initiation Protocol (SIP) network elements determine what services can be supported by SIP networks. These elements must be able to provide customers with appropriate response times for new services, e.g., push-to-talk service. In IP-telephony converged networks, SIP-based signaling elements must meet the requirements of telephony signaling — that is, SIP-based servers will be required to process messages according to real-time constraints. Performance of SIP network elements also determines how much SIP networks cost. To support large customer bases, service providers must provide adequate performance on economical components. In this paper, we examine the main processing requirements of SIP elements and describe a number of performance metrics. We support our analyses with measurements of four different implementations of selected SIP functions.

The peloton bicycling simulator

Peloton is a sports simulator that uses the Virtual Reality Modeling Language (VRML) to create virtual environments for athletic training and competition. In particular, it creates environments suitable for touring or racing along road courses. Users participate in simulations by walking or running on treadmills or by pedaling stationary bicycles. Each user’s exercise equipment is attached to a local computer, and these computers are connected to simulation servers by the World Wide Web. Users may exercise alone or share Peloton virtual spaces with others. Peloton’s virtual environments reproduce sights, sounds, and terrain features from actual road courses. The visual component of this environment is a synthetic, three-dimensional landscape, modeled in VRML, which combines computer-generated graphics with images of actual road courses. The virtual environment also provides simulation participants with force feedback corresponding to a course’s changing terrain. The simulator changes the incline of the treadmill or alters the pedaling resistance to create this feedback. This paper describes the bicycling simulations enabled by Peloton. It also discusses the role of VRML in the system’s implementation, focusing on how its virtual worlds are generated from topological data, how cameras and third-person viewpoints are managed, and how unconventional input/output devices are incorporated. CR

Techniques for handling video in virtual environments

This paper discusses ways to incorporate video displays into virtual environments. It focuses on the virtual worlds created by a distributed multi-user simulator. Still images or video streams represent spaces within these three-dimensional worlds. The paper introduces techniques to deal with avatar movement into and out of video regions. In one technique—media melding—when an object moves from one region to another, the media used to represent that object correspondingly change. In a second technique—object tracing—when an object moves from one region to another, its actions in the second region are represented by a trace object in the first region. Pyramidic panels provide a means of dealing with viewpoint changes so that two-dimensional images and video clips can successfully simulate threedimensional spaces. The paper concludes by suggesting ways to extend our techniques and by listing possible future studies. CR

Director: An Interpreter for Rule-based Programs

CUCS-231-86 Programs interacting with users via natural language interfaces generally require more sophisticated control structures than those needed by programs interacting with users through less flexible mechanisms. This paper describes our development of Director, an interpreter (inference engine) for rule-based programs. Providing an efficient combination of forward and backward chaining, heuristic and user control of inference, and ready access to portions of its internal sttucrure, Director facilitates the construction of systems with natural language interfaces as well as other rule-based systems in which queries are expensive. This research was partially supported by Office of Naval Research grant NOOO 14-82-K-0256. Director An Interpreter for Rule-based Programs

Data to the People - It's a Matter of Control

This paper highlights some early results from our current research. We are building prototype multimedia services, e.g., digital television, video on demand, and teleconferencing, for metropolitan area networks. Our work illustrates that application-aware resource management plays a critical role in providing such services economically. The paper describes a middleware system--called Narnia--that supports the development and deployment of these application-aware resource controls. Using Narnia programming abstractions, application developers can create flexible, dynamic resource managers as part of multimedia service implementations. The paper also presents a brief summary of our experiences using Narnia to build applications.

Visual Techniques to Accommodate Varying Network Performance in Virtual Environments

This position paper briefly reviews how we created virtual worlds for a real-time, multiparty simulator. These worlds are partitioned into regions. Some regions are displayed as three-dimensional objects, and other regions are displayed as still images or video streams. We hypothesize that the size and placement of these regions can be configured to accommodate various user requirements. Such configurations can be based on static descriptions of the computing and communication resources that are available to users. We also hypothesize that dynamic region configurations can be used, in distributed systems, to accommodate run-time variations in computer and network performance. These configuration changes can be based on real-time measurements of system elements.

Pop-Out Videos

This paper discusses a class of multimedia displays that we call popout videos. A pop-out video is a composition of multiple elementary displays. Some of these base elements are video streams, and some are three-dimensional graphic displays. The elementary displays of a pop-out video may be related in two ways. They may be contiguous portions of a single virtual world, forming what we call an integrated video space. Alternatively, they may be parts of distinct spaces, forming what we call a complementary video space. In both cases, the elementary displays are related to each other by coordinated event handling. In our on-going research, we are using pop-out videos to create multimedia virtual worlds and to orchestrate the presentation of data in multiple media. After a survey of related work, the paper outlines our initial implementations of pop-out videos and presents future plans.

Getting the model right for video-mediated communication (abstract)

Videoconferencing, videophone, media space, electronic meeting rooms — these are just some of the abstractions that the research community have evolved for computer-mediated video communications. Each abstraction implies a particular context for communication, supported by an appropriate User Interface design and set of multimedia services. The success of public telephone services owes much to the simple but powerful call abstraction that unifies the services provided by the PSTNS. If video really is to succeed, a uniform abstraction seems necessary both from the users point of view. and for systems developers. Which of the current models, if any, is most appropriate? Perhaps the abstraction that will succeed has yet to emerge or will evolve slowly from a merger of the various approaches. Meanwhile, the engineers forge ahead in the construction of national and international communications infrastructures, so these issues must be raised now. Users will need models of VMC that help them make effective use of the wide variety of multimedia services that will soon be available.