| |
to freejazz_RdJ
Re: CNOC R&V of CRTC 2011-703 and CRTC 2011-704said by freejazz_RdJ:
For the anon, lots of utilities get easements. If competitors were to build facilities, they too would get easements. It isn't a unique advantage available only to incumbents.
Municipal Consent and Easements are two totally different things. Any Canadian Carrier can get Municipal Consent to install facilities along, under or across public roads. Bell has had easements to install over private property (i.e. backyard) for many decades. To assemble those easements today would cost a fortune. |
|
MaynardKrebsWe did it. We heaved Steve. Yipee.
Premium Member
join:2009-06-17 |
to freejazz_RdJ
said by freejazz_RdJ: if Bell has paid $5000 for a concrete pad, they've paid $5K for a concrete pad and it should be recoverable.
Sure. But indies should only pay a pro-rata share for a finite period of time, not in perpetuity. And certainly not compounded over time. |
|
| |
to freejazz_RdJ
said by freejazz_RdJ:I understand the issue with establishing competing facilities. I too would like a more efficient operator. And I've witnessed the rise and fall of CLECs, so I know it isn't easy. But at the end of the day, if Bell has paid $5000 for a concrete pad, they've paid $5K for a concrete pad and it should be recoverable. Yes it should be recoverable. But what about Bell deploying a VDSL DSLAM that has all sorts of additional features built in such as Multicast that are NOT made available to competitors, but we as competitors are required to pay for? Or what about the fact that their optical network to feed the DSLAMS is designed with high end DWDM gear that is only initially configured to pass a small amount of traffic. The rates for the cost for capacity are set NOW, not with a forward look to the capacity that WILL be used, and the lower cost/megabit down the road. said by freejazz_RdJ:I'm still re-reading all of the paperwork filed in all the R&V's, but what does stand out so far is the CNOC AGI report in which we are told that high peak use ISPs will have higher costs. Shocking! That should come as nor surprise. Didn't the ISPs want a capacity model in the first place?
Yes, but the capacity model we got was simply a conversion of cost per gigabyte across a month converted in to a cost per kilobit/second by a conversion factor. This isn't a mathematically valid conversion, as measuring the kilobit/second impact of a gigabyte of data transferred during a 1 month window has almost as much to do with the price of tea in China as it does the cost per megabit/second of peak. Cost for capacity should go down as network utilization goes up. We both know that a 10km stretch of fiber used to transfer 155mbps of SONET data is no where near as efficiently used as a 10km stretch of fiber used to transfer 10GBPS of Optical Ethernet data - simply in cost per megabit/second km of fiber AND cost per megabit/second of optronics. I think the AGI report might be trying to point out that consumers want higher peaks, and that is what the market is demanding (and will demand more of in the future), but I do understand your point. |
|
| HeadSpinning |
to Ott_Cable
said by Ott_Cable :My point isn't Incumbents vs competitors getting their encasement.
Even the Canadian government have to compensate for expropriated property. Why does Bell get away with these encasement without compensating the land owners? To say that they get to build their facilities without the public is not quite true there.
I thin you mean easements... If Bell were granted the easement before the property was subdivided, then the cost is trivial. To get an easement AFTER is much more complex, and costly. This statement holds true for both incumbents and competitors - but Bell has the advantage of getting their easements BEFORE the land was developed. |
|
| HeadSpinning |
to MaynardKrebs
said by MaynardKrebs:Sure. But indies should only pay a pro-rata share for a finite period of time, not in perpetuity. And certainly not compounded over time.
Not completely true. The way telco costing works, the rate is theoretically set with the recognition that the competitor is not actually acquiring the asset, and compensates the incumbent for the front end loading of the expenditure. |
|
| |
PierrePoutin
Anon
2012-Mar-4 6:46 pm
said by HeadSpinning:said by MaynardKrebs:Sure. But indies should only pay a pro-rata share for a finite period of time, not in perpetuity. And certainly not compounded over time.
