Williams, Corvis Lash Back

BALTIMORE – NFOEC – Top executives from Williams Communications Group (NYSE: WCG) and Corvis Corp. (Nasdaq: CORV) banded together at this morning’s opening session at the NFOEC conference in Baltimore to deny the notion of a bandwidth glut.

The position is unconventional, to say the least. The telecommunications industry is scaling back on expansion plans as the financial markets have largely voted against the need for carriers to expand their fiber optic networks while their losses and debt levels mount.

Matthew Bross, senior vice president and chief technology officer for Williams, and David Huber, chief executive officer of Corvis, presented separately at this morning’s session. But both officials blasted the media and analysts for perpetuating what they say is a false claim that too much capacity exists in carrier networks.

In his remarks, Huber went right for the jugular: “The media and analyst community have done a disservice to the public with talk of excess bandwidth,” he said. “It’s true there is a lot of dark fiber, but the fiber cost is only about 5 percent of the total network cost.”

Both Huber and Bross emphasized the difference between dark fiber, raw cabling with hundreds of fiber installed in the ground, and lit capacity, usable bandwidth.

“It’s like saying there are too many microprocessors in the world because there’s a lot of sand on the beach,” said Bross. “Yeah, there’s a lot of fiber in the ground, but you still have to light it and turn it into usable bandwidth. That’s not a long-term oversupply.”

Huber blames the media’s mischaracterization of the issue for exacerbating the lack of capital in the sector. But venture capitalists contend the pervasive notion that there is already too much capacity hasn’t really affected VC investments in new startups looking to create more bandwidth. Instead, they say, the long-haul network was simply over-funded.

“Some areas have definitely been over-invested in,” says Sandy Roskes, investment manager for PentaPort Venture Advisors. “How many more all-optical switching companies do there need to be? Or how many MEMS companies are necessary?”

Bross and Huber aren’t the first executives to lash out at the press over the bandwidth glut. Joseph Nacchio, CEO of Qwest Communications International Corp. (NYSE: Q) ranted on a conference call with analysts and investors last month over the exact same issue. He said that simply calling dark fiber “bandwidth” was misleading, because for every $1 used to dig up a street and install fiber it costs another $2 to actually make that fiber usable (see Sour Grapes of Roth).

“If you want to measure utilization, and you want to measure capacity in this industry, you have to look at the three dollars you have to spend in order to build a network, not the one dollar,” he said.

While these executives say there is plenty of fiber installed in the ground, one lingering question remains: Is there already enough dark fiber lit to satisfy the current bandwidth demand? Williams’s Bross says that demand for bandwidth on his network is growing.

But some industry watchers disagree with Bross’s assessment. “At the moment there is plenty of fiber,” said Peter Tierney, vice chairman of Sphera Optical Networks Inc.. “And these carriers have already lit a lot of capacity. They have enough bandwidth for sale in certain routes; they’re constricted in others. But for the most part they are fine right now.”

These differences of opinion are reflected in the results of Light Research's poll this month, The bandwidth glut. So far, 41 percent of respondents say there's too much capacity, 23 percent say there's too little and 36 percent say it's about right.

- Marguerite Reardon, Senior Editor, Light Reading

Page 1 / 6   >   >>
braddillman 12/4/2012 | 8:06:20 PM
re: Williams, Corvis Lash Back Thanks for the numbers.

Does it make sense to guesstimate what customers are willing to pay for residential bandwidth?

Cases: (all figures pulled from my backside, feel free to correct me, please). Obviously, the actual prices and bandwidth vary, etc.

Local phone: $20/month for 56kbs
Long distance phone: $0.05/minute for 56kbs
Cable modem: $40/month for 500kbs
DSL: $40/month for 500kbs
Satellite TV: $60/month for a few simultaneuous, one-way channels (I can have up to 5 receivers for no addl. cost) of ???Mbs (I dunno - 3-5Mbs I would guess)
Digital cable TV: $40/month, similar to satellite TV
Cell phone: $50/month for < 56kbs, but it's mobile.

I think the pattern is that people will pay tens of dollars per month per service. But I doubt any residential customers would pay hundreds of dollars per month per service. Most people don't have that kind of money.

I don't think the total market or revenue for residential communications services can grow fantasically, simply because the average family budget has limits. I can only afford to spend xx% of my net income on those services.

So, how do the residential service providers grow? By taking market share. How to do that? By offering residents a lower total cost, while protecting their margins (not simply reducing their prices).

Example: customer 'X' pays:
$40/month for cable to the cable company
$20/month local service to the LEC (I think)
$0.05/minute to the long distance provider.

The cable company could offer phone for $20/month or less, using VoIP (assuming that technology works). The customer would have a single bill and company to deal with. The cable company makes money on the phone service if it costs them less than they charge the customer (duh).

Anyway, I doubt that residents will buy more bandwidth at current prices. I think the recent growth in residential bandwidth demand started when prices dropped (does that means it's elastic?).

