Telecommunications networks for remote interactive multimedia

Advances in multimedia workstations, software systems, storage for multimedia and user interfaces which allow the presentation of multimedia information have produced an environment in which interactive multimedia systems are developing at a rapid pace. Most current systems are local in their service range, but there will be an increasing demand for long distance multimedia communications. Multimedia communication issues are discussed in this paper. Present day networks go some way towards satisfying the service support requirements of multimedia, but the future Broadband Integrated Services Digital Network (B-ISDN) will ultimately have the capacity and flexibility required to ensure multimedia applications are completely freed from geographical restrictions.

1. Introduction

Hardware developments have led to the ready availability of 24/32 bit colour display boards for PCs/workstations, allowing the presentation of high quality colour still and moving images, and graphics. "Video in a window" boards complement this by permitting real time video to be presented with other information in an integrated manner. The capability to reproduce high quality sound is also becoming more common. At a more fundamental level, the processing power and backplane speeds of computers permit the manipulation and transfer of large amounts of information, including motion video. Rapid, efficient image compression is possible in a single inexpensive chip. Plug in boards that implement standard motion video coding methods in real time on just a few chips will be available within the next year.

Complementary to advances in PC/Workstation hardware are the software developments that allow these hardware capabilities to be exploited. These include the development of composite documents which combine media (e.g. text and graphics) in the one document. This now allows desktop publishing in which one can actually see what is to be printed. It is also now possible to paste, not just line drawings, but high quality images into other documents, providing a new flexibility for desktop publishing and document preparation. Software systems are emerging that permit shared working by two or more users on the one document. Co-workers can therefore edit the same document interactively and simultaneously. Hypermedia systems offer the potential for information presentation in the most efficient and flexible manner possible in any given application by providing a hierarchical and cross indexed structure to complex multimedia documents. A very recent advance has been the emergence of real time information handling software systems suited to the manipulation of real time information such as audio and video, providing the capability to manipulate (record, store, play, edit, copy, paste) these in the same way as ordinary documents.

Standards are also beginning to be developed to permit the widespread exchange of multimedia information. The ISO is studying stored audio and video systems, database access, multimedia/hypermedia representation, computer graphics systems and others. CCITT covers telecommunication service definition, network support, signalling and coding of communicative audio and video. Relevant activity in CCIR includes the television service definition and CMTT covers standards for coding of broadcast services for transmission.

While some of this standards activity concerns familiar applications, new and vastly different applications will emerge from some of the computer oriented developments. Initially, these will he confined to local applications, using local storage devices and proprietary interfaces and storage formats. At best, some of these could be shared over a Local Area Network (LAN). Increasingly, though, there will be pressure to extend the range of multimedia applications so that even sophisticated multimedia applications, like hypermedia documents, can be transferred across the country and around the world, and information in a variety of formats can be accessed from diverse and remote locations. There will, therefore, be increasing pressure on public telecommunications network providers to support this type of communications.

2. Multimedia communications

• Range of service types. The support of different media (audio, video, data, text, graphics) in a range of different configurations will demand of the network flexibility unlike anything seen previously. A given service may have to be delivered at a range of qualities (e.g. video conferencing at a range of different picture qualities). Different media may be better suited to different network qualities of service. Some services will be interactive, others communicative or distributive. Some will be symmetrical (e.g. two way videophone) and others asymmetrical (e.g. request for video segment from a database).

• Association 'between services. Multimedia information implies some form of linkage between the component parts of the application, and this requirement may impact significantly on the structure and functionality of a telecommunications network. Most obvious, perhaps, are the synchronisation requirements between associated audio and video channels, or between stereo audio channels. Other "links" include tariffing arrangements for a single call comprising several components, and the call admission (ie. allowing establishment of the call) and monitoring of the ensemble. A particularly complex issue arises in the support of hypermedia, where a request for information over one connection could initiate a completely new connection.

• Multipoint. The expansion of multimedia applications beyond local use (on a self contained PC/workstation), by permitting long distance communication, opens up the possibility of multipoint multimedia applications for negotiation, shared working, telecommuting, etc.

