[opendtv] Broadband backwater
- From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Sun, 26 Jun 2005 18:55:44 -0400
Uncommonly good article on this topic.
"So where is the broadband Valhalla we were promised in
1996?"
About where one would expect, without govt subsidies,
is my answer.
Also this bit is on target, about use of passive
optical networks (PONs) for broadband deployment:
"But PON is hardly a silver bullet.
...
"Juniper Networks believes the optimal TV delivery
mechanism for carriers using PON will be to put
broadcast channels and a standard cable package on an
independent wavelength, using an overlay model without
a TCP/IP stack. Custom channels chosen by the
individual consumer then would be delivered via
streaming IP TV. Boland of Juniper said the approach
would yield most of the features of the
multihundred-channel environments that Japanese
carriers are delivering via active fiber, but without
the heavy bandwidth requirements of switched active
fiber."
Sounds right on target. Interestingly, very similar
architecturally to a cable network which provides
broadband Internet access.
Bert
------------------------------------------
Broadband backwater
Loring Wirbel
(06/20/2005 9:00 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=3D164900821
Warren Ryd and Barbara Doyle almost resorted to constructing
a makeshift microwave tower last month in the backyard of
their Black Forest, Colo., home. The couple isn't
particularly consumed with microwave links; they're not even
RF-savvy. They're two people trying to build an Internet
consulting business.
That's easier said than done in Black Forest. The town lies
a few short miles from the sprawling, ultramodern exurbs on
the eastern fringes of Colorado Springs and is just 22,000
feet from the nearest Qwest Communications central office.
But for the area's incumbent phone and cable operators,
Black Forest, Colo., might just as well be Germany's Black
Forest - simply too remote to care about.
An undercapitalized Qwest has struggled to continue
digital-subscriber-line buildouts in its vast but sparsely
populated service area. And Adelphia Communications'
struggles with bankruptcy and a corporate selloff mean that
a hybrid fiber-coax plant won't make it to Black Forest
anytime soon.
"At least Adelphia came out and said they wouldn't be
building out for years," Ryd said, "but Qwest kept giving me
different stories" about when its DSL access multiplexers
would be in place. "It seems the sales folks never really
talked to those making the provisioning decisions, so no one
would give a solid 'No.'"
He and Doyle are desperate. They've resorted to signing up
for 802.11 Wi-Fi service at a local coffee shop, Ryd said,
because they "just can't tolerate 56k any longer."
So where is the broadband Valhalla we were promised in 1996?
The mild upturn in carrier fortunes at the recent Supercomm
show in Chicago could not disguise industry leaders' concern
over U.S. residential broadband penetration trends. In a
keynote speech, Tellabs Inc. chief executive officer Krish
Prabhu cited International Telecommunications Union (ITU)
statistics showing that "we are now in 16th place [in
broadband penetration] worldwide, and the numbers are still
trending down."
"There are tens of millions of us in the U.S. who live with
abysmal bandwidth," Nortel Networks chief executive officer
Bill Owens said during a Supercomm presentation. "Our
vision of regulatory policy is virtually nonexistent."
In his article "Down to the Wire" in the May/June issue of
Foreign Affairs, Thomas Bleha, a former foreign-service
officer in Japan, said that no U.S. carrier has revealed
plans that can match Japan's nearly universal availability
of 20-Mbit/second broadband for $22 a month or that
country's plans for 100-Mbit/s fiber-to-the-premises that
will provide 30 percent of its citizens with active-fiber
connections for between $30 and $40 per month by 2006.
If the President's Information Technology Advisory
Committee worked with carriers to develop usage plans for
existing dark fiber, Bleha asserts, basic broadband access
could become universal in the States, for a subscription
fee of $20 a month, by 2010. By that same year, he says,
two-thirds of all U.S. households could have the option to
switch to 100-Mbit/s fiber services for $30 to $35 a month.
Failure to drive this goal as national policy could make
U.S. companies fundamentally uncompetitive with most Asian
and European nations within a decade, Bleha warns.
Verizon Communications Inc., SBC Communications Inc. and
BellSouth Corp. are not waiting for a sign from the White
House; each is rolling out passive optical network (PON)
installations to enable multimegabit customer access
speeds at minimal cost to the carrier. Strategies range
from all-PON fiber-to-the-home, used in Verizon's FiOS, to
fiber-to-the-node architectures that employ twisted-pair
copper in short VDSL loops.
Still, critics compare the PON efforts unfavorably with
the efforts to provide 100 Mbits/s to all citizens in
Japan and South Korea via active fiber.
"We definitely see more interest in Asia for the newest
upgrades for our broadband routers," acknowledged David
Boland, senior manager for access routers at Juniper
Networks Inc. But don't count out the States, Boland
said. "Have Japan and Korea and other countries gained
momentum by having a policy to bring fiber to every
user? Of course. But you can't assume as a result that
the U.S. has become a broadband no-show."
Telecom equipment OEMs and semiconductor suppliers agree
that the incumbent local-exchange carriers and cable
multisystem operators have become stalled at current
generations of asymmetric digital subscriber line (ADSL)
for twisted-pair access and at Docsis cable modems for
hybrid fiber-coax systems.
Lothar Pauly, chief executive of Siemens AG, said that
carriers are being too conservative when they talk about
"triple play" services' uniting voice, data, and video.
"Triple play is a great defensive play, but a grand slam
is what puts points on the board," Pauly quipped.