I also doubt that residents have unlimited money for new services, even from killer apps.

I think bandwidth demand is there, but only for cheap bandwidth and a limited number of services.
mpl 12/4/2012 | 8:06:22 PM
re: Williams, Corvis Lash Back Some factoids:

Napster accounted for 25% of total US Internet traffic earlier this year.

Universities like Santa Clara, Berkeley, and USC had to place bandwidth limits on the student LAN using monitoring software during the Napster bandwidth-hogging epidemic. Student LAN connections were monitored for traffic abuse and their LAN privileges halted if they allowed significant downloads from their LAN connection.

Napster was shut down during the second quarter of 2001.

Nortel recently announced that the 2nd quarter of 2001 saw an actual decrease in Internet traffic. I wonder what could have possibly accounted for that???

Qwest recently announced that their network is operating at 85% of total capacity.

Livermore labs LAN has bandwidth demands of 5 to 40 gibabits per second in 2001.

@Home's broadband cable network was requiring nearly 5 gbps for their 3 million subscribers at the end of 2000. This capacity was being supplied by two OC48 circuits provided by AT&T.

@Home's average subscriber used 1.7kbps of broadband capacity continuously. This is equivalent to 550 megabytes of data transfer each month or 110 five minute long MP3 songs encoded at 128kbps.

A compressed 2-hour high definition video would require 20 megabits per second of capacity and would consume 18 gigabytes (not gigabits) of bandwidth.

math: 20 megabits/second x 7200 seconds x 1 megabyte/8 megabits = 18,000Mbytes = 18 Gigabytes

The download of a single high definition video would consume 33 times the bandwidth that the average @Home subscriber consumes each month.

The total monthly US Internet traffic estimate of 70 to 200 petabytes was for December of 2000.

Assuming 200 petabytes of data traffic each month, this can be accounted for by a continuous data stream of 620 gigabits per second.

math: 620 gigabits/second x 3600 seconds/hour x 24 hours/day x 30 days/month x 1 gigabyte/8 gigabits x 1 petabyte/1,000,000 gigabytes = 201 petabytes/month.
redface 12/4/2012 | 8:06:26 PM
re: Williams, Corvis Lash Back Guys,

The discussions are excellent. Regarding applications to use the broadband service, I think the most important is a practical high quality video phone (as compared to the current video phone from AT&T). Because I think the ultimate goal of telecommunications is to eliminate the space constraints so that people can interact from anywhere as though they are physically together. 90% of information comes to us through visual means. With a high quality video phone people can see each other daily, so traveling will be much reduced. With it, businesses will move faster. Much of the sales and sales support travel can be eliminated, resulting in major cost savings for businesses.
dietaryfiber 12/4/2012 | 8:06:28 PM
re: Williams, Corvis Lash Back The issue in the local loop is rate of return. The question is: Why should a company invest in local loop electronics? The answer is: To make a good ROI.

Now, CLECs want access to the electronics but do not want to pay for the ILECs to productize this service (DSL implementation takes training, test equipment, back office systems, etc.). So, the RBOCs have slow rolled the implementation of high bandwidth local loops. Think FTTH is going to happen if it requires unbundling at the physical layer?

Do consumers want the service? Yes, but people act as if it is their right to obtain such a service when and where they want it. This says they think this is a right and not a priviledge. Well, if its a right then a structure that provides adequate ROI to the telcos has to be found.

The best part of all of this is that the classic CLECs have little to no interest in residential service (a few do, but most don't). The people that benefit from all of this are the cable guys. Please note the lack of similar regulation to unbundle local cable segments. Is AT&T Broadband not a monopoly? Should they not have to unbundle?

candlelight 12/4/2012 | 8:06:28 PM
re: Williams, Corvis Lash Back Jmd,

Thanks for the tip on Blind Watch Maker. I will try to get hold of a copy.

I was just wondering why businesses are failing
this time around if they can always find a way
out to make money.

1) Is it too much competetion for a small market?
Then what's happening now is process of natural selection and the industry will revive with a few survivors who will monopolize that small market.

2) Or too much optimism leading to unecessary
investments? Were the visionaries hoping to shortcut natural evolution of business by putting too much money too soon based on fancy ideas.

I agree with you there is no good application out there for average Joe to buy DSL let alone fiber to home.

But what will you do when you do have an application? You will need to put too much money to build the infrastructure so that Joe Soap can use that application.

Anyways I still stick by my stand that:

The returns in this industry cannot come overnight. And the reasons are

First that it takes time to hook people to a new idea and internet is still a new idea without any application which is worth fiber to home

(Does Joe need 100 Mbps for buying Airline Tickets online? Does he need to watch news channels on Internet when he gets it on his CATV? Do we need so much bandwidth to search on Google or check Emails or Chat with Sexy Sarah/Macho Gustafson on a Porn Site? Napster was the only time when I thought Cable Modem is worth the money but they killed it).

