Full text of DOI:10.1109/101.876904

Technical
COMMUNICATION
When Users Drive the Network
By Hubert Whyte
Equally important, new service-cre-
Editor’s Note: These are interesting times for networking and telecommunications.
In recent years, the focus has shifted from connectivity to bandwidth. Soon the focus
will shift from bandwidth to services. Bert Whyte, the president and CEO of net.com,
is a leading advocate of open systems for service creation. Whyte has more than 30
years of expertise in the telecommunications industry, much of it in challenging and
often high-risk roles in which his entrepreneurial spirit has been called upon to dra-
matically alter the strategies and the fortunes of the companies he has led. Here he
gives an excellent overview of what to expect in the years ahead. These changes will
have profound effects on the way we communicate. – Harry Schmitz
ation platforms that reside between the
access and core now exist. These open
platforms allow service providers to dis-
criminate by user and by application at the
network edge; i.e., the point where the de-
mand for bandwidth meets the core of the
network. The ability to give every user his
or her own class of service, which can per-
haps even vary with the application that is
requested, provides a powerful means to
he concept of users driving lo-
cal-area networks is well under-
stood, but the idea of users driving
Ask any manager of information sys- manage how information flows through
tems, or LAN manager, if he or she is in the WAN.
T
control of the LAN. The most likely re-
The concept of a user-driven network
sponse will be that they are “ahead of the was demonstrated by net.com at
curve” in supporting the huge growth in Supercomm 2000. Users logged on and
LAN traffic. The motto of the LAN environ- were presented with a portal that detailed
ment is “bigger and faster,” expanding the services available to the user based on
from 10 Mb Ethernet to 100, 1000, and the user’s personal profile. Additional ser-
now G1000 LANs; from bridging to rout- vices were available upon request from the
ing to switching and now a combination. portal. Furthermore, new services could
It is a market that feeds on itself, which is be added by the service or content pro-
why you have so many happy enterprise vider on demand.
the wide-area network is a new concept. As
next-generation broadband networks are
developed, user-driven growth of WANs
will become a reality.
Enterprise networks were not always
driven by the next application that ap-
peared on desktops. Twenty years ago,
when mainframe-centric computing was
the rule, the LAN model was similar to to-
day’s WAN. Every computing task de-
pended on the mainframe. In that era,
computing was determined by the type of
mainframe, its power, and the application
that it could support.
LAN vendors.
This powerful capability allows the
user to dictate the applications and ser-
vices that they want to access. They can
determine dynamically how much the
service is costing. The demonstration in-
cluded billing for user, service provider,
and content provider, supporting the old
adage that a service is not a service until
you can bill for it.
The prospect of users driving the WAN is
something that faces significant hurdles
that are not technology related. Resis-
tance will come from the owners of the
WAN, most notably from the large incum-
bent carriers.
Thankfully, the PC and, more impor-
tantly, PC applications put the control of
computing in the hands of users. The PC
revolutionized network computing, but
not without challenges. The LAN environ-
ment always faced and continues to face
pressure to keep ahead of PC usage, PC
performance, and the vast array of phe-
nomenal applications that continually be-
come available.
Insatiable user demand is driving new
technology, including high-speed access
by cable, copper, and wireless. Optical
technology presently being deployed in
the core will hopefully keep up with cur-
rent demand, and another generation of
optical technology and optical switching
is in the labs to keep ahead of the demand
for the foreseeable future.
A wide range of services and applica-
tions that could be driven by the user is
just emerging. Soon it will be possible to
Harry Schmitz, Editor
■
CIRCUITS & DEVICES SEPTEMBER 2000
35 ■

dynamically request firewall services or carriers already understand the dynamics monopolistic control is less likely.
different levels of security based on the of this explosion in demand, as evidenced
network connection. Moreover, the way by their ability to characterize demand
an application enters or exits the network and the huge investments that they are
can be selected in the same manner that making in their core networks. However,
different levels of service can be requested the unpredictability will not be in the
The critical juncture that major carriers
are facing is now understandable. Based
on the opening up of the local market in
the US, it is becoming more and more dif-
ficult to delay the deployment of competi-
tive access technology. An alternative is to
provision a commodity-like access service
and expand through the acquisition of
other local players, expanding into the
long-distance market and reinforcing
their position in wireless as well. These
choices have merits, but what major carri-
ers really need to do is focus on creating
demand in their networks versus competi-
tive networks by providing enhanced ser-
vices and new applications. They can
remain competitive only by deploying an
infrastructure technology that supports
an open service creation capability, allow-
ing users to drive the network.
from Federal Express.
core, but where the core touches the ac-
cess network, namely the edge of the net-
work. This edge is where most bandwidth
decisions will be made, including where
and how users enter and exit networks.
The core itself is a reservoir of bandwidth
while the real network intelligence lies at
its edge.
Previously, it was easy for the local in-
cumbent, as the only supplier of access
technology, to control the services avail-
able to the user. Unfortunately, in parts of
the world where deregulation has not yet
occurred, this situation will continue.
However, in the new telecommunications
environment, there is new competition
from wireless and cable companies as well
as new service providers who now have ac-
cess to the copper plant, so the possibly for
As more applications become available
across the network rather than residing
on the desktop, service differentiation be-
comes paramount. Indeed, a vast range of
applications is now beginning to appear.
The ability to discriminate among individ-
uals and applications will drive a new gen-
eration of technology to be deployed
within networks. Today’s WAN is more or
less a commodity; i.e., a supply of informa-
tion pipelines. The next-generation net-
work will have to reach a new level of
sophistication.
Today’s WANs are essentially planned
environments that manage demand ac-
cording to the constraints of the network
infrastructure and the tariffing structure.
In many respects, these networks were
simple to plan because they could be
mapped to areas of greatest demand and
between areas of the most communica-
tion. They are similar in many ways to the
mainframe model. However, these net-
works are now becoming distorted as a re-
sult of the PC and the Internet. Now it is
difficult to predict where demand will oc-
cur in the WAN environment, so the natu-
ral tendency is to overprovision the
network.
Hubert Whyte is president and CEO of
net.com, located in Fremont, California.
CD■
Remarkably, overprovisioning is exactly
what is happening in North America and
Europe with respect to the core and in
some metropolitan WANs. Investment by
next-generation carriers and someincum-
bents is at an all-time high as they pave
their way to accommodate the increased
demand that they hope will materialize.
Much of this investment is specifically in
optical technology, which will put in place
an unlimited amount of bandwidth, form-
ing the basic plumbing of the network.
Of course, moving from a predictable
environment where demand is controlled
to an unpredictable one with little con-
trol will be a challenge for the incumbent
carriers, particularly those where the fo-
cus has traditionally been the local-ac-
cess network. Traditional long-distance
■
36 ■
CIRCUITS & DEVICES SEPTEMBER 2000