Hi, ftth. Here's a mind-ramble that may come close to what I meant, initially, and hopefully answers your question. In all fairness, the more I pondered this, the more my view seemed to morph, although it's still on track with my original idea.
------------
You're absolutely correct. Much, probably all, depends on the assumptions made. Before I elaborate on my assumptions, it's probably best to examine the initial circumstances and assumptions that a last mile provider makes (or made) in a more general sense during their initial design stages.
A primary factor tracks with the presumed demand they will be called upon to support and how well it can be satisfied by the type of medium they choose (or have been relegated to, as it were) to employ, whether it's fiber, or copper pair, black coax, wireless, even powerline now, or a hybrid of any of the foregoing. I won't get into the other criteria, except to say that many other factors exist, as well.
Some forms of media are bounded more than others at the upper end of the range (certain flavors of xDSL, e.g., although I think John Cioffi is about to give us a new lesson in 100 Mb/s-plus VDSL, here), and some are not (raw fiber, e.g.) for all intents and purposes, based on existing and foreseeable applications - even those of the most seemingly bizarre nature.
So, any assumptions I make would be case-specific, and we've been over most of those here, before.
Where I'm going with this would seem to be fairly straightforward, were it not for some multivariate factors that come into play, especially with regard to declining costs of bandwidth, increasing usage and the requirements to modify or retrofit infrastructure to accommodate increased loads, with all of this taking place on Moore's front porch.
Today, in the absence of heavy movie downloading, and in the absence of a mature repertoire of enterprise VPN multimedia applications --both are still nascent applications in relative terms, when it comes to isochronous-like applications-- there exists a level of traffic activity that is being satisfied by links that were sized to meet an arbitrary level of demand, with a certain amount of extensibility factored in, for every media type I've mentioned above. Some are provided on a flat rate basis to users, and some, and this is becoming increasingly obvious now with the MSOs, are tiered.
When the ability to download a broad selection of movie and other forms of multimedia titles is added to the mix of flat rate recipients, along with gaming and other gluttonous applications, which also occurs when the blossoming effects of VoIP and MM use by telecommunters becomes prevalent, the interconnect capacity to the SPs' Internet provider(s) along with some other network elements must be sized and at some point re-dimensioned, accordingly.
In short, when bandwidth requirements increase in the local access space, so, too, do the direct and indirect costs of providing capacity and access to the content that comes across the larger Internet. The impact of this can be partially assuaged by the local caching of content and other ways, but those are assumptions for another discussion. And granted, this additional burden is also mitigated to some extent by the declining costs of bandwidth in general.
Are bandwidth costs declining at a rate that will outpace the demand that will be required in order to offset the lower revenues per megabit delivered? Here we get back to the issue of pricing elasticity and the backlash of same when it comes to the bandwidth providers' inability to turn a profit. We read with increasing frequency that there's no profit to be made in pure tranposrt or even plain old vanilla IP, from an xSP's perspective. So the natural thing to do would be for them to add value through the introduction of vertical services, and charge accordingly, or to begin tiering access fees for vanilla services by the amounts consumed.
Also, what is the incentive of the SP to reduce pricing to the end user when bandwidth rates "do" decline significantly. As I'm sure you've already concluded, I'm not referring to the price-cost considerations of the bandwidth availability in the last mile, since those rate components are based on sunk costs and have a twenty-year depreciation cycle to undergo (in the wireline/fiber area, in any event... wireless is another story).
I'm referring here, instead, to the cost of the interconnect bandwidth, as you call it, to the Internet proper, and the incremental capital costs associated with head end gear to process, manage and route it.
It might seem naive to ask such a question (above, based on cost-based pricing), but some municipal last mile providers, ostensibly, have it in their philosophy to pass along savings to users when possible. Unamerican? Probably.
Where I'm going with this is nicely explained, I think, by the dialog that I posted here a couple of years ago between a veteran ISP jock and a relatively junior ISP tech type. I'll try to locate the url in the archive and post it later.
