Enough
with the past. On to the future. We've assembled some of the most
reliable, grizzled (well, a couple of them) and worldly wise pundits
to look through our "seer specs" and let us know what the future holds
in areas from the Internet to telecommunications.
Network
industry pundit Frank Dzubeck is ultrahigh on ultrawideband wireless.
The president of Communications Network Architects in Washington,
D.C., predicts new wireless technologies will finally make it economically
feasible for companies to snip their LAN cables for good.
"You will start to see the demise of wireline LANs - in most premises
for small and midsize businesses,'' Dzubeck says. "The reason is some
of the new local-oriented wireless technologies, especially ultrawideband,
will meet the 100M bit/sec needs of the future.''
Dzubeck says end users are going to demand 100M bit/sec Internet access
at the desktop, and 80% of corporate desktops ultimately will get
that access via wireless. Ultrawideband wireless is scheduled for
FCC approval this spring, and Dzubeck predicts it will become widely
used in office environments by 2003.
"Ultrawideband may destroy technologies like Bluetooth [a competing
radio technology that enables ad hoc networking between portable devices]
because it is more economically feasible,'' he says. "I think we'll
see it first in branch offices and remote offices, but over the horizon
it'll be ready for enterprises.''
Dzubeck is also gung-ho about wireless in WANs, with fixed wireless
services providing broadband speeds to office buildings and small
to midsize businesses. The largest corporations will opt for point-to-point
wireless in locations where it is difficult to lay fiber optics, he
says.
One
wireless initiative that Dzubeck is down on is 3G, the 2M bit/sec
wireless data access initiative launched in Europe. "We don't see
a huge explosion of 3G services,'' he says. "We think we'll see a
refocusing of the industry around [5-10M bit/sec] because 3G doesn't
offer enough new capabilities for the upgrades to occur. And you'd
have to throw everything out when 4G comes along.''
In addition to wireless, the other trend Dzubeck is excited about
is VPNs. He predicts VPNs will deliver more WAN services and on a
much greater scale to companies than they do today. That means service
providers will need to build IP-based networks on which they can deliver
high-quality, guaranteed services.
"By 2003-2004, we're going to start to get a granularity of network
services based on the transaction, with every transaction having its
own quality of service,'' Dzubeck says. "If I'm doing an XML/EDI [electronic
data interchange] transmission to a bank for payment, I may need an
entirely different quality of service than if I'm accessing the Internet
for a database lookup.''
Network
services, he says, will become more intelligent and will be outsourced
to providers that can offer a higher grade of service with less
management required. These providers will bill customers based on
usage, not on a flat rate. Dzubeck prefers a subscription service,
in which customers purchase network services on the level of bandwidth
and quality of service required. Corporations will manage these
types of services, but let carriers implement and deliver them.
During the next three to five years all service providers will
scramble to recover revenues they thought they would get from e-business.
With
traditional retailers finding it less urgent to become e-tailers,
providers that built networks to support online commerce must find
some other way to make money off these new networks, says Berge Ayvazian,
president of The Yankee Group.
"Service providers had been gearing their networks for a very high
growth rate of IP and commerce-related services. How is that going
to be paid for? If the Internet market isn't growing at the rate that
was predicted, that's a big problem," Ayvazian says.
Providers will try to sell e-commerce services to smaller businesses
than they had intended in an effort to make money off the infrastructure,
but they will also branch out into other bandwidth-intensive, IP-based
offerings such as application hosting. "Expect things that build off
hosting as a primary area of innovation. Hosted apps. Hosted enhanced
services. Things that take advantage of the infrastructure but go
well beyond it to take the requirements off customers to develop their
own applications," he says.
If an enterprise customer needs a customized SAP platform, it might
do well to hit up a service provider to write the application for
them. Carriers are hungry enough to work at bargain prices. "Things
would cost an arm and a leg that when there was a very tight market
could be acquired from a third party much more cost-effectively right
now," Ayvazian says.
As they scratch and claw for customer dollars, service providers will
cease network construction in an effort to cut expenses and shore
up the bottom line, Ayvazian says. "The build-it-and-they-will-come
era of the '90s is clearly over. Someone who wants to be a national
provider will have to use facilities built by somebody else and do
some horse trading," he says.
That
includes the major carriers such as the regional Bell operating companies
and AT&T, as well as newcomers such as Global Crossing. And some experienced
carriers formerly insignificant in the U.S. will become major players.
"The Deutsche Telekoms, the France Telecoms, the NTTs that by now
have been able to acquire substantial assets domestically, they're
going to have to be viewed by customers as part of the domestic market.
A lot of consolidation is going to come from these guys," Ayvazian
says.
All service providers, not just the new ones, face the same tough
environment. "What's odd about this time is there isn't any David
and Goliath story here. All companies are being faced with the same
problems. It's not like somebody's being toppled. Whether you're an
insurgent or an incumbent, you are facing the same challenges right
now."
Software: Adrift in a sea of services
By John Cox
Software is losing its moorings, according to Judith Hurwitz,
president of Hurwitz Group and a 20-year veteran of software trends
and fads.
The
traditional anchors of packaged software and application development
have been client/server, Microsoft operating systems and the PC. Now
it's about Internet technologies, such as Web protocols (HTTP and
XML), servers with browser or other so-called thin clients, handhelds,
and reusable software components based on the Java or Microsoft's
.Net component models and application frameworks.
"There's not a center of gravity for software right now," Hurwitz
says. "Microsoft has taken its hits in the market and no one else
has stepped up to the plate and said, 'Follow us, boys.'"
The 'Net and the Web are shifting the enterprise focus subtly but
importantly from the operating system to the network. Software vendors
and enterprise developers are moving applications to the Web. So-called
open source software, such as the Linux operating system and Apache
Web server, are giving some enterprise customers greater independence
from software vendors.
