Comcast bandwidth capping now official

That's the problem. You have to be almost right on the exchange, and then you'll max at 20-30 megs, but the drop-off over distance is terrible. Combined with some of the convoluted routes for phone lines, crappy aluminium wires put in tens of years ago, faulty in-house phone extensions, etc. It's all problems related to levering broadband in on top of an aging hardware layer that was never meant for high-speed internet access. DSL basically treats phonelines as a bit of network with a really crappy bit at the end and that's pretty much what it is.

Even my cable co-ax from 70 metres down the street is strained to max signal fidelity because it's only RG7 and will have to be replaced to RG9 if it gets any worse. A bit of fibre going from my local cab to my house would be able to handle much higher speeds and much more reliably.

 

Twisted pair copper via phone lines is pretty much limited to around 100-200Mbit/sec over short distances with current technology. That's talking on the order of a couple hundred feet at most. That would buy us maybe a decade.

Fiber is the only solution worth rolling out in a "moon-shot" style deployment. The bandwidth capable of being carried is pretty much limitless.

 

That sounds more logical to me. That's one of the reasons why I have a C2D running at 2.5GHz instead of a C2Q running at 3GHz.

 

Right. In the meantime where FIOS is competing directly with U-Verse, AT&T is losing customers left and right.

And on top of that, AT&T is unable to deliver the HD-quality TV that you can get, even with cable, because there isn't enough available bandwidth in 8-10mbit DSL.

Lastly, the comparison between C2D and C2Q is in no way the same with respect to internet infrastructure.

 

Here's the whole article if you want to read it. Looks like there's still ongoing debate whether or not Version's FIOS was the right decision. You seem to think it was and you know where I stand.

http://www.nytimes.com/2008/08/19/technology/19fios.html?_r=1&oref=slogin

 

That quote is just BS. Cost of deployment is all about the time and cost to dig up roads and lay cable. However, if the roads are already laid, all you need to do is a fibre pull from the cabinet down the same plastic pipes that already carry the coax. That takes a couple of guys 30 minutes tops, and then an engineer to come in later and do any internal connecting/deployment of your new fibre service.

The cost isn't about laying a cable to your door (which you usually pay for at least in part when you sign up), it's about laying the pipes down the street so you can do an easy deployment when a customer signs on. If that's done, the fibre pull is simple.

Sure you probably need new cards deployed in your cabinets, but again, that doesn't involve digging up and relaying roads over the course of weeks where the big cost and hassle is.

Like most companies, AT&T want to stunt the market in order to keep making money off existing infrastructure for as long as possible. Up the window dressing speeds, deploy more bandwidth restrictions on customers, use legal spoiler tactics on competitors to the detriment of customers, just anything to keep bleeding the customers without spending money on network upgrades.

 

The only "ongoing debate" is coming from one Craig Moffett who has long been known as a shill for the cable industry. Everyone else, even AT&T, is saying between gnashed teeth that Verizon did the right thing.

The investor community is certainly happy with their decision. And the technologists are pointing out that Verizon has future-proofed their network with fiber.

So.......

 

Is this realistic? Does Verizon really not oversubscribe? That would imply that the cost of operating that 2.4Gb/sec node (and all related infrastructure) is low enough to be covered by 32 homes. How much does one pay in the US for that service? Someone earlier said $50 a month, but let's up it to $100, since it's so state-of-the-art. That's $3200/month they're collecting from the 32 homes. That seems fairly low; is it really that cheap?

The reason I ask is because of the notion that FiOS will save people from bandwidth caps (which surprisingly was the thread's topic). I can only see it happening if they don't oversubscribe. If their business model depends on oversubscription, and I really, really, really suspect it does, then when people start to actually consume the bandwidth they supposedly paid for (below cost), that's when the caps will start to pop up.

 

It's not oversubscription per se. It's the fact that the cable companies are implementing caps because they are limited technologically to the kinds of bandwidth they can deploy. Due to the cable TV & internet infrastructure, they can't offer HD channels, high speed bandwidth usage, VOIP, et al, without serious degradation.

Verizon's infrastructure, on the other hand, has none of those limits. I'm sure that Verizon will oversubscribe. But it's a hell of a lot easier to upgrade the bandwidth of a single piece of fiber-optic glass via wave division multiplexing than it is to try and push more bandwidth through a single piece of coax cable.

All I'm pointing out is that Verizon's ability to meet the bandwidth challenges of tomorrow because of their infrastructure investments today dwarfs anything the cable companies hope to deliver in their current state.

 

I don't think its the citys that are hte prblem. That can be done easily enough. Its when you get to east bumble fuck New Jersey or Hicksville New York that the problem lies. I don't see verizon or anyone laying out cable for 3 miles or so just to get 1 customer or even a handfull o customers. At least not for many years into the future

I wonder how much Japan pays for those lines. how much does a customer pay a month for unlimited access ?

Verizon is going to continue rolling out fios all over the place and they are going to continue picking p costumers. Whats more they are going to continue bumping up the speed of fis to stay ahead of cable and why not , they can easily do it .

The real problem is extremely rural areas. If the goverment was going to pay so that everyone had fiber strait to the house that is one thing but when its a company that has to look at the costs of running a line 5 miles out to hit 2 farms they are going to wonder why they are going to spend that much money.

 

the worst is, the US pretends to do nation building in other countries (after having destroyed them). Why doesn't it at home?