Not completely true. The way telco costing works, the rate is theoretically set with the recognition that the competitor is not actually acquiring the asset, and compensates the incumbent for the front end loading of the expenditure. The way I see it is: The end-user is paying in perpetuity for "line conditioning" (and pays over and over again each time they switch ISP). In effect the resellers customers are only "conditioning" ancient lines for Bell's customers and Bell's IPTV, in perpetuity. The rate of 4.5x the actual cost; This may or may not not take into consideration that Bell is making reseller end-users pay 4.5x times B/W costs and hardware costs due to Bells IPTV. In the cable world, we can use videotron as another example. All their new cell (wireless) runs on the same network adding to the traffic. Again the reseller end-user is paying for this. Rural Bell customers are not getting IPTV. Thus the deferral account submission may be lower by 4.5x just due to this? If so this would mean Bell is making reseller end-users pay for something they shouldn't be. Same with Videotron. In Cogeco's case... well... there is no excuse there. It's just plain highway robbery. HeadSpinning, surely you must see that the "competitor" must be "over-compensating" the "incumbent" for things that it shouldn't be. |
|
| |
to HeadSpinning
I agree it would be good to know more about the costing. If the feature is part of the base image, then everyone should pay for it. How much of a rebate do you give wholesale for each protocol they don't use in the base image? For each FIB entry not used by wholesale? You can't calculate that. You can however if there is an added license to enable it.
The one that really interests me is your DWDM example. How is the unit cost of interoffice DWDM transport calculated? I can't imagine that the first wave that triggers the dark fiber --> dwdm transition is loaded with 100% of the cost recovery burden. If it is, that is bad. All services should share the cost evenly correlated to fill factor with the trigger service perhaps covering some of the early planning expenses.
I'm also shocked that it was a GB/month to $/100Mbits/month conversion. I really need to read the stuff more carefully. I understand they used costing data from the wholesale UBB GB cap era, but didn't think it was a direct conversion. Hopefully JF or someone can point me to exactly where this is stated.
In the end, I'm not sure how much you can force the incumbent (via regs) into building the network that conceptually has the lowest unit cost. If they still want to use 10G sonet instead of ethernet for interfaces for some strange reason, can you really change that? If they were to build with the lowest unit cost technology but that yielded excess capacity, would wholesale share in the initial losses due to the cost of the overprovisioning? |
|
| |
to MaynardKrebs
said by MaynardKrebs:Sure. But indies should only pay a pro-rata share for a finite period of time, not in perpetuity. While the cost of individual concrete pads may have a finite useful lifespan to be recovered from, the equipment that will replace the one originally installed may not be able to reuse the same pad and new accommodations will need to be arranged... which means that as one cost expires, it gets replaced by new costs so you are never done paying. |
|
| |
to freejazz_RdJ
said by freejazz_RdJ:I agree it would be good to know more about the costing. If the feature is part of the base image, then everyone should pay for it. How much of a rebate do you give wholesale for each protocol they don't use in the base image? For each FIB entry not used by wholesale? You can't calculate that. You can however if there is an added license to enable it.
Either that, or let us have access to the features...  said by freejazz_RdJ:The one that really interests me is your DWDM example. How is the unit cost of interoffice DWDM transport calculated? I can't imagine that the first wave that triggers the dark fiber --> dwdm transition is loaded with 100% of the cost recovery burden. If it is, that is bad. All services should share the cost evenly correlated to fill factor with the trigger service perhaps covering some of the early planning expenses. If the commission recognized that, then there would be an adjustment factor over time - or the costs would be re-visited as the systems were filled. The fact is, metro transport costs a heck of a lot less than $23/megabit/month, and the CO to DSLAM link is shared with Internet and IPTV traffic, so Bell can't economically offer an IPTV service if much of that price comes from the non-metro portion of the transport. Heck, I can buy a metro 622mbps OC12 from Bell on a RETAIL tariff for $7/megabit/month - and that's on ancient expensive SONET systems, not packet ethernet! said by freejazz_RdJ:I'm also shocked that it was a GB/month to $/100Mbits/month conversion. I really need to read the stuff more carefully. I understand they used costing data from the wholesale UBB GB cap era, but didn't think it was a direct conversion. Hopefully JF or someone can point me to exactly where this is stated.