(Btw, have you noticed the improvements that the SI Admins have done to the search capabilities? Pretty cool, for a change. Now... what can we complain about, next? ;)
Today we're basically running discreet applications on an interrupted basis, in burst mode, in what could easily be described as a form of traffic that is statistically benign and forgiving. Take the other extreme, where you are filling your pipe constantly not only with topical items at hand, but with background downloads as well.
I'm a programmer working from home, say, and I'm rebuilding a Sybase database, remotely. At the same time I'm participating in a video-conference with the head office. I may elect to simultaneously begin downloading a couple of the movies the kids have been asking for. Also running in background are a whole host of other applications, some of them futures. These include a port from my residential router feeding multiple extensions throughout my home in support of POTS phones, background music, or whatever. The line is never in a quiescent state at this point.
Multiply this activity by the some reasonable number of users on the segment, depending on the demographics, all being fed from the same head end, and the demand for bandwidth on the back end, along with new traffic engineering provisions in the head end, go up in lockstep.
On the matter of South Korea's achievements in these regards, keep in mind that much of their hegemony has been the result of direct government influence is seeing a vision through, both philosophicall and financially. I've posted what follows here before. It's one of many articles over the past year that examines the phenomena that has been taking place along the Pacific Rim and in parts of Asia.
Here's the NY Times article from May of 2003 on S.K. It's a longie, but well worth the read to anyone interested in SK's broadband background:
--------------------------------------------------------------------
America's Broadband Dream Is Alive in Korea
May 5, 2003
By KEN BELSON with MATT RICHTEL
SEOUL, South Korea - As Cho Won Hee zips effortlessly from
one Web site to another, his doting mother at his side, it
is easy to understand why Silicon Valley views South Korea
as the promised land of instant access to the Internet.
The Chos' high-speed digital line - 100 times faster than
the typical dial-up connection in the United States - is
their zippy gateway to home entertainment, education and
shopping, all for $32 a month. And despite the relatively
recent arrival of such connections, the Chos, like many
Koreans, are already as addicted to their broadband hookup
as most Americans are to their television sets.
The Chos are at the cutting edge of South Korea's grand
experiment with all things broadband, the catch-all name
for high-speed digital connections. With a hefty push from
the government, South Korea's telecommunications providers
have built the world's most comprehensive Internet network,
supplying affordable and reliable access that far surpasses
what is available in the United States, even in those homes
that have their own broadband setup.
And now that most of the nation is online at high speeds,
South Koreans are shifting more of their analog lives to
their computers, where they watch soap operas, attend
virtual test preparation schools, sing karaoke and, most of
all, play games.
By embracing broadband so heartily, Koreans have turned
their country into a test case for the visionaries who,
just a few years ago, imagined a future of nearly infinite
digital possibilities. While those dreams have hit speed
bumps in the United States and elsewhere, South Korea -
with Japan not far behind - is racing ahead.
In the process, Koreans are offering a glimpse of what
wired societies are supposed to look like, where fast
Internet connections vastly increase access to information,
help lift productivity and create new markets.
"The killer application of the Internet is speed," said Lee
Yong Kyung, the chief executive of the KT Corporation,
formerly known as Korea Telecom, which controls nearly half
of the country's broadband market. "The money is in the
pipes."
But maybe not yet. Intense competition and overbuilding has
made prying profits out of building those pipes difficult.
And while some content providers have flourished, many
others still exist on government subsidies. Broadband has
also spawned worrying social trends, some say, raising
concerns about children addicted to online games and a
growing digital divide between the young and the old.
This is not unexpected, given the extraordinary pace of
change. Since 1998, telecommunications companies here have
installed nearly 11 million broadband lines, over 5 million
of those in the last year alone. High-speed lines now reach
significantly more than half of all homes with Internet
access.