Over the next few years, Hurwitz says, companies will use these Internet
technologies and standards to make their software infrastructures
more modular and flexible. Early Web and e-commerce projects by enterprise
IT groups often were thrown together quickly. "They built stuff that
didn't scale," she says.
"We have to get to the point of creating an 'information utility',"
she says. "You can see this [direction] reflected in Microsoft's .Net
and in Java, where you have a common set of elements that form an
[underlying] infrastructure, and you plug into it via APIs. This approach
will let you buy 'best-of-breed' applications and have them work together
without crashing into each other."
To
this end, Sun and Microsoft will pick up the pace in incorporating
new Internet standards into their software platforms for smoother
interoperability and integration. XML will quickly solidify its role
as the common language for representing and sharing data over the
Web. Another emerging industry standard is Simple Object Access Protocol,
used to activate remote software components such as Java or ActiveX.
An alternative mechanism, called ebXML, is being positioned for complex
e-business transactions. In the future, a standard called Universal
Description, Discover and Integration, akin to a Web directory service,
will let applications search, find and interact with each other.
But even as companies create this information utility, Hurwitz predicts
it will be a transition point as a new breed of hosting companies
use the same Internet technologies to create services that can take
over big chunks of enterprise computing. Companies, in effect, will
rent software or net services they formerly bought or built themselves.
"Within five years, enterprise users will not build these services
themselves," she says. "Service providers will come into being to
create all this, and then sell metered services to large and small
customers."
E-commerce: Directories, security,
XML at the core
By John Fontana
If Jamie Lewis were a player for his beloved Atlanta Braves, he'd
have a solid bat and a steady glove built on his devotion to the
mechanics of the game.
But
because he's CEO of The Burton Group research and consulting firm,
his professional focus isn't on balls and bats. But his devotion to
mechanics remains. During the next five years, Lewis says, key network
mechanics will include directories, security and XML. They will be
woven into a general purpose IT infrastructure stretched seamlessly
across corporate boundaries to support user management and security
access controls beyond corporate firewalls. The weave will form the
fabric of e-commerce.
By 2006, IT will build this infrastructure.
The next three years will be a landscape of fits and starts as behind-the-scenes
standards and interoperability mechanisms are hammered out. Then,
in 2004 and beyond, tangible products will start to emerge.
Directories, as Lewis has preached for years, are central to creating
provisioning systems for adding, deleting and authenticating users.
"The huge leap is to the next level outside the firewall - where you
have your own internal systems, those of your partners, suppliers,
customers and service providers," Lewis says. "It has to have an integrated,
seamless and secure provisioning environment that we don't have today."
The other question is how do you securely share user rights across
interconnected systems?
"There is a market pressure now to develop a scalable and more manageable
security environment to enable this type of provisioning system,"
Lewis says.
One example may be the recently proposed Security Services Markup
Language, which lets user credential and access rights be shared across
sites regardless of the existing security infrastructures.
The merging of directories and securities will ignite the creation
of business Webs, groups of companies linked through a network to
provide a common function; and business-to-business relationships
and marketplaces held together by an all-purpose communications highway
of XML middleware.
Without this system, Lewis says e-business will have a hard time being
scalable and manageable. Other linchpins also are needed.
"When
you add users, actions have to occur based on that," Lewis says. "That
is where workflow and the message-based middleware kicks in and things
start to happen. An account gets created, your billing system kicks
in, rights are granted or revoked." XML will form the standard communication
layer between applications and services, and between services and
between applications. The directory will have to be able to communicate
in XML.
"In
the short term, we will see lots of struggles over XML standardization,"
Lewis says. But in four to five years, progress toward the definition
of baseline fundamentals should emerge.
Fiber from the desktop to the WAN; server-less companies; minimal
IT departments; rented applications on demand - this is the future
for many enterprise network infrastructures, according to Tere Bracco,
director of e-business infrastructure for research firm Current
Analysis.
"I think the most important thing happening in the near future is
the emergence of the all-optical enterprise," Bracco says. Category
5 copper cabling represents close to 80% of installed network cable
in companies, and network vendors are finding successful business
in pushing Gigabit Ethernet over copper. But Bracco says the trend
in new cabling installations is leaning toward fiber, and she expects
it to soon become as commonplace as copper. Fiber-optic network cable
"will become the Category 5 cabling of the future," she says.
This is no great revolution, Bracco says. But the results of having
fiber from the desktop to the LAN backbone and even to the WAN (thanks
to optical metropolitan-area service providers) will let companies
take on rich, high-bandwidth services that will make the IT department
of tomorrow look different than what many organizations have now.
With huge data pipes not only running into buildings but to individual
desktops, running videoconferencing and hosted applications will become
as commonplace as having an e-mail client, a Web browser and an office
productivity suite installed on a client PC. In addition to hosting
client programs, enterprise database applications, which once required
on-site, back-end servers, will be accessed from off-site hosting
facilities that could be across town or in the next state.
With fiber everywhere and applications being pumped in and managed
from the outside, Bracco foresees day-to-day LAN upkeep tasks being
relegated to less-technical employees. Instead of dealing with application
management and configuration problems, employees will order applications
when they need them and call the service provider when problems arise.
"With bandwidth on demand and applications on demand, businesses will
be able to call their service providers and say, 'We need PeopleSoft
during these hours and this much bandwidth during these hours,'" she
says.
What will transport these applications and on-demand bandwidth, Bracco
says, will still be Ethernet - Gigabit over copper at first, then
fiber running at 10G bit/sec over fiber.
"We're going to see Gigabit everywhere before we see optical everywhere,"
Bracco says.
Related
links