• Distribution of information sources. Many issues arise regarding the location and distribution of information sources once network access is available. For example, access to vast central or distributed databases can be provided over the telecommunications network, and the range of multimedia information used in a particular application might be distributed throughout the network instead of being localised on the one CD-ROM in the local application.

3. Multimedia communications - network options

3.1. Integrated Services Digital Network (ISDN)

For some multimedia applications, standard transmission methods have been defined that will support combined data, audio and video over the ISDN. CCITT Rec. H.221 (CCITT, 1990) defines a multiplexing structure for several 64 kbit/s channels used to support audiovisual services defined by other CCITT Recommendations. However, this structure is not intended for universal support of multimedia services.

3.2. Metropolitan Area Networks (MANs)

By providing transparent interconnection between LANs, these networks will provide the infrastructure to allow remote interactive multimedia applications to flourish. The remote systems will appear to be connected to one very large LAN, and the high capacity will ensure that reasonably fast response times are achieved. The MAN standard (IEEE 802.6) is based on fast packet switching methods in which LAN packets are split into a number of segments for transmission and switching. This means that a wide range of service rate demands can be accommodated and that connection control for different applications can be integrated. MANs are seen as a major stepping stone towards the B-ISDN.

3.3. Broadband Integrated Services Digital Network (B-ISDN)

The use of Asynchronous Transfer Mode (ATM) (again packet based, using short, fixed length packets called cells) as the transport mechanism for the B-ISDN is intended to provide flexibility for the carriage of all known, as well as unforeseen, services. The cell based transport allows adaptation to required service rate requirements and accommodates asymmetrical bandwidth needs easily. Furthermore, low level signalling capabilities permit multiplexing of the different components of a multimedia connection at the cell level. That is, each media has its own cell stream, permitting great flexibility in the routing of the individual media within the network, as well as matching of the cell stream attributes (such as quality of service, and therefore cost) to match the needs of each medium. For example, file transfer data may be much more tolerant of errors or cell losses, because of high level error control, than real time video services.

A discussion of the B-ISDN and its suitability for the support of multimedia is provided in Biggar (1991).

4. B-ISDN Support of interactive multimedia

4.1. Service component synchronisation

Examples of multimedia applications in which synchronisation will be most important include audio/video synchronisation for both distributive and communicative video applications, shared workspaces in which it is essential that the same workspace is seen simultaneously by each participant.

4.2. Multipoint working

A range of issues arise when considering multipoint, multimedia service capabilities. Some of these are:

• Multipoint facility. Existing multipoint capability (for audio or video conferencing) is provided by a central bridge that mixes the signals as appropriate and presents the result to each participant. It is already recognised that new possibilities are available in the B-ISDN. For example, the video signal could be multicast to all receivers, while the audio component only is mixed at a bridge. The diverse range of components in more complex multimedia communication sessions must be considered with regard to how the multipoint capability should be provided and how the components interact in a multipoint environment.

• Dissimilar terminal types. In a conventional multipoint video conference, all participants have essentially the same capabilities and requirements (with the exception of a controlling chairman). In the more diverse world of multimedia, however, some of the "participants it could be databases called into the "conference" by one or more of the other participants. The implications of this must be considered. There will also be diversity in terminal capabilities, some not even supporting all the media. How will networks know of the capabilities of the various terminals, and how will they decide what is and is not a useful connection? Methods of handling this situation in a multipoint environment must be considered.

• Shared control. In a multipoint video conference involving just audio and video media, transfer of "control" between participants is relatively straightforward. Either a chairman determines who is seen and who can speak, or a bridge can make the decision according to the present speaker (voice activation). It has already been noted that other possibilities exist once the B-ISDN is used for service delivery, since more individual control of who is seen should be possible. In a more complex multipoint multimedia session, however, there may be greater need for interaction and individual control. For example, in a shared workspace application, where a document is being discussed and edited jointly, several participants may wish to initiate actions on the workspace without (knowingly) having to gain formal control of the cursor. All requests for action cannot be carried out in an ad hoc way, though; two instructions could be queued for processing, the first of which changes the workspace upon which the second acts. Mechanisms for negotiation and allocation of control of the workspace are therefore essential, and these must be easy to use, not significantly different to the way the operations would be carried out if working solo.

