Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

Bandwidth glut sends packets further for less money

By imrdkl in Internet
Wed Apr 24, 2002 at 11:35:12 AM EST
Tags: News (all tags)

Telegeography, a company which produces annual statistics and anaylsis of the International Telecom industry, has recently released their latest reports. According to a short review from the Nando Times, one of their reports discusses a vast overabundance of bandwidth in the world, which has driven prices below wholesale, causing "hyper-deflation" in the data transmission market.

There are five distinct publications released for sale by Telegeography, Nando has cited the Terrestrial Bandwidth Report (See the Executive Summary in PDF) which states that while only one to two percent of potential bandwidth is active, the amount of lit1 capacity is staggering. New York, for example, currently has 23.5 Tbps running through the city destined for domestic and international networks. "Secondary" cities such as Cleveland also have terabit capacities available, although far from utilized.

This glut has affected prices to the point where providers are having to account for it, according to the report.

Many carriers had counted on falling prices as part of their business strategies, although none was prepared for the rapid declines witnessed in the past few years.
While the declines have been consistent, the report states, the actual prices have varied widely, with the highest price for cross-country leases sometimes being four times the lowest price.

The report concludes that while construction has practically ceased in long fiber lines, there's not alot of upward pressure, or increased consumption likely for global bandwidth. Small and mid-range carriers and providers should be able to look forward to reasonable prices on the major trunks for some time to come. Perhaps this fact will even help to motivate some more work on the last mile problem. With major technology companies going bust, the private consumer may be the last best hope to sell some of this bandwidth.

1. lit capacity is the amount of fiber which is actually in use, as opposed to "dark" fiber, which has been laid but is not carrying signal or data


Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure


Favorite Glut
o bandwidth 44%
o office furniture 10%
o grains/dairy products 8%
o chocolate 38%

Votes: 50
Results | Other Polls

Related Links
o Telegeogra phy
o short review
o Terrestria l Bandwidth Report
o Executive Summary
o last mile problem
o Also by imrdkl

Display: Sort:
Bandwidth glut sends packets further for less money | 17 comments (10 topical, 7 editorial, 0 hidden)
Explanation? (3.50 / 4) (#7)
by epcraig on Wed Apr 24, 2002 at 09:30:52 AM EST

How could you resist referring to the (purportedly) two largest builders and holders of dark fiber, Enron and Global Crossing?
There is no EugeneFreeNet.org, there is an efn.org
Bandwidth not fiber (4.00 / 2) (#15)
by glitchvern on Wed Apr 24, 2002 at 05:05:10 PM EST

Enron didn't actually own all the fiber or bandwith they were trying to sell. They were just acting as a middle man. Global Crossing is mentioned in the article he linked to. In addition he is talking about the amount of bandwith or lit fiber available. Dark fiber is fiber that has not yet been lit. It requires some expensive electronics on the ends of it and some more in the middle. With a bandwith glut there is little reason to light any of the dark fiber.
Programmers are like Mogwai, they hate bright light, direct sunlight is rumoured to kill them.
[ Parent ]
Uneven distribution (4.42 / 7) (#8)
by Chancellor Martok on Wed Apr 24, 2002 at 10:23:31 AM EST

While there may be a heap of bandwidth out there, even an excessive amount, and it might be cheap overall, it's certainly doesn't seem to be distributed very well.

Ask anyone in Australia, and they'll probably tell you the 2 words 'bandwidth' and 'cheap' are mutually exclusive.

In our case, it really is that the 'last mile' (or 'last kilometre' here is a bigger problem here than most places in the world... the same old lack of economies of scale in such a vast place.


Household broadband connections aren't likely to be cheap, or particularly fast for quite a while here, and out in the country, they're unlikely to see any bandwidth at all, let alone overabundant cheap bandwidth.

Chancellor Martok  in Sydney, Australia
"Castrate instead. That can surely rehabilitate. I did it volunatrily, and my grades went up!"  -- Sen

darn those Telcos (3.75 / 4) (#10)
by deadplant on Wed Apr 24, 2002 at 12:47:28 PM EST

Darn those Telcos to heck. They've got all these great backbone networks but they continue to strangle the last mile.

I wish the big phone companies had been a little more daring and gone straight for fiber to the home instead of investing all that money in DSL. I guess it was partially because they wanted to protect the super-high fees they charge for T1 services. The telcos business model is also to blame, they're accustomed to building expensive infrastructure and then slowly paying for it over decades. So when an opportunity comes along to make all their existing infrastructure obsolete they have to hold it back.

The people in small nations/towns like Isreal have all the luck. They don't have ISPs burdened by massive 'white elephant' networks.

Please, sombody bring me fiber! please.. please... wimper..

