Figure 1. General architecture
The combination of multimedia technology and Internet is currently attracting a considerable amount of attention from developers and potential users alike. However, although many new products are being announced and marketed, the majority of the information on the Web still consists of just images and text, and the few "multimedia" items are in reality only short video clips with little interactive functionality.
Our prototype implementation of an interactive multimedia system, based on the integration of commercial products, provides some original functions for enhanced user interaction, whilst still maintaining complete compatibility with the rapidly evolving context of the Web. It also allows the same "product" to be delivered over connections of different capacity by selectively decreasing the bandwidth dedicated to the video part of the data.
Multimedia
Multimedia technology has already been around for quite some time, but with the accelerated pace of development that it has enjoyed in the last few years it can be now safely said to be a part of our daily lives, both at work and in the home. It finds application in fields ranging from encyclopedias to catalogues, from games to video clips, allowing computers to display to users information in a form that is not only textual, but can include sound, moving pictures, or both.
Up to now, the medium of choice for this technology has been the CD-ROM: due to its high capacity it is able - to a certain extent - to meet the needs of digital audio and video, it is stable and reliable and, last but not least, the cost of the device required to read it is low. As a result, the CD-ROM has become the most widely used medium for the delivery of multimedia material.
Internet
In the same few years, another phenomenon has drastically changed the way in which information is distributed, namely the development of digital data networks, and particularly the extraordinary expansion of the Internet, "the network of all networks". This expansion, which can be attributed mainly to the appearance of the World Wide Web with its "browsers", has spawned a whole new range of communication techniques, all of which are characterised by :
Multimedia and the Internet
Whilst it is still largely true that the type of data currently prevalent on the Web is text and still images, there is clearly a definite need to be able to use the same infrastructure/data networks to exchange multimedia data, including sound and moving pictures. A non-negligible part of this perceived need finds its origin in the requirements of advertising: many "service providers" are financed (also) by commercials, and it is well-known to the advertisers that moving images have a much stronger impact on the viewer than any other form of communication.
It is already possible today using the Internet to listen to music and to radio news, to view short video sequences, and to talk to remote users by means of special applications that emulate the functions of a normal telephone. Nevertheless, it is also true to say that the use of multimedia information via a network is still restricted to more or less experimental applications. This, together with the trend towards more extensive use of multimedia information, has created an opportunity that developers are pursuing very actively.
Consequently, the technological area that lies at the intersection of Internet and multimedia is today the object of much interest and intense development, financed - sometimes with sizable investments - both by the Information Technology (IT) industries and by those whose focus is entertainment: film, television, and radio production, and music recording companies.
The eventual end result of all of the effort that is being devoted to this field is still somewhat difficult to predict. What can already be observed is the very rapid pace in the development of new products, which are often marketed or distributed via the Internet itself.
Open issues
All of the development effort notwithstanding, a certain number of problems remain that hinder more extensive use of multimedia data on Wide-Area Networks (WANs). The most significant of these is certainly constituted by the severe bandwidth limitations of the networks: while recent technologies, such as ATM, allow bandwidths of hundreds of megabytes on any length of link, the majority of Internet users are still connected via a dial-up telephone line and a modem capable of operating at 28.8 kbit/s at most. It is also not unusual for other bottlenecks to occur between the user and the service provider, especially at times of heavy traffic, which slow down the overall speed of the link even further. Another closely related problem is the low predictability and high variability of the overall data rate that can be achieved at any particular time on a given link.
The result is that the Internet user wanting to access video material will need the patience to wait for a long time before the whole file is downloaded, and will then be rewarded with the disappointing experience of a video lasting a few seconds and displaying a picture barely larger than a postage stamp.
In addition, the possibility for interaction with the video clip will generally be limited to start/stop commands, such as can be found on a domestic video recorder
ESRIN is responsible for all of ESA's non-operational data-processing activities, including the distribution to external users of the Earth-observation data collected by the Agency's various remote-sensing satellites. In line with the overwhelming general trend, the Web is the medium of choice for the distribution of most of this data. To date the vast majority of the data consist of text and images, but the requirement to support multimedia applications will certainly grow as user expectations increase and the available technology allows it.
Requirements
To gain first-hand experience of the requirements in terms of infrastructure and the potential of multimedia technology, it was decided to carry out a small study and prototype development effort focussing on multimedia data distribution over local and geographical networks. The contract to perform the study was awarded to Cap Gemini (I), following the Agency's usual competitive call-for-tender procedure.
The requirements that were defined for the prototype system implementation include:
A more general overall requirement for the implementation was to maintain openness to the integration of new products, avoiding ad hoc and proprietary solutions.
