1. Web Conferencing, Instant Messaging & Streaming Media
Prepared for the International Communications Industries Association
To support Infocomm 2002
By Andy Nilssen & Marc Beattie
Senior Analysts and Partners, Wainhouse Research, Brookline, MA, USA
Marc F. Beattie has authored public and private reports on product strategies,
distribution structures, emerging technologies and industry applications. Marc is a co-
author in Wainhouse Research’s Conferencing Markets & Strategies series. He is the
principle instructor for the Implementing Streaming Media seminar. Prior to joining
Wainhouse Research Marc was an early team member at PictureTel and Polycom –
holding his last position at Polycom as Director of Business Development; he has spent
15 years working within the industry. Marc has been an independent analyst for the last
three years. Mr. Beattie attended Gordon College where he earned a Bachelor of Arts
degree in Economics.
Andrew H. Nilssen is a co-author of Wainhouse Research’s Conferencing Markets &
Strategies series, and consults for industry leaders, Fortune 500 companies, and venture
capitalists about the multimedia communications industry. Andy has over 20 years of
experience in bringing high-technology products to market. Prior to Wainhouse
Research, he was Director of Marketing, New Business Development at PictureTel, Inc.
where he defined strategies and identified partners to expand business by leveraging
broadband and streaming technologies. Andy earned his MBA and BSEE degrees from
the University of New Hampshire and holds two ease-of-use related patents.
Web conferencing, instant messaging (IM) of the content in a conference call, and the
and streaming media are three emerging new reality that there is access to a network-
“rich media” communication technologies connected web browser virtually everywhere
that have been made possible through the a telephone is located. Web conferencing
combination of computer-generated sells itself on providing a way to augment a
information and the widespread growth of typical audio conference with powerful
the Internet. All three technologies have graphical information exchange.
experienced rapid growth due to the impact
of their powerful yet inexpensive Most web conferencing systems use client-
capabilities. Each has also spawned server architecture to enable simultaneous
dedicated companies that offer the exchange of audio and graphical content. As
component technology in product form or shown in the figure above, a web conference
bundle the products with integrated services typically consists of two parallel
to present a turnkey rich-media conferencing conferences – an audio conference using
solution for the user. telephones on the PSTN overlaid with a
separate data conference using web
Web Conferencing browsers on a network for data. The degree
Web conferencing is built on the premise of audio integration in a web conference can
that graphical data represents a large portion vary greatly, ranging from requiring the user
to set up a totally separate audio call, to
Copyright Wainhouse Research 2002 page - 1

2. displaying the conference number to call in web conferencing systems will simply
to via the web browser, to automatically translate the PowerPoint slides to static
calling out to all participants (scheduled or bitmapped GIF or JPEG images, thereby
ad-hoc). loosing slide transitions and animations, and
also locking the slide resolution into
whatever translation is used. More
Audio Web sophisticated systems use either a
Conferencing Conferencing proprietary object descriptor language or
Server Server
DHTML to translate PowerPoint slides into
graphic objects, and to preserve high
resolution and possibly transitions and
PSTN
intranet or animations as well. Graphical object-based
Internet
systems also allow slide resolution to scale
to take full advantage of the resolution of an
individual participant’s browser, and can be
more bandwidth efficient. Application
sharing and “web tours” have similar
considerations – the basic systems just “peel
& send” the presenter’s screen, while more
sophisticated systems support true multi user
Figure 1 - Web Conference Architecture application sharing – some are even built on
In a web conference, a presenter can present top of the T.120 (specifically T.128)
PowerPoint slides, demonstrate software application sharing engine which is also
applications, use an interactive whiteboard, used in products like Microsoft NetMeeting.
and take polls to gauge the audience’s
reaction. In some specialized customer Web conferencing systems need to get along
support applications, the presenter can with firewalls – or more precisely, need to
actually take control of a viewer’s PC to get through firewalls reliably. If the web
help walk them through a problem. conference involves participants located on
different corporate LANs, and the web
Every web conferencing system requires the conferencing service is not designed to
use of a local software client on the understand a particular firewall, the
participant’s PC to add web conferencing audience behind that firewall will not be
features to a standard web browser. This able to participate. Penetrating firewalls is
usually takes the form of a Java applet, not a trivial issue. The more advanced web
which is downloaded “on the fly” and conferencing technologies will actually try
usually saved locally in the browser’s cache different mechanisms on the fly to figure out
so that it does not need to be downloaded which policy is optimal for getting through a
again for subsequent web conferences. The particular firewall. The web browsing
size, features, and efficiency of this local protocol, HTTP, uses port 80 – which is
Java client are different from one web probably left open by most firewall
conferencing service provider to another. administrators, but it is not the most
The method for capturing and transferring efficient. More efficient than HTTP is TCP,
the graphical image from the presenter to the which is also used for streaming media, file
participant’s web browser can vary. In most transfers, and terminal emulators; however,
web conferences, uploading a PowerPoint most administrators will use a firewall to
presentation to the web conferencing server block ports to unsolicited TCP traffic.
