 w0go.O
join:2001-08-30
Springfield, OR | Internet should be treated as public utility/service Full regulation and definition of requirements for someone to sell and market access to it. Shouldn't be any different than power/phone/water. |
|
| said by w0g:Full regulation and definition of requirements for someone to sell and market access to it. Shouldn't be any different than power/phone/water. No problem, I'll start charging you my full T1 cost plus network overhead costs. That will be $800 per month please. |
|
 superdogI Need A DrinkPremium,MVM
join:2001-07-13
Lebanon, PA | reply to w0g
said by w0g:Full regulation and definition of requirements for someone to sell and market access to it. Shouldn't be any different than power/phone/water. Hogwash!. I also own and operate an ISP. I spent MY money, blood sweat and tears to build and operate it. Who in the hell is anyone to tell me that if I sell something and I am up front about how it operates and what it does, that I can't do that.
Once again, Its mine, I own it and I will do what I want with it. In most cases, without my service (WISP), there would be nothing but dialup for 45% of my customers. My TOS does not even permit P2P of any type. My customers know this up front, BEFORE they use my service. I couldn't afford to operate if I were to let bandwidth intense programs run wild because of the costs.
As another WISP owner already mentioned, there is no way any ISP could operate by paying $600 to $800 a month for a T1 and then give it all to 1 user for $40 measly bucks a month. I am NOT a charity and I don't owe anyone anything.
You could use your argument on the telephone systems in the USA because our tax dollars over the years have paid for it. None of your tax money ever built even 1 mile of any cable operators network, nor did it build any of mine, therefore, I will do what I want, when I want and if they try and regulate me, I will close and a LOT of people will be back on dialup in short order.
Just so you understand, I do agree with an open internet to some extent. As a small provider, I try to be fair to everyone on my network. I don't care what VOIP app. you run or any other type of "Normal" activity. What I do care about is complete fairness when it comes to accessing the bandwidth I have.
If I would allow P2P apps. like Bittorrent, LImewire etc. to run, 2% of my users would hog all of the bandwidth. How in the heck is that fair to my other users?????
I could hear the phone call now: "I'm Sorry Mrs. Smith that your internet is crawling tonight, but little Johnny down the street is updating his World of Warcraft game with a few Gig worth of patches and since the CO that makes the game is to cheap to buy dedicated bandwidth so that he can stream it from one source and only use his own allocated bandwidth, he is using Bittorrent and opening up 1000 separate connections to his single computer. It should be done in a few hours though, so just go knit a sweater until he is finished, and then things should go back to normal. Have a nice night Mam!"
If we want to talk about fairness, lets talk about Blizzard and a few other CO's that have decided to go "Cheap" and distribute their money making ventures over a P2P network rather than buying their own bandwidth. That kind of crap really pisses me off.
Anyway, its just not as easy as it seems to get that internet connection to your PC, and the day anyone tells me that I have to give it away or stop blocking certain programs, I will do 1 of 2 things, either close up or charge each user based on usage and get back from the hogs that take. If I can't do that, then back to step one, CLOSE. 
--
»www.wavecrazy.net
|
|
|
|
patcat88
join:2002-04-05
Jamaica, NY
kudos:1 | Why don't you get something cheaper than a rip off T1 from the ILEC? Why not go with a regional fiber metro ethernet provider that has their own fiber for $2K-3K per mo for 10mbps? Or use P2P microwave to a POP? |
|
 JKM
join:2009-06-08
Seymour, MO | I have 2 fiber providers near me. One 7 miles and one 19 miles. Cost of bandwidth is about $300.00 per MB per month from either. Then there's the $350.00 per month tower lease for the PTP link. Figure it out, that's $416.67 monthly per MB for 3 MB. As I keep preaching, the real problem is the last mile technology is far ahead of the middle mile technology, bandwidth availability and fair pricing.
--
Begin with the end in mind! |
|
patcat88
join:2002-04-05
Jamaica, NY
kudos:1 | said by JKM:I have 2 fiber providers near me. One 7 miles and one 19 miles. Cost of bandwidth is about $300.00 per MB per month from either. I assume you mean mbitps per month?