4.3. Service integration

The most obvious application of this concept is in the most demanding service type - video. Work is already well advanced in the definition of hierarchical, or layered, video coding architectures that allow interworking across a broad range of services. This will mean that one terminal can accept and present a video signal almost independently of the original source and transmitted resolution, frame rate, etc (Scott, 1990). Complementary studies are required on integration of the various other service components.

4.4. Additional presentation issues

These issues include: how to provide control of the connections, which may be to different sites, in a manner that is "user friendly"; what tariffing feedback should be provided to allow the user to monitor and "control" the price paid for a particular multimedia connection; and how control of a multimedia connection is to be divided between the user and network operator, and whether the arrangement should vary between media.

4.5. Media conversion/translation

An example where this capability might be required is the interrogation of a text based electronic mailbox via an audio only telephone. If this capability is to be permitted, then text to speech conversion will be required somewhere in the link. Another example might be the inclusion of a participant with videophone capability into a multimedia conference where computer graphics files are being exchanged. In this situation, the means to convert the graphics image to standard video signal format would be required.

Investigations are required to determine the range of media conversions which may be required, the feasibility and importance of providing them and the important question of where the conversion should occur. Must it be provided within the network or is it a terminal/database issue?

4.6. Interworking across networks

While interworking across networks is often considered to be solely a network consideration, many service issues also arise, such as how much of the multimedia multiplex should be passed through a network constrained by bit rate or functionality. How much of the decision making process should be given to the user, and for how much will the network intelligence be responsible? How should a multipoint connection be implemented if one of the participants can only be contacted over a low rate network or on a terminal of limited capability (single media)?

Integration of each of the various media is also important in this situation since, for example, the layered "Universal Video Coding" architecture would permit lower rate (and quality) subsets of a video signal to be passed over a rate constrained connection (Tan, 1991).

5. Conclusions

At the same time, switched network capacities are increasing (and developments in other areas such as signal processing occurring,) to the point where we can consider support of more traditional media, such as television and high quality audio distribution, that has not been possible on the public network in the past.

These developments indicate the need for an overall framework for the support of these services under the banner of "multimedia" (which can include single media as special cases). Identification of the special requirements of these services, how these can be satisfied, the impact on the network and the most effective means of presenting the service to the customer are important areas for further investigation.

This paper has provided an initial view of the impact that multimedia services may have on communications networks, what the network options and evolutions may be, and a more detailed look at some of the issues that must be considered as the B-ISDN evolves to support this important service category. The B-ISDN has the transmission capacity and flexibility suited to the support of all the components of a multimedia connection, and should be able to deliver very fast response times and a high perceived performance (potentially indistinguishable from operation with local peripherals). However, there remain several important areas to be studied to ensure that this vision can be realised.

6. Acknowledgment

7. References

CCITT (1990). Frame structure for a 64 to 1920 kbit/s channel in audiovisual teleservices. Recommendation H221.

Day, A. and Dorman, D. (1990). Towards an Australian broadband network infrastructure. Telecommunication Journal of Australia, 40(2), 3-14.

Pretty, R. (1990). FASTPAC: Telecom Australia's high speed LAN interconnect network utilising QPSX fast packet switching technology. Proceedings Australian Fast Packet Switching Workshop, Melbourne, 9-11 July 1990, pp. 200-204.

Scott, C., Biggar, M. and Dorman, D. (1990). Getting the Picture - Integrated Video Services in B-ISDN. Telecommunication Journal of Australia, 40(2), 21-25.

Tan, W. (1991). Advanced Video Service Features through the UVC Architecture. Proceedings Australian Broadband Switching and Services Symposium, Sydney, 3-5 July 1991, pp. 8-14.

Authors: M. J. Biggar and C. J. Scott Telecom Australia Research Laboratories 770 Blackburn Road, Clayton 3168, Victoria, AUSTRALIA Please cite as: Biggar, M. J. and Scott, C. J. (1992). Telecommunications networks for remote interactive multimedia. In Promaco Conventions (Ed.), Proceedings of the International Interactive Multimedia Symposium, 313-319. Perth, Western Australia, 27-31 January. Promaco Conventions. http://www.aset.org.au/confs/iims/1992/biggar.html