Slashdot Mirror
Bandwidth Shortage And The Telephone Company
FasterThanLight writes: "This article from USA Today regarding (non)usage of existing fiber and its impact on bandwidth in the semi-near future ... more doom and gloom. Why? Greed, of and by the (surprise, surprise) large telcos." Remember, this story is about a predicted shortage, not a current shortage.
When I see billboards around town suggesting a second phone line for internet use, I suspect the telcos' plan is to get lots of people addicted to having long internet sessions on their modem, after which the telcos will go running to Congress saying that they need to start charging local calls by the minute due to the excessive connect times in the Internet Age.
Why, why am I so cynical? Oh, well. It was fun while it lasted. I now return you to the scheduled rants...
Sheesh, evil *and* a jerk. -- Jade
Seriously, this is just another business cycle. Lots of people jump in, overbuild for the current market, the market crashes, the survivors consolidate, demand marches onward, business picks up, lots of people jump in, ... Happens to everything: wheat, memory chips, telecom, you name it.
-- ;-)
Kuro5hin.org: where the good times never end.
Argh, a spam story, sssca and a telco story in one day, slashdot is really pissing me off today. :)
I can totally see my local telco using this to their advantage. These are the same jerks and assholes who charge an extra $15 a month for a static IP. All I want is 1mbit up and down, and no restrictions on what services I can use (ie. servers). Is it too hard for these broadband 'providers' to offer something so simple? Is it so hard?
NO, they'd rather spend their cash revamping their websites every 15 days with more 'features', or telemarketing me to death - and we already know that you're only service from a dsl provider is dslreports.com.
No sympathy from me - I'll continue to get screwed and like it, along with everyone else.
I have lived in three neighborhoods in Texas during the past 4 years, and during all but the past years, everytime I went on a drive, I could see construction trucks with spools of thick fiber lines being buried underground.
However, this all stopped about 4 months before I got layed off from Alcatel, one of the largest suppliers of fiber-optic cable in the world. There's just not as much business as there used to be, and people aren't willing to speculate on putting fiber down when they won't immediately see profits from it in this kind of market.
I don't, however, think we are going to see a shortage of bandwidth anytime soon...at least in big cities and suburban America (judging from the state of things in Texas). There's more cable underground than I care to think about, and I know for a fact you can get more bandwidth up and running in less than 9 months--the timespan that this article suggests. If there's a big enough market for it, telcos will have it in tomorrow! We threw up huge testbeds at Alcatel in under a month that could easily have served a small city a good amount more bandwidth.
Basically, this article is a bunch of speculative horseshit supported by quotes from people that either don't know what they are talking about, or have alterior motives for giving the quotes.
1. DWDM -- OC768 is coming your way and a lot of badly beat down telecom providers want to sell it.. BAD. Think they'll give ya a discount if yo buy a bunch of units?
2. Theres so much fiber in the ground on long haul routes that if there is ever anything resembling a shortage in my lifetime, I'd be impressed.
3. It's not nearly as expensive as the article poses to setup electronics on fiber. It's not $1 to install a fiber and $20 to put electronics on it. It just doesn't work like that.. Dug up the street lately? It costs a fortune.. Attaching the fiber to the Gigabit Ethernet port is far less.
4. Greed -- If there is money to be made, the bandwidth will be created.. It's called the law of supply and demand..
"Why? Greed, of and by the (surprise, surprise) large telcos."
What does this have to do with the shortage of services? Commodity prices are determined by the market. Aptly put by the article, fiber is like farmer's seed. Farmers are actually subsidized by the government to not grow certain crops, because it make it harder to make money when *everyone* is growing the same thing. When it costs 20 times to actually use dark fiber compared to just laying it down, this makes it hard to make money running a fiber service. The big telcos can afford to always undersell the startup. Such is the nature of the market. And don't spew any of that "there's no competition" crap, look at how many telcos there are. If that number is more than 1, you have competition.
