IONA Technologies Ltd - Position Statement

Real-Time Multimedia and the Web - RTMW '96

Martin Chapman

Introduction

As part of ongoing product development, IONA Technologies are investigating a number of issues related to real-time, multi-media, and telecommunications. The challenge for IONA, and the OMG, is to embrace the existing work in these areas and incorporate them into the Object Management Architecture (OMA).

This summary document outlines a number of positions that predicate our work.

Position 1. Use of CORBA

The following is a (non-exhaustive) list of consortia active in interactive multi-media standards:

  • IETF - Internet Engineering Task Force
  • DAVIC - Digital Audio/Visual Council
  • TINA-C - Telecommunications Information Networking Architecture Consortium
  • MMF - Multi-Media Forum
  • IMA - Interactive Multi-Media Association.
  • With the exception of the IETF, all the above rely on the basic technology defined by the OMG , the smallest subset being use of IDL (Interface Definition Language) and IIOP (Internet Inter-ORB Protocol). The combined members of these consortia represent, almost completely, the entire interactive multi-media industry. As the WWW converges on this industry, it is our belief that, the next generation of WWW protocols and architectures should be based on this existing work, and hence CORBA should play a substantial role.

    The alternatives are:

    1. To invent a new set of protocols/architecture - do we need another one? We think not.
    2. Use COM/DCOM - great for clients, however, it is not based on widespread consensus, and we believe it does not scale.
    3. Use Java's RMI (Remote Method Invocation) - this is fine but only if everyone programs in Java.

    CORBA is suitable for all. It works on a range of platforms, comes in a variety of sizes (for set-top boxes to mainframes), it copes with legacy systems, and allows for evolution. CORBA has also demonstrated integration with other technologies, such as OLE/COM and Java, allowing for seamless cross technology development.

    Position 2. Audio/Video is not just retrieval

    Traditionally the WWW has been used to retrieve information. So far the extent has been retrieval of clips (either downloaded, stored locally and then replayed, or replayed over the network), or tuning into radio/tv 'broadcasts'. A fully interactive environment should also be envisioned whereby users can setup conferences, join exiting conferences, and invite others to participate in a conference. We include in this definition simple point-to-point activities such a 'phone' call, as well as multi-party activities involving any number of people. Furthermore, such an environment should be capable of dealing with people and/or terminals on the move. It should also be possible to easily combine such communication facilities with processing facilities, such as to enable multi-party gaming (i.e. streams and logic).

    This requires session control models and protocols, as well as stream control models and protocols.

    Position 3. Separate out the control from the stream flow

    The setup, release and modification of audio-video streams can be accomplished in a manner that is independent from the stream protocols. This separation is well supported in other fora - DAVIC's use of DSM-CC (Digital Storage Media Command and Control) is a prime example - and has been effectively demonstrated in interactive TV trails and systems. Good examples here are the BT trail in the UK, and the Hong Kong Telecom Interactive TV service due to go into operation in 1997.

    The WWW would benefit from a standard way to accomplish control and management of streams, while allowing any network, transport and encoding technology to be used. To this end, the OMG has recently issued an RFP on the control and management of audio/video streams (OMG, telecom/96-08-01).

    Position 4. QoS is not just a network issue

    A lot of work in the multi-media world focuses on the network aspects of QoS. While this is a major factor in the performance of the system, the computing systems and applications are additional factors. QoS must be considered from and end-to-end perspective i.e. from application to application. What is the point in throwing all your money at a network if, at the end-of the day, a message gets delayed in the application itself.

    In a similar manner, realtime must also be considered from an end-to end perspective.

    In achieving this, it is important that the right APIs are provided, and that there is full architectural support. Within the OMG, the Realtime SIG and the Telecoms Task Force are addressing these issues. IONA is active in both of these groups.

    Position 5. Diverse configurations

    There is, and there will continue to be, a diverse set of terminal configurations. This is particularly true when it comes to multi-media gadgets. A good example here is a video camera. A camera can be plugged into a board within a terminal, or it can be plug into a board directly connected to a network (e.g. ATM adaptation cards). In both cases, the camera is part of the (logical) terminal, however the setup of streams will require different mechanisms. Approaches to multi-media in the WWW must allow for such diverse configurations.

    In a similar vain, the display of video has two approaches. Either directly in the browser itself or via a plug-in/helper application. Both have their merits, and hence we believe a single approach should NOT be adopted.

    Summary

    Customers, developers and the industry in general would benefit from a convergence in protocols and architecture to build and deliver multi-media services. This area is relatively new to the WWW, and evolution of the WWW should take into account the trends in the multi-media, on-line industry. CORBA is already being used in live systems to deliver multi-media services, and provides the base set of interaction mechanisms for a number of large industry fora.

    IONA is constantly evaluating its product range for their suitable to the on-line industry, and to this end are investigating issues related to pluggable transports protocols, realtime, QoS, and audio/video streams. One of our philosophies is "one size does not fit all", and hence we will ensure CORBA can be tailored to different environments.

    IONA are active in the OMG's Realtime and Internet SIGs and in the work of the Telecommunications and Electronic Commerce Task Forces to ensure that real-time and multi-media topics are properly reflected in the OMA. IONA is also active in TINA-C, which are defining application frameworks for interactive multi-media services and are producing requirements for an underlying distributed platform.

    Contact

    Martin Chapman
    Standards Liaison Officer
    IONA Technologies Ltd.
    8-10 Lower Pembroke Street
    Dublin 2
    Ireland
    email: mchapman@iona.com
    tel: +353 1 662 5255
    fax: +353 1 662 5144
    www: http://www.iona.com