An all-encompassing vision must consider local- and
wide-area wireless services as part of an integrated
whole with other broadband consumer offerings, Pauly
said. Phone, cable and utility companies, he said, are
not thinking long term when they develop services as
defensive measures.
Every carrier today has a different view on how
broadcast and streaming Internet Protocol TV channels
will be carried, and each approach has direct
implications for the bandwidth that must be delivered to
the user.
PON: No paradise, no panacea
With region-wide programs like Verizon's FiOS and SBC's
Lightspeed, PONs are being adopted with fervor by
carriers looking for a low-cost way to to drive fiber
speeds closer to the customer. By limiting active-fiber
switching to a neighborhood node and using passive
fiber components in the last mile, PON promises to give
carriers a faster path to profitability than
traditional FTTH programs. PON also can be used in
conjunction with twisted-pair schemes that use copper
wire from the home to the optical line termination
device, and the faster speed of the backbone allows use
of faster DSL alternatives, such as VDSL and VDSL2, to
accelerate customer download speeds to 25 or 50
Mbits/s.
For some business and even residential networks, it
might make sense to use active switched fiber from a
broadband loop carrier, as Occam Networks Inc. and Telco
Systems Inc. have demonstrated. Utility companies have
found such networks in "greenfield" residential builds,
where they can also use the optical terminal for meter
reads.
But PON is hardly a silver bullet.
Double play was an easy step up, even when extended to
digital cordless handsets, said Faraj Aalaei, CEO of
Centillium Communications Inc., which makes xDSL chips
and also has formed an optical-components group to
design PON chips. But video support in triple-play
requires a new way of thinking about how content is
delivered, Aalaei said.
"There are different approaches for mixing broadcast
overlay and IPTV models, but we have to assume that a
single home platform must support multiple streams of
multicast IP traffic," Aalaei said. "When I say
'triple play,' I can't refer to support for a single
broadcast channel."
In general, PON semiconductor makers are either
following the ATM-oriented evolutionary path of APON
to broadband PON and then gigabit PON or are leaping
straight to the Ethernet-only EPON and GEPON
environment (the approach favored in some Asian
nations). BroadLight Inc. and Freescale Semiconductor
are high-profile GPON suppliers; Teknovus Inc. and
Passav Technologies have cast their lot with EPON.
At the end of the day, the real issue is how
packetized services are handled. That means
VDSL/VDSL2, or even ADSL2+, can be used as a temporary
last-mile feeder for PON service; but, long term, PON
must extend deeper into the network toward the end
user.
"If I have a copper infrastructure I want to utilize,
it may make sense to have a fiber-to-the-node
architecture that uses VDSL or VDSL2," said Matt James,
senior marketing manager for PON products at Alcatel
North America. "But when service providers want to
address IPTV, they have to look seriously at a network
transformation.
"Yes, you can do true triple play with VDSL2, maybe
even with the 26 Mbits/second of VDSL1, but you'll
need GPON to the premises for speeds above 100 Mbits a
second - and I think that is the end game most aim at."
Juniper Networks believes the optimal TV delivery
mechanism for carriers using PON will be to put
broadcast channels and a standard cable package on an
independent wavelength, using an overlay model without
a TCP/IP stack. Custom channels chosen by the
individual consumer then would be delivered via
streaming IP TV. Boland of Juniper said the approach
would yield most of the features of the
multihundred-channel environments that Japanese
carriers are delivering via active fiber, but without
the heavy bandwidth requirements of switched active
fiber.
Niket Jindal, program manager for PON at Freescale,
said such an architecture may be compelling, but
"every time you talk about an RF video overlay, there
are costs associated with distinguishing different
types of TV delivery. Ultimately, we need to get fiber
all the way to the home, and eventually there may be
more active fiber components in the network."
At the same time, Jindal said, "we cannot look too
closely at Asian markets with dense populations as
models. First of all, there's a lot less focus in the
U.S. on upstream speeds than in many Asian markets.
Second, with a large and distributed population in the
U.S., network evolution always takes longer than we
anticipate."
Is a shift from ADSL2+ to VDSL2 perhaps too big a
jump for networks with copper extending from optical
network nodes? Tony Zarola, strategic business manager
for broadband products at Analog Devices Inc., noted
that DSL customer-premises equipment manufacturers
have just upgraded to the 24-Mbit/s ADSL2+, so
expecting them to migrate now to VDSL2 may be asking
too much. A better option may be bonded ADSL2+
channels, Zarola said. In bonding, two channels are
logically combined to provide the aggregated bandwidth
of the two channels. Thus, two 26-Mbit VDSL1 bonded
channels would provide the equivalent of more than 50
Mbits/s.
Infineon Technologies Inc., for its part, is betting
that the bonding option will prove just as viable for
VDSL2, where it would provide consumers the same
100-Mbit speeds they could get from active-fiber
connections.
Tom Starr, chief standards director at SBC, said
carriers regard VDSL2 as a breakthrough for
complementing passive fiber strategies that move to
the neighborhood. Star, who had worked on various DSL
standards, said the work to finish VDSL2 has
generated more excited than he has seen since "at
least the very early days of ADSL."
The newest use of DSL comes from Aktino Inc., a
startup founded by PairGain Communications executives
that uses multiple-input, multiple-output frequency
division multiplexed channels for copper binding. The
company launched backhaul and metro-extender roles
for its MIMO-over-DSL system, based on T1/T3 channels,
in early May and showed an Ethernet version of the
same system at the Supercomm show.
All material on this site Copyright 2005 CMP Media LLC.
All rights reserved.
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