Unfortunately they have (4.33 / 3) (#12)
by El Volio on Wed Apr 24, 2002 at 01:36:35 PM EST

DISCLAIMER: I work for one of the "big phone companies", but I work in the infosec department of the International arm and don't have anything to do with the telco stuff other than the fact that it affects my 401(k). Please don't construe anything I say as representative of my or any other company: this is entirely in my private capacity. I'm too low-echelon to be able to speak for them anyway...

On the one hand, I agree: fiber to the curb would be awesome. I can just imagine having true 10Mb (or even 100Mb) Ethernet to my home; it would be great. That said, there seem to be a lot of complaints about "if the phone companies weren't so interested in making money, we could have a great telecom infrastructure!" That's not really true: if the telco companies go under or get to the verge of bankruptcy, they won't be able to build and maintain those networks. The demand has to be there.

That's one reason why we've seen so many companies go under, like the dotcom crash: they'd build up these huge infrastructures but have no real way to make money off of it.

And believe me, there's plenty of financial problems in the telcos as it is. Seen stock prices lately? It's not pretty. I'd love to see the same things, but in a capitalist society there has to be profit there, and frankly the telcos don't make much profit as it is -- POTS has virtually zero margin, all the growth and profit is in wireless and data.

[ Parent ]

Why isn't it used? (none / 0) (#11)
by fortytwo on Wed Apr 24, 2002 at 01:01:00 PM EST

There should be loadbalancing, so packets going NH->NY can take a detour for Austrailia if all other bandwith is used...

This is used (3.00 / 1) (#13)
by steveftoth on Wed Apr 24, 2002 at 02:43:17 PM EST

This is why they invented routing. In fact, this does happen a lot, it's just that once you get to the 'real' internet it's just better to wait for a really fast link ( for a short time ) rather then try to go around. Most of the latency or 'slowness' of a connection is always in the 'last mile' portion of your link to the internet. Take your 'high speed' access for example. A packet probably has to go through about 5 - 10 hops before it gets to the destination. Also, most backbones do not have 20-30 links coming in/out of them, so a large part of the time, there is no option.

[ Parent ]
It's not that simple. (4.00 / 2) (#14)
by Kyle on Wed Apr 24, 2002 at 04:06:46 PM EST

I worked at a small to mid-sized ISP about two years ago.

On a LAN-sized network, the routing protocols (such as OSPF) are capable of taking congestion and link quality into account when choosing routes.

On the big bad Internet, where a router must track constant changes in tens of thousands of routes (on each connection it has), there's already a ton of information and no time to think about how good the routes are. The routing protocol of choice (BGP4) only cares whether a route is there or not. I remember hearing of some work to get a BGP that considers link quality, but it's a hard problem.

Our ISP had three upstream connections, and we wanted our incoming traffic to balance on them to get best performance. Unfortunately, that's just not possible. We can't control which of the three some router in Japan uses to get to us, and we can't ask them to pass to each other if their links to us are saturated. Balancing our outgoing traffic is easy since we control it, but you can't get optimum behavior everywhere all the time.

[ Parent ]

Traffic Engineering (3.00 / 1) (#17)
by supine on Tue Apr 30, 2002 at 06:34:40 AM EST

Traffic engineering is a way of shaping traffic flows to best suit the underlying infrastructure.

There are many ways to achieve this (some routing protocols have built in "link costs" which can incorporate performance metrics; add on controls eg. MPLS enable tweaking of routing protocols to meet requirements) but it is hard when the traffic flow in question traverses multiple backbones and is therefore subject to different policies and varying traffic engineering decisions made on each backbone.

It is much easier to route around congestion when you own the whole network.


"No GUI for you! Use lynx!!!, Come back, One year!" -- /avant
[ Parent ]

Predicted problem (4.00 / 1) (#16)
by Graymalkin on Fri Apr 26, 2002 at 03:47:50 PM EST

I think saying carriers had not expected a rapid decline is prices is a specious assumption. Carriers knew how fast the technology to increase capacity was outstripping the demand for that capacity. It wasn't like one day everybody had a ton of OC-48 links and stood around scratching their head as to how they all got there. The only expectations there weren't met were the ones in terms of demand. That demand didn't material because local carriers weren't offering affordable services to take advantage of the backbone bandwidth there was now available. Instead of fat pipes going to homes and businesses meeting their data needs they were focusing on extending the capabilities of the wiring they already had in place. They can't exactly be demonized for this because it is probably the most economical option they had available.

For the most part DSL is an interim technology, it bridges the bandwidth gab between widespread POTS and widespread fiber. There's also the nagging problem that getting a high bandwidth connection from your local carrier in any form costs you oodles of money. It would cost you less money in the long run to light some dark fiber between two locations than go through various local carriers. With the amount of raw capacity on IXCs have available it is a shame that it costs so much to get at it locally.

Bandwidth glut sends packets further for less money | 17 comments (10 topical, 7 editorial, 0 hidden)
Display: Sort:


All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!