Comcast bandwidth capping now official

Oh it has nothing to do with faith. Faith is belief in that which is unseen and unproven.

It has to do with what Verizon is doing with their infrastructure and what the cable companies are not doing.

 

You can't limit the situation's scope to prove your point. You've stated numerous times that you have confidence in them continuing the service... sorry to tell you but that's faith.

 

As I've noted earlier, upgrading the available bandwidth within a fiber cable via WDM technologies is FAR easier than ripping out old and outmoded infrastructure and installing new. What do you think these corporations have been using to upgrade their backbones for the past decade or so?

One requires a few million dollars each time a major upgrade is required. The other requires a few billion. The orders of magnitude difference in cost, not to mention the fact that that has been one of the reasons why Verizon undertook this infrastructure upgrade in the first place, leads to the fact that this is their stated business goal.

They are uniquely positioned in the US to ramp up their infrastructure easily and quickly to meet the bandwidth demands of tomorrow. Cable and DSL companies? Not so much. If they could ramp up bandwidth demands easily, they would. Why? Because it would significantly increase the value of their product and allow them to offer many more services to the end customer.

Just look at what's happened with AT&T and U-Verse, for example. They're throttling bandwidth to end users in order to provide a base level of QoS. A single HDTV stream on a 10mb DSL connection takes up nearly the entire available bandwidth. So if you're watching TV, you're SOL on your downloading. If they could upgrade their bandwidth easily they wouldn't resort to such measures.

 

http://www.law.com/jsp/legaltechnology/pubArticleLT.jsp?id=1202422769174

There are many more such suits. We can't let people decide they want to pay for something and have good connections, oh no that would be terrible. Instead we will charge them more for worse service.

 

And those are infinite? What sort of granularity is available here? Can the same hardware produce every single wavelength of light at the flick of a button? What technology is being used to create this light? Let's move beyond the pedant's point that vast is not infinite, I still think the bandwidth is a far cry from being even nearly infinitely expandable.

You're being terribly optimistic about this. Not to mention that you're completely waving over the actual cost to provide this bandwidth. It's not free, and upgrading it doesn't magically make it cheaper -- what it does it shift the available bandwidth available per consumer. Companies will oversell; when they hit a bottleneck, there's two places to go: they can expand, or they can decide to throttle connections. I don't think that believing in the essential goodness of a for-profit company is wise, for one, but for another I don't think bandwidth can expand forever.

 

I'll say the same thing I've said to skyring. How is it that global companies have upgraded their fiber backbone bandwidth so easily over the past decade? WDM technology.

And yes the possibilities, as far as we're concerned, are infinite. Why? Because each wavelength of light can carry enormous levels of bandwidth, and the current technologies allow companies to compress hundreds of wavelengths into a single fiber.

In the end, the capacity upgrades are handled at the routing offices of the 'net companies. Not as they are right now for DSL and cable companies, which typically require running new wiring as well. It's far less expensive and time consuming to do capacity upgrades in this manner. And it's how companies have been doing it for years on their backbone.

Verizon has simply brought this to the home.

As for the cost to provide the bandwidth, it's a matter of the equipment that does the actual switching at the router level and splitting of light waves. As the cost of that equipment falls, as it always does with new computer hardware, so does the cost of providing bandwidth.

Oh and btw, to an earlier point you brought up, the reason why T1s, T3s, et al are far more expensive is because the phone company runs a dedicated circuit to your home whereas with other technologies you're sharing with everyone else at the CO. They are also guaranteeing a level of service that you won't get with other internet connections.

That, and that alone, is why those lines cost far more to run.

 

Please, I need a source for this. Just 'look up WDM' isn't sufficient. I need analysis that projects WDM growth as effectively infinite. Just going 'but it's WDM, it's different wavelengths of light'! doesn't do it.

It's not zero, though. It's not cheap, either, I bet. And you can't argue that equipment will continue to get cheaper forever any more than you can say that bandwidth will continue to get greater forever. It's very nice to believe it, but belief doesn't have much sway over the laws of physics.

Source, again. That's a strong statement, and I have plenty of different reasons, which you are saying don't factor into the price. Don't forget, these connections also provide far more bandwidth than any ISP will give you. What you talk about service, that's exactly my point. Service means not just a support line, it means that they guarantee that a high percentage of the bandwidth you've bought is available. THat's what I'm saying when I say that with a dedicated line, you get all the bandwidth the connection actually says (as opposed to a Verizon's '50 mbit' lines).

A 50Mbit connection can do what, >10Tbyte a month, nonstop? Even a T1, which is only 1.5Mbit will give you >400Gb a month, nonstop. Do you think Verizon would be cool if people started consuming an OC-1's worth of bandwidth, each?

Like Skrying said, you're operating off hope.

 

You can google WDM obonicus. It's not that hard, and the technology has been around for over a decade.

If you don't want to type in the searches for yourself, here you go.

