Open Source business -
Last week I promised that there was a huge market for commercially
minded Open Source developers. This week I want to talk about it. And
for those who think that Open Source should mean "free," I've attached
some non-profit centers to look at.
Over the last couple of weeks, an odd kind of pessimism has continued
to grip the Open Source marketplace. At the heart of that pessimism is
paradox. The overall Open Source and Linux markets continue to grow
like crazy. Many firms involved with Open Source continue to rack
growth rates of 200 to 300%. But skeptics point out that the
best-known commercial vendors of Linux-based products are having a
time trying to generate anything that looks like a profit.
The problem is that pessimists are looking in the wrong places.
As I mentioned last week, if they looked in East Asia, they'd see a
Linux market with the potential to displace Microsoft as the operating system vendor
of choice. If they wanted to see Open Source companies with solid
lines, they could look at contract development houses that specialize
Linux -- or "rent-a-geek" I.T. consultancies that have embraced the Open
Source movement. They're doing just fine. In addition, Linux is
positively contributing to the bottom lines of firms who manufacture their own hardware. Makers of servers, routers, PDAs and desktop
can afford to help support Open Source development and write-off the
costs as a promotional expense. Over the long term, contributing to a
robust Linux environment is probably cheaper than putting up with
Microsoft. Finally, there are clusters of niche applications, notably
the next generation of electronic cash registers, where Linux could
become the de facto standard.
But there's an even bigger market out there-- a market that will
grow to an estimated $700 billion over the next couple of years, and that's
just in the United States alone, if Yankee Group and IDC estimates are to be
believed. That market is in the midst of one of the most fundamental
technical changes it has experienced in its history, and Open Source
will inevitably play a major part in its future.
A couple of weeks back I slapped wireless digital "WAP" data
telephony as immature, if not an inherently dumb idea, and a risky
business for the Open Source developers who were focused there. But the larger voice/data telecom market is another story entirely.
No less a personage than Intel's Craig Barret has been
telling the world that telecom is the future, and that future
requires an Internet based telephony based on "open standards," if not
fully Open Source code. The days of $100,000 proprietary telecom
switches is long over. So are voice-only networks. The future belongs
voice over Internet (VoIP) systems. And the glue that's going to hold
them together "almost has to be open source- based."
There are a number of reasons for this. The first is that the old
cozy telecom market has finally broken apart. Wireless telecom vendors,
particularly those outside the United States, have introduced real price
competition, and are generally more convenient than their old wireline
competitors. In the States, flat-rate Internet has had a similar impact
on the traditional metered long distance market, and the Web comes with
its vast array of expanded services. As a consequence, there is
tremendous price pressure on major carriers to cut costs, exponentially
increase bandwidth and add service options.
It can't be done on old legacy telecom platforms. The cost of the
NT-1 and NT-2 compliant big iron telecom switches is far too high, and
the technology is far too primitive; these old dinosaurs cannot provide
the variety of capabilities
(interactive/text/voice/video/streaming/duplexed/multicast) needed by
modern communications carriers- at least not at a reasonable price. The
legacy communication protocols used by the carriers (ISDN, etc.) also
presumed that voice, not data, would be the guts of telecommunication
traffic; as soon as data began to overwhelm voice on long distance
lines, the limitations of these old telecom protocols began to become
Increasingly, the market has begun to embrace a new view of the
telecommunications future -- one that sees voice as a small eddy in a
massive data stream. The future is voice over packetized data network
protocols; voice over DSL and voice over IP (VoIP). The future is
already taking hold in telecom backbones, and is spreading closer to
neighborhood switch. Not long ago, voice over IP meant poor voice
quality. Today that quality is so high that switch vendors have had to
introduce "comfort" noise onto their lines to assure customers that
they're using a telephone!
Traditional vendors of voice-centric providers of time-sampled
"circuit switch" technologies like Lucent Technologies are beginning to
get creamed in the market by packet telephony pioneers like San Jose,
Calif.-based Sonus Networks. Only the dead weight of the massive
investments made in these old systems have kept carriers from
them altogether. In some places, they are abandoning these old systems
now. Cities like Dallas have begun to tear out their old
telecom infrastructure and replace it with VoIP systems.