is the norm – though some systems (ex. Finally, IP ports, which act like a sub-
WebEx) actually process and distribute the address to an IP address, may also be
presentation from the client itself. Basic blocked or restricted by the firewall. If a
firewall restricts port 80, web browsing will
Copyright Wainhouse Research 2002 page - 2

3. be blocked. Because web conferencing and are quickly moving to augment
systems involve a local client, a good system basic text-based chat with file
will be able to try each successive method to transfer, data sharing, PC remote
get through the firewall in the preferred control, and even real-time IP-based
order – TCP, HTTP with dynamic ports, and voice and video.
HTTP via port 80 – in order to establish the
most efficient web conferencing connection As with other forms of conferencing, IM can
through a firewall. be implemented by purchasing an IM server
or by using an IM service. The big three
Two last considerations are the browser commercial IM service providers –
itself and standards. Web conferencing AOL/ICQ, Yahoo IM, and Microsoft MSN –
vendors spend considerable effort debugging are not compatible, meaning a user on one
their clients to run with each new revision service cannot communicate with a user on
from major browser vendors to be sure a another. “Commercial-grade” IM software
particular web conferencing service works server packages are available for
reliably with most popular browsers. For organizations and service providers to
most web conferencing implementations, the purchase to host their own IM sessions for
vendors’ attitude towards standards is often security or cost reasons (ex Jabber, Bantu,
“who cares?” since server communications Microsoft Exchange Server). The presence
are web conference specific and the clients features of IM are also beginning to merge
are lightweight on-the-fly Java downloads. with web conferencing in products such as
Most vendors believe there just isn’t too the Polycom WebOffice and Lotus
much investment or interoperability to be Sametime.
concerned with. However, this will start
becoming a little muddy as IP audio and The introduction of Microsoft Messenger for
video become part of a web conference and Windows XP marks a major move by
as users start using many different web Microsoft to expand their MSN IM service
conferencing services and multiple web into a full conferencing service. Messenger
conferencing Java clients accumulate on for XP is IM-based, but features integrated
user’s PCs. data conferencing (file transfer, whiteboard,
application and desktop sharing), and taunts
Instant Messaging users with “Add Voice” and “Add Camera”
Instant Messaging (IM) is becoming an icons. Wainhouse Research has found that
influential factor in the rich media Messenger for XP’s wide-band audio can be
conferencing services space for two reasons: excellent and the software-based video
compression using a USB camera is
1) IM systems uniquely use presence adequate for effective desktop
to display personal contacts along videoconferencing. Messenger for XP can
with their real-time availability rendezvous with presence information only
status. Thus the probability of when connected to either Microsoft’s .NET
successfully connecting with services (which requires each user to get a
someone is known before trying to Microsoft Passport account) or to an
make the actual connection. No enterprise-hosted Microsoft Exchange server
more telephone tag, no need for – so clearly Microsoft is playing on both the
voice mail – IM is bringing a client and the server side. The service also
welcome change to the age-old hit- requires the network to support a new
and-miss paradigm for calling network protocol called “Universal Plug-and
someone. Play”, which may take many years to
enable.
2) IM systems are increasingly
becoming much more feature rich,
Copyright Wainhouse Research 2002 page - 3

4. Streaming Media ! Content Management. This is also
Streaming media technology is a relatively known as digital asset management.
new vehicle for extending multimedia These products and services make
conferences to any user with a network- libraries of digital assets more
connected PC and web browser. valuable and more accessible. One
of the key elements of this market
Streaming technology involves creating, segment is the area of audio and
serving, and playing a mix of audio, video, video indexing, which enables
and/or presentation data over a network intelligent searching on streaming
without having to wait to receive the entire media files. Another major segment
multimedia file of the event. Because IP is represented by systems which
networks introduce variable delay, enable owners of digital content to
streaming requires the receiver to control access, manage libraries,
accumulate pieces of data off the network at track usage, and generate detailed
the top of a buffer while simultaneously reports and billing statements.
