October 1999
MOVE OVER PBXS
BY BRIAN STRACHMAN
During the past six months, the PC-PBX has achieved more industry and media attention
than any other CTI product in recent years, and rightfully so. The PC-PBX epitomizes all
of the good things about computer-telephony integration: lower costs, sophisticated
applications, and in many cases an orderly integration with existing data networks.
Perhaps a clarification of terminology might be in order, as many vendors of these types
of products don't refer to them as PC-PBXs, but as communications servers, IP-PBXs, or
integrated communications platforms. To make life a bit simpler and to avoid any nasty
phone calls from angry vendors claiming some sort of product positioning faux pas, the
all-encompassing and completely contrived term "Next Generation Communications
Platforms" (NGCP) will be used for the duration of this article to describe any
product that controls office switching but does not fall under the umbrella of a PBX or
key system.
That having been said, for all of the attention it has received the NGCP market is
actually quite small, totaling around $100 million in end-user revenues in 1998. This is a
drop in the bucket of the PBX industry, which amounted to over 7 billion dollars last year
in U.S. sales and service revenues. Up until just a few months ago, the NGCP market was
basically a three horse race running neck and neck between Artisoft, AltiGen, and Picazo.
However, things are just now starting to get exciting, because by the end of 1999 there
will be more than 25 vendors actively shipping products, including some familiar names
such as Lucent and Nortel. Simply by virtue of the number of new entrants, this market
deserves notice.
Despite the obvious industry optimism, there is a single reason that the PBX market is
70 times bigger than the NGCP market, and that is reliability. NGCPs are plagued by the
FUD factor (fear, uncertainty, and doubt) inherent in new technologies. Similar to the
phrase "No one ever got fired for buying IBM" is the idea that "No one ever
got fired for buying a PBX." The NGCP industry has yet to prove to skeptics that it
is possible to achieve a five-nines level of reliability in a Windows NT, PC-based
product. To combat this perception, many of these forward-thinking vendors have created
some very innovative solutions to the issue of PC uptime:
- Use industrial computers with Windows NT: Provided that no extraneous applications are
loaded, fault-tolerant PCs have proven to be highly reliable.
- Build the switching into the boards: By taking the operating system out of the equation
and building the basic call functionality into the boards, a high level of uptime can be
achieved.
- 98% + 98% = 99.999%: Using redundant Windows NT servers running at a very conservative
98 percent uptime, the five-nines of reliability can be met.
- Use a distributed architecture:
A proprietary PBX operating system is used for the switching and Windows NT is used for
the applications. While both OSs are technically within the same chassis, the PC is
removed as a dial-tone point of failure.
- A more reliable OS: The simplest way to alleviate concerns about Windows NT is not to
use it. Several vendors are utilizing alternatives such as DOS, embedded systems such as
VxWorks, and some are even considering Linux.
Over the next several years, the market will likely adopt an attitude of acceptance
towards NGCPs, and the FUD factor will become less pervasive. The demographics of the NGCP
buyer can be easily segmented into two categories: the small office with fewer than 100
extensions, and the large office with greater than 100 extensions. Currently the market is
small and will remain so, far into the next year or two, because most NGCPs are sold to
the small office where the primary motivator is price. They compete most directly with key
systems and Centrex, while only occasionally going up against a low-end Lucent or Nortel
switch and usually only winning the sale on the basis of cost. Looking solely at the small
office market, the NGCP will achieve moderate growth at best.
However, the greatest potential lies within the seven billion-dollar traditional PBX
market. Today, very few companies considering the purchase of a PBX are potential
customers for a NGCP vendor. This situation will slowly change, and with the entrance of
such notables as Lucent, Nortel, Cisco, and HP into the market, there will be a
significant rise in perceived legitimacy of NGCPs in the eyes of switch buyers. Given a
few years of reliable operation and huge advertising dollars, the NGCP will begin to
seriously cannibalize PBX sales.
Cahners In-Stat Group has forecast that the Next Generation Communications Platform
market will exceed five billion dollars in U.S. end-user revenue by 2003. This is not to
say that the PBX industry is dead, but a slow decline is certainly in the cards. Using a
familiar computer analogy, the PBX is like a mainframe whereas the NGCP is much like the
PC. Closed, proprietary, black monoliths reminiscent of the movie 2001: A Space Odyssey,
PBXs typically are produced entirely by a single vendor from operating system to hardware
to applications. NGCPs are much more open to third-party applications, and in many cases
use voice boards, PC chassis, operating system, and applications from entirely different
sources.
In the final analysis, the ultimate advantage of NGCPs will be their ease of use and
myriad of applications: traits only found in an open development environment. The industry
can look forward to high growth rates as well as technical innovation and the Next
Generation Communications Platform will be the driving force.
Brian Strachman is an Industry Analyst for the Voice and Data Communications
Services at Cahners In-Stat Group covering computer-telephony integration (CTI),
interactive voice response (IVR), call centers, and voice messaging. Cahners In-Stat Group
is a high-technology market research firm covering the consumer and convergence,
networking, wireless, telecommunications, Internet, and semiconductor markets. For more
information, please visit their Web site at www.instat.com.
|