Is $300/mbitps/mo with internet access or is that loop, and does it include free construction or not? What about faster links than 10mbitps or long term contract pricing? Why not get P2P metro ethernet to some place with cheap or spare bandwidth (colo center or office building, with a hosting company willing to oversubscribe their gigabit link)? Cogent is only $10/mbps/mo on network, although Cogent is on the cheapish end. What about the local cable company and their business division or they don't serve ISPs? |
|
iansltx
join:2007-02-19
Golden, CO
kudos:2 | reply to nevtxjustin
Hmm, sounds like people will just get T1s then... |
|
iansltx
join:2007-02-19
Golden, CO
kudos:2 Reviews:
 ·Comcast
| reply to JKM
If I recall correctly superdog just switched to a cheap Tier 3 with $65 per Mbit bandwidth, if not less.
I've been quoted $30ish per Mbit on a 100Mbit commit in my hometown.
If you had access to $30 per Mbit bandwidth ould you sell unmetered, unlimited, unthrottled connections at 1 Mbit for $45?
The T1 comparison is absolute hogwash if you're not using a T1 for your own pipe. If you are then fair's fair, however don't defend companies that have their own national backbones and have effectively zero marginal cost for 1GB of traffic. |
|
| reply to superdog
You're not the problem it's the big boys who control the middle. |
|
| reply to iansltx
said by iansltx: however don't defend companies that have their own national backbones and have effectively zero marginal cost for 1GB of traffic. This... |
|
| reply to iansltx
said by iansltx:...however don't defend companies that have their own national backbones and have effectively zero marginal cost for 1GB of traffic. Do you have any idea how much capital is required to carry traffic nationally? How about the opex costs for the facilities, power, lambda/optronics for a national backbone all over the US or world wide? You also need to scale that for 40-50% / year traffic growth which is what happens even with slowing customer net-adds. |
|
iansltx
join:2007-02-19
Golden, CO
kudos:2 Reviews:
·Comcast
| I said marginal cost. Once the infrastructure is in place, you merely have to swap out the lasers at the ends to upgrade capacity. One user's gigabyte of data (or thousand gigabytes of data) is a tiny fraction of the capacity of these backbone links, and thus doesn't really show up on the cost radar.
From what I hear, fiber routes cost $5 per Mbit to keep up. Again, less than a penny per GB. So practically zero. Also, a corollary of Moore's law has to do with bandwidth; equipment will be made on the fiber side to cope with demand. |
|
| said by iansltx:I said marginal cost. Once the infrastructure is in place, you merely have to swap out the lasers at the ends to upgrade capacity. One user's gigabyte of data (or thousand gigabytes of data) is a tiny fraction of the capacity of these backbone links, and thus doesn't really show up on the cost radar. From what I hear, fiber routes cost $5 per Mbit to keep up. Again, less than a penny per GB. So practically zero. Also, a corollary of Moore's law has to do with bandwidth; equipment will be made on the fiber side to cope with demand. You should cite your source as you heard wrong.
The perception of "once a network is in place" is highly inaccurate. While a large scale network does have incremental upgrades every year (which are no where near zero or pennies) , it needs major infrastructure upgrades every few years.
Growing capacity isn't a matter of "swapping out lasers." There are many different components required to bring bandwidth from point A to Z. To properly budget one should not ignore existing spend and only look at one aspect (a laser) they need to look at the loaded costs as eventually each component has to be upgraded to grow capacity needs.
These components are:
•Space and power in each facility (for a US only fiber based backbone figure about 400 or so facilities with amps, regen equipment and routers). This is JUST for the backbone.
•"the lasers" numbers are very large to add another Nx10G or Nx40G to the network all across the entire infrastructure. Also don't forget any regeneration equipment for spans which exceed transponder distances.
•10G Xenpak, Nx10G router cards, Chassis slots multiplied by the number of 10G links to get from point A to Z. Don't forget the redundancy for that link.
Marginal... No. This is actually very, very expensive and only accounts for the backbone costs. You have a similar type of analysis for the regional area networks and the last mile fiber or DOCSIS infrastructure.
You are also talking about "One user's gigabyte of data" and need to multiply that user times multimillions. |
|
iansltx
join:2007-02-19
Golden, CO
kudos:2 Reviews:
·Comcast
| We're arguing about different things.