Second, infrastructure is expensive and you cannot use the usual business strategy of giving freebies to each and every person to get his/her attention.
(Consider the failure of so many ISP's who were offering free connections to subscribers over dialup lines. Wonder what will happen to guys who give fiber to home as a freebie)

The only way I see carriers can make money with
current business strategies is to stick with very cheap low bandwidth solutions for home customers and target big businesses with fibers.

My second request in the post was if you really want average Joe to give you money for Fiber to home what applications do you need?

I can add the following application:

1. How about education over internet. Live classes taught by Experts. You can make it interactive too. Learn French from an expert in Paris... Learn
to play piano from a internet renowned teacher..

I would like readers to add more applications. Then we can convince the VC's again to throw some more money at us..

jmd 12/4/2012 | 8:06:30 PM
re: Williams, Corvis Lash Back There is no chicken and egg problem. Chicken and egg problems only arise because people assume the chicken or the egg (which ever) magically appeared in a single event GĒō it didnGĒÖt. Read The Blind Watch Maker.

In this case you are saying the network providers wonGĒÖt build out because there is no demand for bandwidth and the end consumer is not consuming bandwidth because thereGĒÖs a lack of bandwidth to hook them. I disagree. Things start with a reason for people to wanting more bandwidth. That reason must distil into a value received by the consumer. If the application is there then businesses (either new or existing ones) will find a way to extract the profits.

optical_survivor 12/4/2012 | 8:06:30 PM
re: Williams, Corvis Lash Back Dear Colleague,

I had the following post on a different track and got very few follow-up posting from the audience. I would welcome your thoughts on the subject matter. care to share your insight with me???



I have a subject matter I am really interested: I
assume I am not the only one. Recently, there was a
small track on whether 40G really makes sense any time
soon. Granted, I know system guys can solve the
problem shortly: there are a lot of really smart folks
in this industry and I am sure they will be able to
figure this out. The issue is the ability of the
existing fiber infrastructure to support ULH/LH 40G
systems in a cost effective manner. Some tid bits I
saw lately and some speculations on my part:

1. as services providers consolidate (already
happening), the combined, embedded
fiber/infrastructure is sufficient to support all the
projected traffic growth during next several years.
For instance, if the wall street still encouraged
360G, Level 3, QWest all to build their OWN network,
each of them may be compelled to build out the
network. In the current climate, if they consolidate,
the existing combined carrying capacity gives them
very little incentive to upgrade their 10G to 40G.

2. I have seen that as much as 85% of the existing
fiber infrastructure will not be able to support 40G.
Is this correct??? If not, what kind of special
compensation is needed for PMD, chromatic dispersion,
etc, and how does this affect the economics of 40G?
Note that Corning announced earlier this year that the
new fiber deployment this year will be less than 50%
of last year. This is another bad news for 40G. In
order for 40G to be deployed widely, you would want to
see more and more active build up of infrastructure
with the state of the art fibers available now.

I saw a statement that there is a 10 year fiber glut.
Is this true??? I also saw that fiber infrastructure
built before the end of 1997/beginning of 1998 cannot
support 40G. Ture??? Roughly, what % of the fibers
currently deployed and to be deployed up to YE 2003
can support 40G economically??? Any estimates???

3. General accepted wisdom is that 4X capacity comes
at a 2.5X price in the optical transmission business.
However, this is the case when the cost reduction
curve for the new technology follows the normal
when 10g first came out, I heard that the initial
price per 10G was not 2.5 times that of 2.5G, but
rather close to 5 times of the 2.5G. It was justified
with the "network simplications" argument (it's easier
to provision/operate one 10G than four 2.5G links). I
believe (I could be wrong) it took about 3 years for
10G to really economically prove over 2.5G in terms of
the initial deployment cost (2.5 times the price of
2.5G). However, this was during the go-go years when
everybody was merrily building up their network at a
furious pace. It takes a large scale sales traction
for theh cost/price curve to follow a steep downward
slope. Now, it's 2001, and everybody's CAPEX has been
drastically reduced. we may not see, any time soon,
the kind of craziness that drove the industry last few
years. Given this, isn't it a fair assumption that
40G economics will take a MUCH LONGER time to prove in

jmd 12/4/2012 | 8:06:31 PM
re: Williams, Corvis Lash Back Are there any hard numbers as to how much typical delays are with respect to access, local loop and peering points?
mikeward 12/4/2012 | 8:06:31 PM
re: Williams, Corvis Lash Back a similar scheme is currently being mooted in australia. i'm hazy on the details but you can catch the story in comms week international.
mikeward 12/4/2012 | 8:06:31 PM
re: Williams, Corvis Lash Back disclosure: i'm not mike ward of corvis. i have no professional interest in or affiliation to corvis. i made the comments in my capacity as an observer of the networking space - not as a partisan.

Page 1 / 6   >   >>
Sign In