Wavelength-division Multiplexing: http://www.google.com/search?hl=en&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=nR1&q=Wavelength-division+Multiplexing&btnG=Search

Dense Wavelength-division Multiplexing: http://www.google.com/search?hl=en&...+Wavelength+division+Multiplexing&btnG=Search

But if you want something without having to educate yourself in general, try these:

http://www.commsplace.com/Knowledge/ITcs/html/tutorials/fibre_optics/introduction_fibre_optics.htm

http://www.glenair.com/qwikconnect/vol6num3/coverstory.htm

Whoever argued that the cost is zero? Certainly not I. However, if you want to argue that computer equipment won't continue to get cheaper and cheaper, then I have about 50 years of the reduction of the IC to prove you incorrect.

Wasn't it only 20 years ago that people were saying the microprocessor would run out of steam by now?

So unless you're predicting the end of the computer age, bandwidth will continue to get cheaper and cheaper and cheaper.

You want a source? Try any phone company running a T-line. Ask them why it's so expensive. I've had a T1 run before.

It's actually 16 Terabytes per month of data usage if the download connection were going nonstop. Sorry, but even with bittorrent you'll find it impossible to max out an internet connection in that fashion. There isn't any data large enough to sustain that kind of transfer rate today.

Second, what was the backbone of the internet 20 years ago? ~1-2Mbit you say? That's a slow home connection these days.

By the time everyone has a 50mbit connection in the US, i.e., another 10-20 years at the current rate , what do you think the backbone will be then?

One only needs to look at the history of bandwidth growth to project out with some semblance of understanding.

 

Perhaps Usenet. If you opted to run a NNTP server at your home with direct connect to another feed and pulled every single message in every single group.

 

I'm not arguing about the existence of WDM, I'm questioning the notion of 'infinite' bandwidth. Infinite, for starters, is a ridiculous buzzword when used in the context of anything real. I can totally accept that it offers a lot of bandwidth, but I will take exception to the notion that this infrastructure is infinitely expandable.

http://books.google.com.br/books?id...&hl=en&sa=X&oi=book_result&resnum=9&ct=result

Also, I realize how light operates. I'm questioning, though, how many actual wavelengths can you use? How are they generated? How many wavelengths can you detect reliably and cheaply? There are physical questions behind this that you can't just handwave away with 'but it's SCIENCE!'. Can you use the entire electromagnetic spectrum? Or even just what we call 'light'? I know the answer to some of these, but then I'm the one skeptical about promises of infinite bandwidth.

I love these predictions. 'Everything will be fine because it's always been fine in the past, and it still is fine'. Even Intel doesn't think it can go on the technology it's on forever. Will carbon nanotubes, a completely diferent be the replacement? Well, there's certainly no economy of scale for that technology in place yet, and I have my doubts about it being ready in 2016, when Intel predicts it'll start having tunneling effects in its gates. Will there be some other technology? Maybe, but what? Will it have the economy of scale we have with CMOS? But hey, we have 50 years of history to prove that the future will be perfectly fine! Sounds perfectly reasonable to me.

You do realize that Moore's Law isn't an actual law of the universe, right? Hitting a plateau would not mean the end of the computer age, either. Cars spent decades with mostly incremental improvements, it wasn't the end of automobiles.

So, yeah, I don't think things will improve indefinitely. I think there's physical limitations to all things, and I, being pessimist, do believe we'll see them sooner than later, but I'm just guessing now.

And a telephone company will tell me 'yes, all the cost of renting a T-1 is service, none of it is the actual bandwidth'? That's what you're saying, and that's what's not true.

It's certainly not impossible, otherwise there'd be no use for OC-3 etc. It's not easy, I'll give you that. (Notice how you're essentially arguing RudeCurve's earlier point that an arbitrary amount of bandwidth is impossible to fill.)

But you seem to be missing my point, and I'm not sure if it's on purpose: what you're paying for is nowhere near an actual 50Mbit connection. Period. That's their business model; as soon as your consumption starts to exceed whatever portion of the bandwidth they actually assigned to you, the service providers will begin to become unhappy.

Much higher, naturally. Will your connection stop at 50Mbit for 10-20 years, though?

 

1) I stated that fiber has a limitless carrying capacity. You asked for evidence of this and I provided links to WDM, DWDM, and two links in particular that state precisely that fiber's carrying capacity is infinite.

You being a skeptic is one thing. Talking about the end of the computer age and all that stuff is another. Yours is a circular logic argument that fails under its own weight in this circumstance.

2) With respect to Moore's law, it says that transistor density will double every 18 months, and yet that's already slowed down from what it used to be. Now instead of trying to cram into a smaller and tighter space, engineers are building wider in the form of multiple cores. That's precisely what they've done with WDM technologies and light in a fiber.

That said, this is also where carbon nanotubes, quantum computing, et al come into play. Those technologies are dealing on scales that are far smaller than anything we're dealing with today and will be ready to go within a few decades with regard to mass production. Technology is always pushing forward, and there are always doom sayers at every step of the way who say "no no no".

