Is There Hope for VDSL?/ article
VDSL is fast and affordable, but widespread adoption isn’t a sure thing.
by Tom Nolle
Network Magazine
08/04/00, 6:08 p.m. ET
Most of the Digital Subscriber Line (DSL) debates focused on how
to achieve low-end-of-the-broadband performance over relatively
long copper loops. This resulted in an industry bias toward
Asymmetrical DSL (ADSL). However, a new RBOC project may
open the door for a faster, more exciting form of DSL known as Very
High Bit Rate DSL (VDSL).
VDSL can deliver up to 14.5Mbits/sec over loops just shy of a mile long (1.5 kilometers), and an
eye-popping 58Mbits/sec on loops less than 1,000 feet. The new RBOC plans will support
broadband access through fiber remotes rather than home-run copper loop, and it's likely that
Competitive Local Exchange Carriers (CLECs) will adopt similar fiber-based practices. Carrier
fiber remotes would terminate a copper loop at the remote concentrator, which is a box or hut
close to the customer, not back in the Central Office (CO).
GET SHORTY
The RBOCs expect to reduce the average loop length to less than 1,000 feet. Many-perhaps
most-homes and businesses would thus be in range of the highest-rate VDSL services. In effect,
this would provide the bandwidth of a T3 connection at a price that could be within reach of all
small businesses and even some residential users. But is it too good to be true?
The key prerequisite for VDSL availability is the deployment of fiber remotes to shorten loop
length. SBC's (www.sbc.com) Project Pronto is the best-published RBOC effort in this area, and
SBC has an aggressive deployment schedule that would deploy nearly 10,000 such remotes by
the end of 2000. Overall, the company expects to deploy nearly 30,000 remotes.
I estimate that the total number of fiber remotes that could be deployed by 2005 in the United
States will exceed 100,000. Such a number would put VDSL within at least technical reach of as
many as 80 million sites.
VDSL HURDLES
As important as the mass deployment of fiber remotes would be to VDSL availability, it doesn't tell
the whole story. Two other factors could limit the extent to which the deployed fiber remotes
could offer VDSL service.
The first factor is the bandwidth between the remote and carrier serving offices-the “uplink.”
Current fiber remotes from Alcatel (www.alcatel.com)-the feature player in the SBC
deployment-include a model with two OC-3 uplinks and one with an OC-12 uplink. Two OC-3s
offer about 300Mbits/sec; an OC-12, about 600Mbits/sec.
If each user has 58Mbits/sec, there won't be very many users per fiber remote. This would make
the cost per user unacceptably high. While costs could be reduced by assuming that users shared
the uplink for nonvoice traffic, bandwidth oversubscription would reduce the chance that VDSL
loops could support video applications.
The second factor is total infrastructure cost. Running a fiber connection to each remote would
require massive fiber deployment. The answer is a hierarchical network; for example, power
companies generate high voltage and deliver it to distribution centers and subcenters, from which
it eventually works its way to the home. In information networking, the architecture that shows the
greatest promise is Passive Optical Networking (PON).
PON fiber networks, unlike our power-grid example and today's traditional networks, don't require a
device at each fiber junction. Instead, the fiber is installed like a tree, with “branches” off from the
“trunk” (which connects to the carrier's serving office), and branches from other branches. Each
branch point is a passive optical splice.
The beauty of this approach is that a very large service area can be covered by a PON at a very
low equipment cost; equipment is required only at the head end of the trunk and the serving point
at the end of each active branch.
APON ON THE LINE
The problem PON generates, however, is obvious: The laser transmission from each user point
collides with others at each splice creating mass confusion. Fortunately, Dense Wavelength
Division Multiplexing (DWDM) can keep individual PON users separate. ATM technology can
format information on each wavelength, creating an ATM PON (APON) system. APON is the
subject of an ITU standard, G.983.
One of the Alcatel re-motes is compatible with APON, according to Alcatel literature. SBC and
BellSouth have both indicated plans to deploy APON technology in their new broadband access
networks. APON trials also are being conducted internationally. However, it's not clear whether
any, or all, of the 10,000 SBC remotes scheduled for deployment in 2000 will be on PON fiber. If
you want to believe in a VDSL future, you'd better hope that all of them are.
If PON/APON is deployed selectively, it may foreclose universal VDSL because of the cost of
retrofitting non-PON remotes. Likewise, a large commitment to low-speed fiber remotes (OC-3)
would starve the VDSL loops of the uplink bandwidth needed to support high-demand
applications like video.
There may be no issue as important to broadband applications as the success of VDSL. Users
should monitor the fiber access products and carrier deployment plans to ensure their interests are
served.
Tom Nolle is president of CIMI (www.cimicorp.com), a consulting firm for strategic technology
planning. He can be reached at tnolle@cimicorp.com. |