When prices do rise due to "shortage" then as the article predicts, those who raise prices because they do not want to use more fiber will be undersold by new companies that will find it suddenly profitable to provide comms services. The article is basically predicting that the 20:1 cost ratio of use vs. creation will decrease.
This article completely discounts the obvious innovations coming to market for pushing more bits down the same fibers over time.
I also dispute the notion of an approaching shortage. Of course James Crowe wants you to believe that there is an impending shortage - his company is on the ropes and he desperately needs to foster the notion of the bandwidth boogeyman to keep investors interested in his moribund stock.
Yes if the "last mile" problem is solved, there will be a tremendous spike in demand...but lets be realistic about developments in the last mile - many telcos are scaling back or cancelling outright plans to push high bandwidth deeper into their networks. Case in point, SBC's "Project Pronto", which would have given 80% of SBC customers the equivalent of a 5-7MB/sec connection, has been cancelled for good. If SBC has to provide cut-rate access to their networks to companies like Covad, they simply aren't going to bother with upgrades...and forget about the government forcing companies like SBC to sell off the local loops, this isn't Cuba...hell will freeze over first.
The sad fact is that the regional Bells are only going to make major upgrades when they no longer have to subsidize the competition. It sucks, but its the inevitable fact.
Well, in the case of your telco, it loooks better on the books not to have capital depreciating, so their move made economic sense, from a business finance perspective.
I have serious doubts about this article though. While they make the single valid point that it costs money to light up a network, there were and still are valid mechanisms for financing that activity. Most of the telcos that have gone under were in debt to the equipment manufacturers like Lucent, Ericson, and Nortel Networks. That is all bad debt now, which causes these manufacturers to cut costs by (among other things) reducing R&D expendatures. This means advances in the industry will not come as fast as they were, but they will still come eventually.
These manufacturers are still willing to finance the lighting of fiber networks, as needed, (in that such activity requires purchase of multiplexing equipment and switches, the sale of which these companies have financed for the past two decades).
When it comes down to it, the industry is returning to a pre-tech-bouble state, not dying completely. The determination of which companies will still be standing will be which are able to adapt quickly enough. Unfortunately, this is made more difficult with a Wall Street backlash against the telecom industry, but such things happen and will be overcome in time.
There will be no shortage because the market is capable of meeting demand. The required financing will become available because it is the only way the manufacturers will stay in business.
In closing, let me just say that I always get my technology news from McPaper because after all they're known for their technical expertise and research prowess.
--CTH
--Got Lists? | Top 95 Star Wars Line
Stop and think about how much bandwidth costs.
It's a lot more than routers and fiber. In fact, compared with the costs of upkeep and support, the infrastructure is almost negligible. That's why there's so much infrastructure already built, but so little utilized.
Why not use it all? Because people are not yet willing to pay for what they get. Standard business practice is to charge the customer 5x the actual cost of a product or a service.
Broadband service is so desparately trying to compete with the low cost of dialup, that it's not making the margin it needs. Of course it doesn't scale linearly, as a 128kb DSL connection doesn't cost 32x a 4kb dialup, but a 128kb DSL connection for only 2x or 3x the cost of that dialup sure isn't making the DSL provider the same margin as the dialup gives the dialup provider.
Bandwidth is expensive, we want -- no, we demand 100% uptime, no slowdowns, this, that, etc. Until people are willing to pay the true cost of this service, none of the greedy Telcos are going to make any money out of this, and will have no motivation to build new infrastructure, make new plans available.
You are quite accurate. A true shortage only occurs when regulation or some similar non market force changes output. An excellent example would be the gasoline shortage in the 70s. It only occured because of price ceilings placed on the price of gas. No ceilings no shortage, less would be sold at the higher price, but there wouldn't be a shortage. This is just markets responding to stimulii. First everyone wanted to invest in telecommunications, debt and equity markets were happy to loan and invest in anything related to telecommunications. Now noone wants to invest in telecom at all. Soon the extra investment will be used, as demand increases and the price increases, and as prise increases, investment will return. Markets are pretty effective at things like this. What created the problem was the flood of investment following the removal of regulation on the industry. Not to say that the regulation was good, but that it being there slowly built up an imbalance and the removeal of the impediment to balanced markets caused such rapid movement that it created an imbalance in the other direction.