Same issue as the power grid : it's infrastructure. Having the government build and maintain infrastructure is not that communist or socialist. What is going on in your country is far worse (the permanent war economy, and the rescue of the financial debacle, where US workers will end up paying trillions to benefit the higher casts.)

About rural areas, can we have fiber cables hanging on the poles that already bear the copper lines? from my understanding the main cost would probably be labor. Have the gov. recruit people in the local population and actually do something?

 

Here in Garland TX (near Dallas TX) ppl are jumping to FIOS in large numbers. FIOS is better and even cheaper too compared to Time Warner cable. The cost for me with Time Warner cable TV, Verizon DSL and phone was about $150 a month and the TV was over $65 a month with only ext basic and the DSL was a 768k down. The FIOS package was for over 300 channels with HD channels, a 15/2 internet, and phone service with unlimited long distance with no charge for only $109 a month. The Time Warner internet 3/128k was $75 a month at the time and you would be luck to see a download speed reach 35k a sec to with it. Here in less then a month there upgrading the 15/2 to a 20/5 for no charge too. As you can see here every one here is jumping to FIOS when its available to them because its much better and cheaper then to other options.

 

Again, how much does it cost to operate that node? That's the key question. Overselling isn't a problem because of technology issues. It's not that you won't get the bandwidth you're promised so the ISP will cripple people's connections to make up for it. ISPs never guarantee 100% of the advertised bandwidth, so most people already don't. Overselling's a problem because at the price the average consumer pays, they're not really paying for the bandwidth advertised. It's what makes it all affordable.

Your link is what, 20mbit? I'm pretty sure that what a cable consumer pays is nowhere near what a dedicated, unlimited 20mbit link actually costs. What you're paying for is a projected average, how much they figure the typical user will consume. If they're wrong though, and people consume more than that, which I suspect is what is happening, then their profits are in danger. That's when they start to impose arbitrary caps. (This can mean the average person using more than they expect, or just more people than they predicted on the upper end of the consumption curve.)

Now, don't get me wrong: ISPs are nice and greedy, so they may be a little trigger-happy in protecting their profits. This is all compounded by net neutrality (which I'm not against). People are using tons of their bandwidth, and if they're not buying the ISPs content, then they're not getting an extra penny for it. I think this situation is untenable; I think we'll move to cellphone-like tiered pricing plans, with caps and metered usage.

Now, maybe Verizon does not oversell, or they accurately projected usage and actually make a profit off people using hundreds of gigabytes a month. This whole argument hinges on that, so naturally, take my premise with a grain of salt. Verizon's investment in infrastructure might mean that they can offer more bandwidth at a lower cost, but that just means the thresholds have shifted, not that they're not there.

 

Estimates that I've read are roughly $70-80 for 100mbit symmetrical connection.

 

I don't know what it costs to operate a node for Verizon, but considering they're upgrading their network from BPON to GPON as quickly as they can, I don't think it could cost too much. They are, after all, a business looking for profits.

My connection is 50mbit/20mbit. As for the profits aspect of the argument, I don't believe that's the case. Why? Because they're already putting in place the infrastructure to handle millions of users with connections that are actually faster than the one I've got.

Additionally, fiber is inherently much easier to upgrade end bandwidth cheaply than any other transmission medium we've got.

 

I think you're still focusing on technology. My point is that because of oversubscription, you're not actually getting the bandwidth you're buying. But that's okay, because you're technically also not paying for it. I had numbers from a quick internet search on dedicated link costs, but they seem to be conflicting, so I'd rather not bring any up (things like $500/mo for a T1 line, but elsewhere 8k/y to 45k/y for an OC-3, a pretty big difference). But regardless, all of them are significantly more expensive than Verizon charges for 50/20 ($144 w/o phone), even if you divide the OC-3 by 3 (get 3 neighbors) to get a dedicated 50mbit link.

The issue is that dedicated links are expensive. And that's not considering all the extra infrastructure necessary, if any. So based on those prices it's implicit that based on what the average consumer pays, they're not all expected to use the full bandwidth advertised. And the companies certainly never guarantee the full bandwidth. So again, if it turns out that consumers are actually using this bandwidth that the ISPs never actually thought they'd use, and the ISPs are not getting paid enough for all this content consumed, and if Net Neutrality becomes a law (as it probably should) then we'll probably see some sort of cap.

The fact that Verizon is expanding means, to me, that users are not 'over'using their bandwidth just yet.

 

The reason I've focused on the technology is that it is the technology/infrastructure that determines whether or not a company is forced to institute caps. The cable/dsl companies are forced to cap because their infrastructure is unable to meet the demand. They can't upgrade easily either, which is why they're trying to squeeze as much as possible out of their existing infrastructure as they can.

Verizon on the other hand can easily scale their fiber network to meet demand because fiber has limitless bandwidth carrying capacity as we know it.

 

We'll see how switch performance increases over time with demands. It's funny that you put so much faith in Verizon but have none for other companies.

 

Limitless? Are you sure? Do you have actual source for that? It doesn't sound right. Regardless, you seem to think these things are priceless, and I don't think so. As soon as people start to really consume the bandwidth they supposedly paid for, caps will come in. The only thing that's saving them is that at 50Mbit, the bandwidth you could consume in a month is far, far above 250G (a few orders of magnitude greater, in fact) so it's quite possible that no one would reasonably come close to the consumption of a dedicated link.

 

You should look up WDM technologies. Run out of bandwidth? Add more light wavelengths.