The only costing data the incumbents (except for MTSA) provided was based on $/GB from the UBB process, so the only option the commission had was to come up with a conversion factor. They didn't really have a choice unless they wanted to force the incumbents to re-cost their services using the capacity model. said by freejazz_RdJ:In the end, I'm not sure how much you can force the incumbent (via regs) into building the network that conceptually has the lowest unit cost. If they still want to use 10G sonet instead of ethernet for interfaces for some strange reason, can you really change that? If they were to build with the lowest unit cost technology but that yielded excess capacity, would wholesale share in the initial losses due to the cost of the overprovisioning?
No, but when the commission uses meaningless units of measure like $/GB to measure network costs that goes down as you load it, then convert it to a capacity rate, and the incumbents submit cost studies based on who knows what level of efficiency and network loading factors, you're going to end up getting numbers like $23 per megabit/second. There is no transparency in the process. Everything is submitted in confidence, and the Commission has neither the expertise required or the budged to hire subject matter experts. They have to do the best with what they have, and this is the net result. |
|
 DavesnothereChange is NOT Necessarily Progress
Premium Member
join:2009-06-15
Canada
2 edits
1 recommendation |
said by HeadSpinning:....There is no transparency in the process.
Everything is submitted in confidence, and the Commission has neither the expertise required or the budget to hire subject matter experts.
They have to do the best with what they have, and this is the net result.
I've just read this last page of posts, and have quoted the statements which stand out the most to me as being the crux of the matter. Everything else stated/speculated here is simply a consequence of this #### confidentiality process, and the fear running deep in the hearts/minds of corporate execs, that if they were to disclose the numbers, that someone would catch them in a lie or several.  CNOC is pushing hard for disclosure because they can smell the lies, even from the other end of an overpriced fibre haul.  |
|
1 recommendation |
to HeadSpinning
said by HeadSpinning:But what about Bell deploying a VDSL DSLAM that has all sorts of additional features built in such as Multicast that are NOT made available to competitors, but we as competitors are required to pay for? What is the fractional cost associated with the DSLAM's multicast capability? Probably $0.00 since I doubt there is any option to NOT have that and many other features even Bell might not be using. When you buy any switch or router for your own company's use, the equipment likely has dozens if not hundreds of features you will never use even in the base feature license and you have no option to pay for a license that only has the features you use there either. As for making features like multicast available to 3rd-party ISPs, this would require switching to a more TPIA-like setup where the incumbent would be handling GAS traffic as IP in order to handle multicast groups in an efficient and meaningful way... which would mean goodbye MLPPP, native IPv6 and the ability to manage your own IP space with all the ramifications that come with it. Considering how slowly incumbents are implementing IPv6 and how badly 3rd-party ISPs are starting to feel the IP block squeeze, I would imagine PPPoEv6 and the ability to manage IPs on a per-login basis are far more valuable than multicast. |
|
| |
They can deploy 2 L2 domains. One for transporting PPPoE packets for Internet access, and another for Multicast. No need to go to a TPIA-like environment. |
|
|
 andyb
Premium Member
join:2003-05-29
SW Ontario |
to HeadSpinning
I dunno nothing about sonet and metro links but assume internal network would be a metro yes? If you can get me a link or a pdf of the sonet prices you said I will either file them myself or have some else do it during this R&V bullshit |
|
| |
Page 16 - Item 301.3(e) - 622mbps OC12 - $4400/month between any two wire centres within a rate centre. The rates in question were last revisited in 2004. |
|
andyb
Premium Member
join:2003-05-29
SW Ontario |
andyb
Premium Member
2012-Mar-4 8:38 pm
