Issue No.05 - May (1995 vol.28)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/2.384116
The state-of-the-art in computation, communication and algorithm, and maturation of relevant standards has reached to the point where the development of large distributed multimedia systems providing compelling services to a large number of people is becoming feasible. These services include interactive entertainment, distribution of news and information, on-demand video services, home shopping, interactive multi-user games, and digital multimedia libraries accessible from home. Cable and phone companies, as infrastructure providers and companies owning content to populate the infrastructure, have realized that a capable network coupled with improved computing and compression techniques can soon be capable of delivering interactive services in a profitable manner. A number of alliances have recently been formed between entertainment, cable, phone, and computer companies with the main focus to architect a wide area information infrastructure. Universities and industrial laboratories have been intensively working to define this architecture. A number of standard groups have also emerged to bring some order to the undertaking. Out of all these activities, a picture of future information infrastructure is emerging. In this paper we first describe a distributed multimedia system architecture that can support on-demand, interactive TV applications. The architecture is composed of a hierarchical configuration of multimedia servers and network switches. In such a multimedia system, multimedia data must be compressed, stored, retrieved, transmitted over the network to its destination, and decompressed and synchronized for playout at the receiving site. The actual transmission of media must be managed by considering the time dependencies and relations between the different data types to ensure that media delivery to the receiver site guarantees the presentation of synchronized media to consumers. Fundamental concepts and techniques in these areas include video and audio compression, multimedia data storage and retrieval, networking, and synchronization. In the second part of the paper we address the main issues in designing a terminal device for a multimedia network. This device, called a TV Set Top Box (STB), decodes the information (e.g. video and audio) at the customer premises for display, and provides the interactive capabilities to the user. We present its hardware and software architectures. A major driving force for the designers of an interactive TV STB is how to achieve the desired functionality at the right cost. Although it might be difficult to meet that price with existing components and technology, cost-reduction strategies, such as integration of major functions into a small number of VLSI chips, may satisfy the price goal. Multimedia storage servers provide access to multimedia objects including text, images, audio, and video. The design of such servers fundamentally differs from conventional servers due to: (1) the real-time storage and retrieval requirements, as well as (2) the large storage space and data transfer rate requirements of digital multimedia. In this paper, we present an overview of the architectures and algorithms required for designing digital multimedia storage servers. The objective of error-free multimedia processing and communication it to manipulate and transmit media under time constraints by providing guaranteed services such as the colloquial telephone. Hence, the control-management level of the host and underlying network architectures has become a key issue of any distributed multimedia system. This paper discusses 'resource management' at the host and network level, and their cooperation to achieve global guaranteed transmission and presentation services, which means end-to-end guarantees. The emphasis is on 'host resources' (e.g., CPU processing time) and 'network resources' (e.g., bandwidth, buffer space) which need to be controlled in order to satisfy the Quality of Service (QoS) requirements set by users of a networked multimedia system. The control of the specified resources involves three actions: to properly allocate resources (end-to-end) during the multimedia call establishment, so that traffic may flow according to the QoS specification; to control resource allocation during the multimedia data to adapt to changes when degradation of a system components' capacity occurs. These actions imply the necessity of: (a) new services, such as admission control, at the hosts and intermediate network nodes; (b) new protocols for establishing connections which satisfy QoS requirements along the path from sender to receiver(s), such as a resource reservation protocol; (c) new mechanisms for delay, rate, and error control; (d) new resource monitoring protocols for reporting system changes; (e) new adaptive schemes for dynamic resource allocation to respond to system changes; and (f) new architectures in the hosts and switches to accommodate the resource management entities. This article gives an overview of services, mechanisms and protocols for resource management as outlined above. This paper describes MultiMedia MedNet (MedNet), a distributed collaborative medical multimedia system. MedNet combines neurophysiological, audio and video data in a real-time remote diagnostic and monitoring system, and is used on a daily basis in seven hospitals at the University of Pittsburgh Medical Center. MedNet provides an interesting example of a working distributed and collaborative multimedia system in an exacting environment, involving data types and manipulations that are broader and more demanding than those found in a typical teleconferencing application. Phase I MedNet incorporates digital signal processing, along with distributed systems tools for group communication, multimedia file access and storage, and software for the manipulation and display of neurophysiological information. Communication control in Phase II is managed by a multimedia group process protocol called the Distributed InterProcess Communication System (DIPCS) that is scalable to a WAN. Audio-visual material can provide useful aids for learning when integrated into computer-based teaching systems. However, a teaching system is only useful if the learner remains active and motivated. It is well-known that page turning or browsing does not ensure effective learning. To learn, students must want to learn and must be involved and active. They must be challenged to reason about the material presented. Flashy graphics and simulations are not enough; the experience must be authentic and relevant to the learner's life. Multimedia, when coupled with other improvements in educational software design, can support effective and quality instruction. We propose a model for instruction which moves beyond the 'Tyranny of the Button' and includes use of intelligent simulation, dynamic links (on-line generation of links based on a student's behavior), and multimedia composition and creation. We discuss technological barriers which prevent the widespread development of effective systems along with ways in which these barriers can be broken down. Finally we suggest that truly interactive systems can evolve into multimedia pedagogues and can facilitate a shift in teaching and learning.
Borko Furht, Deven Kalra, Frederick L. Kitson, Arturo A. Rodriguez, William E. Wall, "Design Issues for Interactive Television Systems", Computer, vol.28, no. 5, pp. 25-39, May 1995, doi:10.1109/2.384116