The numbers are startling, given that South Korea was among
the nations hardest hit by the Asian financial crisis just
half a decade ago. But rather than retrench, the country
turned a disaster into an opportunity. Spending on
broadband and other high-technology gear helped lead a
transformation of the economy, pushing the overall
information technology sector to about 13 percent of
economic activity and making South Korea much less
dependent on heavy industry.
"In Korea, there was a sense of crisis and they needed to
take aggressive action to keep up with globalization," said
Izumi Aizu, who runs the Tokyo-based Asia Network Research
Inc. "In the U.S., the Internet has turned into a very
conventional business."
By racing the fastest down the information highway, Korea
has highlighted how far the United States has to go. Though
broadband connections are increasingly common in America,
service is comparatively expensive and coverage spotty.
Telecommunications companies in the United States, from
start-ups to long established businesses, spent hundreds of
billions of dollars to build fiber optic networks, but many
ran out of cash before they brought those lines the "last
mile" to people's doors.
When it comes to high-speed penetration of the home, the
United States lags well behind South Korea and Canada, and
has slipped below Japan.
America's uneven adoption of broadband has Silicon Valley
executives looking at South Korea with envy. While often
disdainful of government intervention, many high-technology
leaders in the United States now argue that Korean policy
makers got it right by actively promoting the technology.
The Korean government built a nationwide fiber network to
get students and others hooked on high-speed service. To
keep prices low, it encouraged rivals to compete with the
former state-run monopoly, KT, and it provided loans to
software ventures.
By contrast, the effort to bring broadband to the American
home is bogged down in the fight between the regional Bell
companies and their rivals.
Fee-based online services are now blossoming in Korea. Once
a novelty, home shopping now makes up 8.7 percent of all
retail sales, a rate that is expected nearly to double by
2005, according to Accenture, the global consulting
service.
For the Chos, their experience began, like so many Koreans,
when Won Hee, now 20, visited one of Seoul's many PC bangs,
or Internet cafes, while in high school. Soon he was
studying for his college entrance exams online and shopping
for music, videos, furniture and a vacuum for his mother.
Though he attends lectures at his college, he does most of
his other school work online, including making
presentations with his classmates. "The speed is the
biggest difference," he said. "Because all my friends have
broadband, we tend to use the Internet even more."
Though Korean parents often fawn over their sons, Mr. Cho's
parents grew jealous of Won Hee's connection. Since he was
always online at night, his father stayed late at his
office, where he had his own broadband line. His mother
wanted to study for a real estate broker's license online.
So the Chos leased a Wi-Fi base station from KT so their
other computers could gain access to the high-speed
connection. Mr. Cho started coming home earlier and Mrs.
Cho signed up for her course. Wi-Fi, also known as wireless
fidelity, is a technology for providing wireless Internet
access.
Including their home phone, three cellphones and cable
television, the Chos spend about $200 a month on
telecommunications fees, which is not atypical for wired
families in Korea and a hefty expense in a country where
the median annual household income is under $20,000,
roughly half that in the United States. Despite the
relatively low cost for high-speed Internet, that overall
sum, most experts here agree, is close to the limit of what
ordinary families will pay. So companies are now focusing
on providing even faster connection speeds and new services
without raising prices.
Like many American broadband users, the Chos started out
with what is called A.D.S.L., for asymmetric digital
subscriber line, which is best at downloading data over
broadband telephone networks. But they recently upgraded to
a system that is even faster in both directions, making it
easier to use interactive games and other two-way services.
The lines, capable of speeds of up to 40 megabits per
second, are much faster than anything commonly available in
the United States, where 1 megabit to 3 megabit
transmission rates are typical.
High-speed digital access is creating businesses that were
unworkable with ordinary dial-up connections. The Korean
company Megastudy, for example, has built the country's
biggest online test preparation school for college entrance
exams, while KT and rival Hanaro Telecom sell accounting
services over the Web to small businesses.