Degaussing scares the bad magnetism out of the monitor and fills it with good karma.
Nobody is forced to subsidize anyone, they are just forced to resell at fair prices. It is amazing how much people bash this mild antimonopoly provision.
The last mile problem is not being solved because cable and dsl are much less popular than everyone thought they would be.
People are not flocking to cable and DSL and communication companies are seeing little point in releasing something better.
I've been involved with a WAN project, based on in ground fiber. At the time, people started with Cisco based fiber routers, because of networkability - these cost $30k on both ends for a T3 line worth of bandwidth. I personally think it was a horrible idea and would far prefer 100Base-FX on both ends - 1/20th the price (I estimate), and greater functionality for our situation, which is mainly data traffic.
Phone companies simply can't use ethernet - it doesn't have QOS, or guaranteed packet delivery times, dedicated channels, etc... These are what telcoms want/need, and the technology that does it costs a ton. For your joe blow data network, ethenet makes sense - telcom is a different issue entirely.
BBK
fiber is dirt cheap. it costs approx $100.00 for every fusion splice to be made (that is what we pay the contractor.. $100.00 per splice) and it costs $6500.00 to light up 1 (that's ONE) 50km fiber to a paltry 100mbps Full duplex. Bringing up a 1000mbps linx is 5 times that price and bringing up a multiplexed 5000 mbps (that's 5 of the above multiplexed over that one single fiber over a tiny 50 km distance) costs $100,000.00 in all supporting equipment. I know as I just lit one up (I didnt pay for the fusion splicing... I did them myself! Dang cool machine to play with!) We had ran a 16 fiber run laid using directional boring equipment on utility right of way. The cost of the equipment to light it up was at least 20 times the price of the fiber,and the cost of shoving it into the ground and the 256 splices made (I did them so that cost was gone... except for the $50,000.00 for the fusion splicer but that's a long term investment.)
It makes you sick holding onto a piece of equipment that is no larger than a paperback book and has less than a handful of electronics on it and knowing that you were bent over and made to bark like a dog for the fiber module company to a tune of $3500.00 each for the cheap stuff.. (single mode is the way to go, cheaper and you can get decent distances with it compared to multimode.) Now if you are a large telco or company that has to overpay your "fiber technical engineers" to do what I did last week, you'd spend another $15K in salaries...
Oh BTW, if you install 16 fibers... you light up 8 of them MAX... you never just start using your spares, that's a big no-no.
Do not look at laser with remaining good eye.
All I want is 1mbit up and down, and no restrictions on what services I can use (ie. servers). Is it too hard for these broadband 'providers' to offer something so simple? Is it so hard?
No, it's not hard, in fact it's pretty easy, though it takes a few weeks to set up.
If you want to run servers, you need to get a real ISP.
You call up bandwidth.com. You say, Hi, I'm Joe Blow and I want a megabit up and down. They will say "great, a T1 is just what you need".
The helpful staff will then compile a list of packages you can buy, ranging from about $800-1600/month for 1.54 Mbits/sec in most areas.
You see friend, in the real world, people have to pay for bandwidth. Your ISP has to do the above when they buy bandwidth, and logically, they can't sell it to you for less than they pay for it, otherwise they won't be in business very long.
They may get a little discount if they can afford a T3, but that's still $20,000-$30,000 a month, and is about the equivalent to 30 T1s.
Are all you people that whine that they want 1Mbit up and down, unrestricted, for $30/month math challenged, or just stupid?
I've had enough abrasive sigs. Kittens are cute and fuzzy.
For every $1 spent to put a fiber in the ground, (Score:1) a company has to spend $20 to attach it to all the equipment, configure it and turn it on.
Sorry, but my bogosity meter is banging against the peg on this one. Anybody have any real figures on this?