emptying the bottom of the buffer for real-
time playback of audio and video. Since ! Content Delivery Networks. If a
receiving data from the Internet in particular user is delivering streaming media
can be unpredictable, the entire streaming on their own LAN, chances are they
process can introduce a fair amount of delay, will have ample bandwidth and
easily measured in 10’s of seconds; thus control. But if the stream is to be
streaming technology is uniquely applicable delivered to far-flung locations, or
only to receiving one-way live or stored to sites not on the LAN, or if the
presentations, where real-time, two-way event is to be accessed by the public
interaction is not a priority. The audio and at large, then some type of delivery
video stream can also be accompanied by network or systems architecture
synchronized presentation data in the form should be used. A variety of content
of still images or Flash-encoded visuals. delivery networking schemes exist;
most use the public Internet for part
There are many different ways to look at the of their solution, some use satellites,
streaming industry. Wainhouse Research some use intelligently placed servers
has divided the industry into four segments, to move content closer to the end
each of which includes hardware, software, users, thereby minimizing the traffic
services and components, each of which that has to move over the Internet
represents a distinct application set. Most backbone.
vendors operate in just one of these
segments, although some of the vendors and ! Client-Server Products. There is no
service providers are expanding their recognized standard for streaming
offerings to cover multiple pieces of the media today. The major client-
puzzle. server choices today include
Apple’s Quicktime, Microsoft’s
! Content Creation. Products and Windows Media Technologies, and
services in this sector enable Real Network’s RealSystem. Some
customers to create content that is client software systems are designed
optimized for streaming media to be used as stand-alone player
delivery. We include here all the applications; others are designed to
steps taken to capture, edit, and reside inside a standard web
encode content for streaming browser. Some streaming products
delivery. use no client at all; instead a Java
applet is loaded automatically when
needed to play the streaming file.
Copyright Wainhouse Research 2002 page - 4

5. Other important elements in the The streaming hardware and software
client-server segment are the data servers available to streaming service
compression/decompression engines providers vary according to a wide variety
or codecs, which are usually of user needs within these two categories:
implemented in software though
some codecs use special hardware to ! Number of simultaneous streams
increase their speed. Some of the supported – this can range from
most popular streaming codecs are single digit numbers to several
Sorenson Broadcaster, RealVideo, thousand
MPEG-1, MPEG-2, MPEG-4, and ! Number of files available –
H.263. Most of the client-server primarily a function of storage space
products on the market support ! Bandwidth supported – most of the
multiple codecs. consumer/Internet users today are
connected at 28.8 kbps, but
corporate LAN users are in a
Content
Manage Delivery position to take advantage of the
Create Network much higher data rates possible.
& Server Client
Encode
Higher data rates translate to much
higher audio and video quality.
! Security features - authentication
features ensure site security or
enable pay per view for direct
revenue generation. A server
Figure 2 The Four Streaming Functional
interface may connect to Web-based
Blocks
backend credit card and transaction
processing systems.
Figure 2 illustrates how these four functional
! User Accounting – provides the
blocks fit together. All streaming
hosting service the ability to know
applications start with a content creation and
who watched what video when.
encoding process. Sometimes, the content is
! Interfaces to live video feeds such as
indexed, managed, or subject to access
news, weather, etc. Since these
control, although this is not always the case.
feeds are generally NTSC analog
All streaming files eventually are placed on
video, the server also may contain
a streaming server from where they are
an encoding solution which digitizes
delivered to the users over some type of IP
the video and then compresses it to
delivery network, corporate LAN, extranet,
the streaming format required. The
or Internet connection. The content is
encoding solution may be software
eventually viewed (or heard) with some type
residing on the host computer or it
of client software that must be compatible
may be a codec hardware accessory
with the protocols, file formats, and
on an add-in board. Industrial-
compression techniques that are supported
strength streaming applications
by the server.
often use hardware codecs which
enables a single PC to encode
Streaming media services fall into two broad
multiple streams simultaneously.
segments: consumer streaming services,
! Management systems that present to
which uses the Internet for delivery, and
users an organized list of what is
professional corporate streaming services,
available for viewing.
which uses both private intranets as well as
the Internet.