That should not be a basis for an argument on your part, however.

3) T-line's are a dying breed for precisely the reason I've stated. It's the service and reliability that are the costs of a T-line, not the bandwidth. My connection has been up 98% of the time and provides FAR superior throughput on upload and download than a T1, for example, yet it costs me hundreds of dollars less than a T1.

What T1 service provides you is 99.999999% guaranteed uptime and a dedicated circuit to the CO meaning that it is impossible for them to oversubscribe you to the point where you'd ever see less than the bandwidth that you've paid for.

Given that most DSL companies give the same kind of performance levels, it's moot for anyone other than a business that absolutely needs to have those service level guarantees. For small businesses such as the one I used to be part of for 6 1/2 years, DSL/Cable/Fiber has the SLG that is required.

Additionally, 10 years ago the bandwidth provided by T-lines was exorbitantly expensive to provide. Today, not so much. A T-line 10 years ago cost my company $4,000 a month. Now? You'd be hard pressed to pay more than $300-$400.

And that's the point.

4) With regard to the impossibility of filling a 50mbit connection and comparing it to an OC-3, you do realize, hopefully, that one is an internet backbone and is carrying traffic from all over the world while the other is an end node that is used by 2 people?

That comparison is utterly absurd on the face of it.

And as for RudeCurve's assertion, the difference between 250GB/month and 16 Terabytes/month is greater than an order of magnitude! How can you POSSIBLY attempt to compare the two with a straight face?

Additionally, you clearly don't understand Verizon's FIOS business approach. The stated business purpose of their service is for you to use bandwidth as much as possible. They encourage you to download HD movies, to use VOIP and Video streaming, to use Bittorrent, etc. I should know because I've called them about it, asking about caps and whatnot. They BUILT FIOS for this purpose. The ONLY reasons companies are starting to cap is because they don't want to invest in their infrastructure. Instead they want to squeeze their current infrastructure for everything they can before having to lay out a dime.

Verizon went in the completely opposite direction and decided that instead of capping internet usage they would future proof their network. Sorry, but you're thinking in the old infrastructure way of things, and that has nothing to do with how Verizon has changed their business model to the end user.

Last, the fastest home connection you could get 20 years ago was a 9600 baud - 14.4 modem, if you were lucky. In other words, the backbone was about 150-220 times faster than the fastest internet connection available then. Today? The backbones are generally in the range of 40Gbit/sec. Considering my connection is the fastest available in the states currently, that's about 819x faster.

In other words, backbone bandwidth is outstripping bandwidth available to end users by several times over the past 20 years. Why? Because the companies have implemented WDM technologies on the backbones and have only recently begun that process to the connections to the home. Hell, they've already got terabit speeds going across some parts of the internet backbone as well, so in truth the carrying capacity far outstrips what can be sent to the home today. And it's only accelerating over fiber.

Quite frankly, you're just thinking as the cable and dsl companies are, i.e., that we'll always be stuck with outdated technology that cannot scale. Fiber scales easily and quickly. Copper line transmission? Not anywhere close.

 

I have to laugh every time I see a Comcrap commercial advertising "the nation's largest fiber optic network"

 

Japan is now getting some 1Gbps fiber-optic service for about US$56.50. (Press Release).

KDDI Corp will launch a fiber-optic communications service with upload and download speeds each of up to one gigabit per second on Oct 1. The new service will target people living in single-family homes and low-rise apartment buildings. The traffic speeds will be the fastest in eastern Japan, up drastically from the current 100 megabits per second.

KDDI will charge 5,985 yen in basic monthly fees for Internet and telephone services, down 1,155 yen from the current price, if a user subscribes for two consecutive years.

 

I think it's perfectly ok for an ISP to provide only metered billing if there is healthy competition in the market. But this piece of reasoning is just hilarious: "traffic on our backbone is growing 60 percent per year, but our revenue is not." Well, duh! Neither is the amount they need to invest in their backbone...

 

First of all, your links said that the BW is essentially infinite. More importantly, however, is that fibres aren't the only thing you need for communications (and I don't think most of the infrastructure uses single-mode fibre anyway).

 

Thing is, it only ever works one way. Try asking the ISP if you will have to pay $0 if you go away for a week and don't use the internet. Ask them if you get a discount for not using their mail/web/new servers. Won't happen, because they only want to raise the price for extras that used to be free, rather than develop a truly modular billing model where people pay for what they use.

What it boils down is getting more money out of the customer by offering less, and hoping to not have to spend money investing in infrastructure whilst still increasing window dressing speed for marketing purposes.

 

I can freeze my account, if I am away. But only up to a month in a year which is sucks but understandable.

Anyway I think the ISPs are just preparing for HD content. Making sure it doesn't happen or if it did, they'll at least get alot of money for it. In short term is a win for them. I am suprised they didn't do it sooner.

 

A quibble not even worth mentioning.

Naturally. The point is that upgrading a fiber network to the home is FAR easier and cost effective than the old copper and coax networks.