That's a reasonable number.
When you're burying the fibers it costs you a bunch to dig a trench. That's because the trench goes around the whole continental USA, plus a few cross-paths. Every major city has to end up with two trenches connecting it to two other cities (so if the fiber going one way gets cut the signals can be rerouted to go around and come in the other way).
It costs you a LOT to dig that trench. If you put the minimum two fibers you need in it, when you need more you'd have to DIG ANOTHER TRENCH. So you put in a BUNCH of fibers. You don't want to dig another trench for a century or so.
The extra individual fibers cost next to nothing compared to the trench, even after you include the cost of the splices. The total cost is still enormous. But once you divide the cost of the trench among all those fibers the cost PER PAIR OF FIBERS is small - approaching the cost of the fiber itself.
But now you go to light them up. This means a box at every city or town along the way where you want to hook up, plus maybe several repeaters in access boxes EVERY MILE along the fiber (the spacing and cost varying according to the type of fiber and what type of signal is going through it - but it's not cheap). For packet switching you need maybe a million bux worth of box in each town for each fiber pair you light up (though lighting extra fibers can be done for maybe a quarter of that by adding cards to existing boxes.) For phone calls and raw pipes, maybe an eighth of that (though that just lights and protection-switches the fibers and hooks them to the local signals with a few fat pipes). And a rule of thumb is that the rack space for the box costs more than the box.
20-to-1? Might be low.
Of course the repeaters and boxes both get better roughly by Moore's law. So if you can hold off lighting a fiber for 18 months, it costs you half as much. So of COURSE you don't buy boxes and repeaters until you're ready to light the fibers.
So the analysts looked at all the dark fibers, forget about all the non-installed boxes, and started prattling about a "bandwidth glut" - as if nobody would buy another box for 20 years.
Meanwhile the tellcos had bought enough boxes to light their first fibers (plus a couple spares). So there was a dip in box purchases while they switched to finding customers to use the bandwidth on the first lit fibers. (One set of fibers has a LOT of capacity, so purchases are lumpy.) And the tellcos got into a price war to get those customers - with the little guys getting squeezed out by the old former monopolies with their buckets of cash.
So the investors paniced. And investment in "telecom" dried up. And without investment the squeezing proved fatal for the new little guys. And without the little guys nipping at their heels the big guys started taking their time (though they're still installing and still filling up their current boxes - plus the ones they got for near-free from the dying little guys). And without investment nobody bought more of ANYTHING, creating a self-fulfilling prophecy and crashing the stock of anything tarred with the name "telecom".
But even with the big guys dragging their feet the boxes ARE getting filled up, just like the article says. Some time soon the surviving companies will need to buy more. And the surviving suppliers will have less competition and ZERO inventory (having long since switched to build-to-order as a belt-tightening measure, and also sold off all their pre-made stuff at fire-sale prices).
After all - do YOU have broadband yet? On slashdot, probaby yes. But don't you know somebody who doesn't, but wants it and can't yet get it?
They're STILL doing the INITIAL buildout - and the "broadband" pipes are still tiny compared to what people WOULD buy if it were available and the price were right. Only a couple percent of the country has broadband. A LOT more people want it, and will get it as soon as it's available. (As of a couple months ago the tellcos expected to install as much DSL this year alone as had been installed since it was invented.)
Seems to me the article is right on. The current "telecom crash" is at least partly due to a panic reaction by investors, and the result will be a bandwidth crunch as the boxes fill up and there is a sudden need to buy more - creating a seller's market for the surviving suppliers.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
ATM/Sonet/SDH switching equipment is damned expensive. Fully decked out switches can easily cost $250,000 - $1,000,000+ depending on port density and speed (OC3 - OC192).
You also have to deal with what TYPE of fiber is in the ground already. Older stuff can't support the big DWDM equipment. Zero-dispersion fiber was popular until we cranked up the speeds and found that certain problems occurring about OC-48 result in exponential loss (no data making it thru). Newer fiber is dispersion-shifted, with erbium or another rare-earth doping.