For conferencing applications, streaming
servers are specifically configured to work
with videoconferencing systems and real-
Copyright Wainhouse Research 2002 page - 5

6. time input. The first generation streaming training sessions, or recordings of live feeds
conferencing servers use the video and audio can be stored on a server, to be reviewed on
signals directly from the videoconferencing demand by users from any browser-enabled
system as the input to the streaming desktop. Only a single copy of the original
subsystem. Future conferencing servers are event is maintained on the server, which
likely to have unified user interfaces and simplifies updating, access management,
digital trans-coding to make the and distribution control.
conferencing to streaming connection
seamless. Match the Media to the Message
Streaming media clients, which are designed While many of these conferencing
to run as both browser plug-ins and stand- technologies are new and often more
alone applications, can be quite large and are exciting to work with than a standard audio
thus downloaded and installed on the call, Wainhouse Research encourages clients
receiver’s PC prior to a streaming media to evaluate the message they are trying relay
event. and apply the appropriate medium to match
the message.
Streaming video can span a wide range of
bit rates from 20 kbps up to MPEG1 Figure 3 presents two curves that one should
(typically 1.5 Mbps) and occasionally keep in mind while choosing the rich media
MPEG2 rates (more than 8 Mbps). conferencing vehicle for a particular event.
Determining the optimal bit rate requires a The first curve illustrates the trade-off
trade-off between the network bandwidth between Cost and “Connected-ness” – for
available versus the fidelity/quality needed. example, if cost is not an object, then why
Network bandwidth can be segmented into not use videoconferencing for everything?
three categories: Internet/modem speeds (< Videoconferencing provides maximum
56 kbps), ISDN speeds (< 256 kbps), and visual feedback with your audience, which
intranet speeds (> 256 kbps). is important if the meeting requires
persuasion, selling, or consensus building on
Audio- and video-on-demand technology a controversial topic. If this level of
enables service providers to complement connected-ness is not required, then a lower
real-time streaming offerings with cost conferencing vehicle could be used.
conference storage and retrieval. Training The second curve illustrates the trade-off
videos, briefings, audio conferences, between Interactivity and Reach, which
earnings calls, employee benefits programs, fortunately is consistent with a natural
Video Conferencing:
Meetings
“Connected-ness”
Web Conferencing or
Streaming:
Interactivity
Briefings
Audio Streaming:
IM Broadcasts
Cost Reach
Figure 3 Match the Media to the Message
Copyright Wainhouse Research 2002 page - 6

7. meeting dynamic: the more people involved audio can be delivered at low bandwidth
in an event (“reach”), the less interaction using standard streaming platforms such as
there tends to be. Hence an interactive Real Networks G2 or Windows Media
Meeting should use a real-time conferencing Technologies. Each individual stream will
technology, while an informational typically use less than 8kb/sec of network
broadcast can use streaming. A briefing bandwidth per user. Audio over 28k modem
could use either technology depending on connections was considered little more than
the degree of real-time interaction required a toy as few as 24 months ago. Today
to cover the content. effective streaming audio can co-exist with
graphical data updates and still be delivered
It may be that a simple audio conference very well over modem speed connections. If
achieves 90% of the intent. If the sharing the network supports multicasting, it can be
data is important many easy-to-use web used to minimize network utilization for the
conferencing services are available to add to one-way audio stream as well.
your audio calls. While streaming can
provide impact and reach, knowing what While each of these technologies has its
your target audience is capable of receiving independent capabilities, they can also be
- PC clients on a modem are not often good used together. By way of example, IM can
candidates for streaming - will determine if be used to determine the “presence” and
they ever get your message. If the message status of colleagues. If your IM client
involves critical persuasion, the added indicates their availability, you can request a
“connected-ness” benefit provided by a true web conference to share a document or
two-way videoconference may be most object on your PC. While engaged in that
effective. web conference participants can IM one
another on the subject being discussed while
If the web conference is a presentation not disturbing the conversation of the larger
where one site is doing the majority of the group – this would be similar to leaning over
talking, audio streaming of the presenter’s to someone in a conference room and having
voice is becoming a popular option. In this a sidebar conversation. If some team
case, audience feedback to the presenter and members were not available for the web
to other sites is limited to text chat, polling, conference the meeting could be captured –
and “mood” indicators – which, for both the audio and data - and archived for
structured presentations, can be quite viewing at a later time via streaming.
effective. Very high-quality human voice
Copyright Wainhouse Research 2002 page - 7