Approach
In the first phase of the activity, a fairly detailed state-of-the-art analysis was performed, supported by some early breadboard implementations, to define the best technical approach for the project. Given the requirements fixed by ESRIN, this had to comply with the general guidelines to:
Given the rapid evolution in this general area of technology, it was decided to follow a "rapid application development" approach that would allow the system's design to exploit emerging products, and this contributed greatly to the eventual successful outcome of the overall activity.
The application chosen for the demonstration of the system's capabilities was a general presentation of ESA activities, based on existing public-relations material for various missions. The material was chosen on the grounds of its ready availability, and also because the information contained in it is sufficiently interrelated to allow meaningful navigation, thereby demonstrating the interactive capabilities of the underlying system.
Technical solution
The prototype system is completely integrated with the WWW technology, and the client application is indeed a normal Web browser to which specific "plug-ins" (functional modules that interface in a predefined way with the core software) have been added. This approach, which required a change to the initial direction of the work, was chosen after an analysis of the state-of-the-art in order to take advantage of the Web phenomenon (then in its infancy), and to guarantee easier future integration with other ESRIN Web-based applications used for the various data-distribution services.
Two classes of service are foreseen, which reflect the two situations of a user connected to the server via the public Internet (typically via a 28.8 kbit/s dial-up connection), or via Intranet (normally a Local Area Network, or LAN). In the first case, the use of video material is precluded by the bandwidth limitations of the connection. Consequently, the system only provides high-quality audio, which can be integrated and synchronised with text, still images, and animations. In the second case, the system also supports high-quality video, whilst the ability to integrate and synchronise it with other types of material remains unchanged.
The architecture of the prototype consists of a server acting both as a Web and multimedia server, connected with the local client workstations via a switched Ethernet LAN, and accessible from remote stations via the public Internet.
Figure 1. General architecture
It is also possible to include a live television broadcast in the multimedia material accessible on the server. While this option is not one of the deliverables of the project, it has been successfully demonstrated via the integration of a real-time MPEG video compression card. The authoring of the application (graphics, animations, audio and video editing) is performed on a PC-based station, equipped with an MPEG video compression card (not real time, to allow better quality at affordable cost), and commercial authoring software (Adobe Premiere for the video, and Macromedia Director for the animations).
MPEG was chosen as the format for the video data as part of the general policy to favour public standards over proprietary ones in the hope of maintaining openness towards new developments. The term "hope" is used because, in a domain as active and as broad as Internet, one cannot exclude the possibility of a proprietary standard becoming the "de facto" choice at some stage. MPEG 1 was chosen in preference to MPEG 2 because of the latter's greater bandwidth requirement and the definitively higher cost of the hardware and software required for authoring and displaying MPEG 2 files.
The MPEG 1 compression allows the visualisation of full-motion (30 frame/s) video with quarter-screen format and very good resolution. A full-screen display is usually still fairly good, but is more sensitive to the quality of both the source material and the compression process. In the prototype system as implemented, it is foreseen to equip the client workstations with a hardware decoding card, which allows the quality to be maintained even with less powerful PCs. In the near future, the need for a dedicated decompression card will disappear, as the newer generations of processors will provide good software decompression.
The commercial product chosen for the audio streaming function is RealAudio by Progressive Networks.Two complementary approaches have been followed for the distribution of video data:
User interaction
The functionality supporting the interaction with the user is very high compared with that for the customary Web applications. In addition to the usual browsing and navigation functions, it is possible to:
The plug-in developed to carry out these functions is based on Java and Javascript for the management of the on-screen controls and the interactivity with the user, and on the MCI standard interface (Multimedia Common Interface) for the integration of the multimedia drivers.
One interesting feature of the system, which is a direct result of the design choices that were made, is the ability to provide the same navigation paths for both classes of service foreseen. In the case of access via a low-bandwidth channel, the video material is simply replaced with a selection of stills extracted from the video itself, which remains synchronised with the original audio stream in unaltered quality.
Figure 2. Multimedia-capable Web client
In addition to providing an interesting technical solution to the problem of the development of interactive applications for the distribution of multimedia data over the existing telecommunications network infrastructure, the prototype system's capabilities suggest some very interesting potential applications for this multimedia technology integrated with Internet, such as teaching, self-paced training, entertainment, etc.
As the system has been designed from the outset based on the most widely used standards and using commercial components, it is "open" to the integration of future new technologies and products. One possible further development could be integration with the emerging services for data distribution via satellite (e.g. Direct PCTM). These systems promise to overcome the bandwidth limitations of the current Internet by exploiting the technology and infrastructure that is being put in place to support direct satellite broadcasting of digital television.
ESA Bulletin Nr. 91.