But entertainment, as expected, is the big attraction,
especially games and videos. In 2001, SBSi, the interactive
division of the Seoul Broadcasting System, started charging
500 South Korean won (about 40 cents) a show to watch soap
operas and other streaming video programs. The service has
attracted 1.8 million registered users; 4,000 more sign up
every day. The drama "All In," the true story of a Korean
gambler who beat the odds in Las Vegas, drew 1.6 million
viewers during its initial 24-episode run online; now
10,000 Koreans a day pay to see reruns on their computers.
"On the basis of this new infrastructure," said Hwang Eun
Ju, a manager in SBSi's strategy division, "we could
develop and benefit from new broadband content."
While content providers are taking advantage of Korea's
broadband network, the companies that built it are
besieged. Growth of new subscribers is leveling off and
providers, locked in a price war, are cutting installation
fees and giving away modems.
KT has the deepest pockets, but its continuing investment
in the new super-fast interactive technology is expected to
keep its broadband division in the red for at least another
year. The chief of the No. 2 player, Hanaro, resigned in
March in response to the company's mounting losses. The
third-largest provider, Korea Thrunet, filed for bankruptcy
protection from creditors, also in March, after failing to
find new investors.
"Turning a profit is not the issue; it's whether they can
survive or not," said Song Sauk Hun, an analyst at Gartner
Korea.
After encouraging rivals to enter the market, the
government is now quietly endorsing consolidation.
The United States has gone through a similar shakeout,
except it happened before the broadband network was
extensively built. The Telecommunications Act of 1996 set
off a surge of expansion that collapsed when the Internet
bubble burst, driving many of the broadband start-ups, like
Rhythms NetConnections and NorthPoint Communications, out
of business. While fixed-line operators in Korea and Japan
were cajoled into making D.S.L. service available at low
cost, analysts say that the Bells are reluctant to cut
prices.
At around $50 a month, broadband costs about twice as much
in the United States as in Korea and Japan. Worse,
broadband in the United States is slower and less suited
for interactive entertainment and other two-way uses
because it relies on an asymmetric system that receives
data much faster than it can send it.
The Bells say they are doing everything they can to promote
broadband. But critics say the phone companies view
broadband as more of a threat than an opportunity, so they
have done little to rectify these problems.
The phone companies "are a very powerful industry that
spends enormous amount on lobbying," said Charles H.
Ferguson, a senior fellow at the Brookings Institution in
Washington who is working on a book on broadband. "They've
been able to retard progress and competition."
Even so, after a slow start, the United States is catching
up, mostly because the cable industry has picked up the
ball. At the end of 2002, about 16.4 million homes had
broadband lines, up 52 percent from a year earlier. Cable
companies, which provide more than half of all connections,
will invest at least $10 billion this year in new
infrastructure.
America's suburban expanse certainly adds to the expense of
connecting every home, but until broadband costs less,
supporters say, consumers will remain wary.
"If the prices for high-speed access were $25 or $27," said
Joseph A. Crupi, vice president for broadband communication
at Texas Instruments, "it would be a no-brainer."
But while few question the advantages of having the nation
hooked to the Internet by high-speed services, many argue
about whether the benefits are worth the cost. The
Telecommunications Industry Association, which represents
Intel, Cisco Systems and Lucent Technologies, among others,
wants Washington to take a larger role in shaping a
national broadband policy. A bill now before Congress would
provide tax breaks of $2 billion over 10 years; it faces an
uphill struggle at a time of budget deficits.
Having seen how South Korea has turned itself into an
Internet powerhouse, broadband advocates say that the
United States risks losing out by not moving faster.
"People must have access to high-speed Internet from home,
as well as work," said Paul Saffo, a director of the
Institute for the Future, a technology industry forecasting
group in Menlo Park, Calif., "or we won't be full players
in the global economy."
nytimes.com
---------------------
FAC
frank@fttx.org |