I'm saying that the cost per megabit on teh backbone,w hile stll ammking a profit on the backbone service itself, is $5 per Mbit per month. Or about a cent per GB.
Also, I don't need to multiply 1GB by multimillions. We're talking about usage based billing (charging tons extra because you're two drops in the bucket instead of one) not network sizing.
On the last mile, I agree that non-fiber architectures are bnadwidth-starved and relatively expensive to upgrade (okay, DOCSIS 3 is less than $100 per customer, but whatever). However on the backbone there are plenty of companies selling $5 per Mbit and making a profit at it. Therefore costs on the backbone must be lower than $5 per Mbit.
Again, I don't hear backbone providers like XO and Level3 complaining about capacity. They just build more as needed. The fiber is already in the ground. The endpoints just need to be upgraded. Whereas last-mile providers are too cheap to lay down the fiber and thus decrease their upgrade costs from there on out by a large degree. |
|
3 edits | said by iansltx:We're arguing about different things. I'm saying that the cost per megabit on teh backbone,w hile stll ammking a profit on the backbone service itself, is $5 per Mbit per month. Or about a cent per GB. On the last mile, I agree that non-fiber architectures are bnadwidth-starved and relatively expensive to upgrade (okay, DOCSIS 3 is less than $100 per customer, but whatever). However on the backbone there are plenty of companies selling $5 per Mbit and making a profit at it. Therefore costs on the backbone must be lower than $5 per Mbit. First, I am having de ja vu of a previous conversation I had with you on a similar topic where the inaccurate $5 number came up.
Also you cannot mix GB served and Mbps, as networks are not built that way. That said there is correlation between a heavy GB usage and Mbps built to support it. Assuming you are on a 12Mbps plan and networks need to be built for at least a 95-98%ile. A heavy user will normally peak at 12Mbps for over 5% of the month which even with unsustainable $5/Mbps maps to $60 and this doesn't factor in any other costs to run an end to end network.
said by iansltx:Again, I don't hear backbone providers like XO and Level3 complaining about capacity. They just build more as needed. I would not use ISP examples of providers that are not profitable. Level3 has bled money since they started and has ridiculous debt. XO had 1 profitable year in the past decade.
said by iansltx:The fiber is already in the ground. The endpoints just need to be upgraded. Whereas last-mile providers are too cheap to lay down the fiber and thus decrease their upgrade costs from there on out by a large degree. This is partially correct. The fiber is in the ground, but if you build fiber networks you understand that new equipment has to be added every ~80Km to light new fiber and ever 1000Km to add capacity within lit fiber. Capacity also requires every add/drop city for router capacity, space, power.. but I am repeating myself with fact based data.
Your also incorrect on the last mile issues. While Verizon has done a great job at marketing FTTH, the reality is that fiber is not needed today to serve 99+% of the residential user base needs. Fiber is really only required (with today's technology) when you get well above 100Mbs which very few ask for (except some minority users on this site  |
|
iansltx
join:2007-02-19
Golden, CO
kudos:2 Reviews:
·Comcast
| 1. Coaxial plant from what I gather is MUCH more complex than fiber. Copper plant is also more complex than fiber. Therefore from a long-term ROI perspective fiber is better.
2. I can't get more than about 4.5 Mbps over copper to my location without multi-bonded T1s...and Im in town! So there copper is a dead tech.
3. $5 per Mbit is what I'm lifting directly from WebHostingTalk. There are folks over there (Mzima for example) who run their own backbones. Therefore I think they'd have a pretty good idea of costs.
4. If the cost is in megabits per second and not GBs then why are providers capping? Why aren't they either offering "free night and weekend GBs" or throttling people down during peak hours but still offering unlimited bandwidth?
5. If it's so expensive to run a fiber network etc. etc. then why are FiOS users able to get 20/20 for $60 per month with no caps, overages, etc.? Don't tell me "because Verizon is losing money on those customers" because I'm sure the company did the math and saw that the ROI is positive. I can name provider after provider in the US that have fiber networks, decent symmetric speeds, decent pricing and no caps. These providers will never have to dig up their last mile again; that problem is solved.