This is why ATM never caught on in the LAN, even with cheap OC-3 cards -- switches cost a friggin' fortune!
Learning HOW to think is more important than learning WHAT to think.
You've hit the nail on the head: the telcos don't understand the business that they are really in.
Today, customers of the telcos want to moves bits from physical location A to physical location B. If a telco is concentrating on any goal other than this, they are doomed: dumb analog services (e.g. call waiting, message boxes, redial) are just milking a dying cow. Telcos must understand that the end-points now have the smarts - they should just be in the business of moving bits. That means leasing dark fiber, providing redundant circuits, whatever it takes.
Telcos are selling a commodity: work on that, you stay alive, forget it, you die.
You know, first of all, most modern homes (within 10 years) are 'prewired' for a second phone line, right? They just switch which wired pair your using on whatever phones you want... Not to mention they charge you near $100 for the 'install'...
Second, Not all of us have the option of using DSL... Verizon (who became the local phone company) sees my town (of 5000 people) as a non-market so they refuse to offer anything beyond phone lines & super expensive ISDN 56k to businesses... The local switching center is state of the art & they could add anything they wanted simply by tossing in the controller box... 99% of the town could be served from that one switching station for DSL...
we are all invisible unless we choose otherwise
Never sign a contract for more than one year on a leased line, or 6 months for colo.
I ran an ISP for about three years, until around mid '99. It's not my main business any more, but I still have a couple of hosting clients for high bandwidth sites. the ISP business is TOUGH. The competition is insane, so the approach most ISPs take these days is to advertise really high prices while offering competitive rates only to those who haggle and know how to shop around.
Here's what I've been spending, year by year on Internet service. I've switched providers several times over the years due to changing needs wrt colo vs leased line, and varying costs. I've now been with Hurricane Electric for over a year. They are outstandanding, but you'll have to haggle to get a good price.
1997-1998 - 3 bonded centrex ISDN lines from Brainstorm, 384Kbps: $750/mo
1999 - shelf and 1Mbps at Above.net plus a ptp T1: $2000 + $450/mo
2000 - shelf and 1.5 Mbps at maxim.net: $700/mo
2001 - ptp t1 to Hurricane: $650 ISP, $350 XO for the line
2002 - shelf at Hurricane and 2MBPS: $650. PTP T1 to my shelf: $350
As you can see, over the years the cost of connectivity has fallen from $1822/mbps to about $500. That's not just per MBPS, I'm talking about a complete package - remote connectivity for 1-2MBPS upstream.
The cost of installing fiber is still outrageous, but the fluctuations in demand have resulted in a surplus of strands in the ground. I've coordinated fiber installations before - trust me it's a BIG deal. Trenching, conduit, permits, dealing with the city and the fscking retarded telcos. It's no fun, it's EXPENSIVE, and it can take upwards of three months just to get 100 yards of fiber in the ground. But now that the fiber is there, ISPs and telcos can start using it as soon as there's demand, just by connecting the needed equipment.
Also don't forget that the same strands can usually be used for OC3, OC12, GigE, etc. So it's not just that there are unused strands in the ground, there is also a ton of equipment that can be upgraded to increase the capacity of the strands we're using.
Bandwidth costs still have a long way to fall!
Furthermore, wasn't the 10GbE or whatever the acronym for 10Gig ethernet standard is, supposed to be finalized like last month?
And if I recall, part of the standard was that 10GbE devices would be specifically geared for use with dark fiber. Am I wrong here? Anybody on the IEEE standards committee like to comment on this?
And finally, as an overseas American in Asia, I get the distinct impression that consumer telecoms in the US are falling behind what's going on in other parts of the world and I just can't believe that's going to continue. We've got no bandwidth problems where I'm at and we've got lots of options. Even the monopoly networks offer really cheap and fast DSL and then there's the competition offering fiber to the desktop or CAT5 into your aprtment for like twenty bucks a month. I can't imagine the US really has it so bad if we've got it the way we do here.