Carriers in trouble, most notably AT&T, may not be able to delay
the adoption of IP across their networks. AT&T's Jim Daugherty
"We're not buying any more Class 4 toll switches after this year. We're
building IP backbones as large as anyone else's, and we'll arrive at
packet-based telephony as quickly as anyone else." Business class IP
telephony is on the horizon, as are new, complex voice/data services
that take the best of voice and Internet services and mix them together
in new and creative ways.
The market is still in its infancy. VoIP services currently generate
about $100 million in annual revenue in the United States last year, but a
analysis by Frost and Sullivan project that market grow to $2 billion
the end of next year. After that, who knows....
The reason that any of this matters to the Open Source and Linux
communities is that this future network almost has to be both open, and
Unix-, if not Linux-based.
It is critical to telecommunications providers that they be able to
inter-operate their set of services with other operators around the
world. In the old days of proprietary systems, adding even the most
basic services to telecom systems took forever. Services like call
waiting and call forwarding took years to implement as
carriers held onto proprietary technologies and fought over standards.
In today's Internet-meets-voice environment, there is no time for such
The core telecom protocol for Voice over IP, a standard called ITU
H.323 has been opened up. Open H323 is a standard with the ability to
handle complex for audio, video, and data communications across
networks, including the Internet. A new class of "telephony application
service providers" like Pagoo is embracing
Open H323 as the fastest way to develop new applications that can be
used across the global voice and data network. The v.p. of product
development at Pagoo, Phil Piernot, said recently that the company's
"Voice Express" system would embed open source into its architecture.
Computer-Telephony pioneer Dialogic, now part of Intel, is
looking to the Open Source community to create an array of secure IP
systems as well as voice-based Internet and e-commerce applications
voice portals, multimedia messaging and hybrid phone/web "contact
centers." Other telephony firms such as Voivida, and Quicknet are also
beating the bushes for Linux developers and partners.
And it's not just VoIP; firms like Cisco and Motorola are
the development of a whole new generation of computer telephony
solutions based on Open Source platforms. Open Source firms like Lutris
Technologies are taking Internet application server technology and
adapting it to the telecom business. Next generation (post WAP?)
from Motorola will depend in large part on server technologies
by Open Source vendors and Motorola has made a point of cultivating
platform system developers like Lutris
Even firms who aren't necessarily committed to Open Source are
flirting with it in order to drive demand for new computer-telephony
applications. A case in point is SpeechWorks International, which recently
announced a deal with VoIP developer iBasis. The two firms are
developing an architecture that could allow your local telecom to digitally
connect to voice -activated, automatic call centers hundreds or
thousands of miles away. Instead of making a long distance "toll free"
call that the merchant has to pay for, users could make a local phone
call to their ISP. The ISP could use VoIP via the 'Net to connect with
"world wide voice web" of services, that sound a lot like
a traditional (if somewhat smarter) call center. Via the Web, the
customer could then handle bill payment or pick up customized audio
services, all without anyone spending a dime in long distance charges.
That would work for wireline and wireless phone users alike. It
certainly seems like a simpler proposition for a mobile phone user than
squinting at a WAP display terminal the size of a postage stamp. The
two keys to driving mass acceptance? A superior "voice server" and an Open
Source VoiceXML-based audio browser.
Admittedly this looks a lot different than the vision of free and
open code envisioned by Open Source pioneers. But it works, and there's
a business here.
And for those that want to try their hand at building a telephony
platform that's open and free, check out Bayonne, the
telephony server of the GNU project. Source for the telephony drivers
can be found at the openh323.org. Other Open
telephony sites include linuxtelephony.org,
OpenTelecom. Brian Wiles at Speakfreely.org is
building an Internet Telephony device, and you might also want to check
out the Gphone project
headed up by Roland Dreier. Some of these efforts may not be as
as the systems being envisioned by commercial developers, but they'll
give you a chance to look at some of the technology, and you might save
a buck or two on long-distance charges.
NewsForge editors read and respond to comments posted on our discussion page.