6. Note that I said "dig up". I'm fully aware that endpoint and midpoint equipment will have to be upgraded, though realistically until you see multiple providers competing fiber vs. fiber GPON will be plenty.
7. I can get Cogent 100 Mbps ethernet, including a Time Warner Cable loop, for $3000 per month for 100 Mbps symmetric...dedicated...to an area that at the time of the quote had no active compettion for the last mile on the fiber side. Put another way, there was no fiber but TWC's. Yet I was able to get $30/Mbps including internet transit. So that's $22 per Mbps between a large city and a small one (pop 10,000 or so). Don't tell me TWC is losing money on this because that's either not true or everyone i the communications business is in the wrong industry and we'll see bandwidth prices skyrocket in the next 30 days as companies go bankrupt. Fat chance. |
|
| 1. Maintenance of a new fiber plant is cheaper than a coax plant, but the cost to build it is very expensive. Verizon started 5 years ago and spent over 20B (that is billion) to build out FioS for the current homes passed. Even if you factor in maint savings, how many customers do you need to pay that off?. It might be a decade before anyone really knows whether Verizon’s bet on FiOS is a smart investment, but I personally think it was the right choice for them given the fact they can pay for it with other lines of business.
Comcast spent 1B to upgrade to DOCSIS3 and go all digital reclaiming tons of spectrum in 1-2 years. Comcast can now reach more customers with 50Mb advertised speeds than all other US telcos combined. Can FTTH offer higher speeds than DOCSIS... sure.... Will people buy that in masses anytime soon?... no. Was DOCSIS3 and Cavalry the right choice for Comcast vs FTTH... yes.
Also telcos have to build fiber all the way from the CO. Most cable companies (when needed) only need to build from the HFC fiber node point to the home. BUT (and that is a big but) they don't NEED to do it now.
This is a bit of a leap frog and one can successfully argue that FTTH has long term benefits, but in a real world business if you do not have to spend 20B, you don't send 20B. Techies of the world scream for FTTH, but when 99+% of the users don't need it, it is just plain STUPID to do it. VZ did not do FTTH because they want to please .001% of their customer base, they did it because they had to upgrade their plant from twisted pair. Cable companies do NOT have to do this to get the speeds customers are asking for.
2. I agree twisted pair copper is dead. Verizon realized this and decided to go FTTH. Their ROI on that investment is going to be very long, but it was do or die for them and the only way they can possibly do this is to pay for it with their wireless revenue. FTTH is not a profitable standalone business in the majority of US geographies and at current costs. Period.
3. Mzima is yet another unprofitable business. I think I have explained the reason for the unsustainable $5 per Mbit price point several times before.
4. This is simple. Average consumers don't understand 95%ile billing and Product folks want to make this simple. They do analysis on users in the high GB and rough estimates on the costs to support and draw various tiered structures around it. The billing systems are expensive to change too. Is it a perfect model? no. Is it pretty close?... yes.
5. Sorry to say it but, "because Verizon is losing money on those customers". This is a long term investment for VZ being subsidized by their other lines of business. They did the math and saw it was a do or die investment in their wireline business.
I stand by my two points with the facts that support them
•$5 / Mb price point is below the cost to build and operate end to end networks. Spot market pricing for transit has other factors around it and does not reflect the true costs. If you ask people that build networks they would agree.
•FTTH as a stand alone business in the majority of situations around the US is not a profitable business under today's economics. VZ is making a multi year (decade+) investment to continue in the wireline business and funding is subsided by their wireless business.
"Can you hear me now?" |
|
JKM
join:2009-06-08
Seymour, MO
1 edit | reply to patcat88
Yes that is Mbps. About $506.00 for a T1. Those are the options. No cable or DSL provider offers resale and the closest cable is 11 miles, and DSL 7 miles. They don't intend to supply bandwidth to their competitors. The real problem is not the last mile providers per say. It is the select last mile providers that also control the middle mile infrastructure. They don't intend to wholesale bandwidth at a price that will allow fair competition. When I can use 40Mbps service, I can purchase that for $4000.00 per month on a five year contract.
--
Begin with the end in mind!
EDIT: Just Internet bandwidth. |
|