So even dividing that by 6 is way cheaper than current BUT is only 2 wire centers in a rate centre.Not sure how that rate centre shit might look. |
|
| |
to HeadSpinning
said by HeadSpinning:Heck, I can buy a metro 622mbps OC12 from Bell on a RETAIL tariff for $7/megabit/month - and that's on ancient expensive SONET systems, not packet ethernet! But metro transport is fixed point-to-point within a serving area while GAS/AHSSPI is point-to-multipoint across the whole service territory... not quite the same costs. As far as SONET on "legacy" equipment goes, Bell started moving customers to SONET emulation on Ethernet back-haul years ago, only customers who require true SONET from end to end are guaranteed to remain on it for the time being while the legacy equipment is still serviceable. |
|
| |
said by InvalidError:But metro transport is fixed point-to-point within a serving area while GAS/AHSSPI is point-to-multipoint across the whole service territory... not quite the same costs. Quite right. Not the same costs, but Bell has argued that the reason GAS aggregation costs more than transit is because it is on lower density Metro routes, not high density long haul routes - so the majority of the costs of GAS transport should be related to Metro. Based on that, and the density of transport of GAS, I would think this is a reasonable comparison. Given advances in technology and optronics, the price for the Ethernet transport should be lower than the SONET transport. said by InvalidError:As far as SONET on "legacy" equipment goes, Bell started moving customers to SONET emulation on Ethernet back-haul years ago, only customers who require true SONET from end to end are guaranteed to remain on it for the time being while the legacy equipment is still serviceable.
They're not doing SONET "emulation" on Ethernet. They're overlaying SONET on DWDM systems. The current generation of Ciena (nee Nortel) gear they're installing takes a SONET multi-port transponder blade that maps up to 16 SONET (16xOC3 or 4xOC12 or combinations to a maximum of 2.5G) circuits on to a 2.5G wavelength in either linear or ring configurations. It is true next gen SONET, and isn't running on Legacy equipment. I know from the maintenance releases that several of my OC3 circuits have been migrated to this technology, and I have full SONET End to End visibility between my MUXes at both ends. |
|
| HeadSpinning |
to andyb
said by andyb:So even dividing that by 6 is way cheaper than current BUT is only 2 wire centers in a rate centre.Not sure how that rate centre shit might look.
So take a typical city like London which has 4 wire centres within the rate centre. Lets say you picked the Clarence CO as your hub location, and then purchased transport from Bell to the other 3. You would then need 3 OC12 links - one to each of the other 3 COs - to complete your network which would deliver 622 mbps to each CO. In a city like Toronto might have 25 wire centres, so the network would be larger - but the cost/megabit/second remains the same. |
|
MaynardKrebsWe did it. We heaved Steve. Yipee.
Premium Member
join:2009-06-17 |
to InvalidError
said by InvalidError:said by MaynardKrebs:Sure. But indies should only pay a pro-rata share for a finite period of time, not in perpetuity. While the cost of individual concrete pads may have a finite useful lifespan to be recovered from, the equipment that will replace the one originally installed may not be able to reuse the same pad and new accommodations will need to be arranged... which means that as one cost expires, it gets replaced by new costs so you are never done paying. Do the math.... $5k pad under a remote....which might be (in an urban setting) supporting say 300-400 pairs, which are typically used for both POTS & data. For the purpose of this investigation, let us say that POTS is strictly Bell....so the cost of the pad for the data side is 1/2, or $2500. Is the economic life of a pad 10 or 20 years? It's probably about the same as a sidewalk, which is typically more than 20 years, but let us be generous to Bell and say it's only 20 years, or 240 months. Divide that into $2500 and you have about $10.42/month. Now divvy that up among the say 300 data users/month and you get about 3.5 cents per month. Pads are generally made bigger than the size of the currently installed gear - to accommodate newer bigger boxes. A few new holes drilled, a few new Hilti anchors and some epoxy and the pad is repurposed for new gear. |
|
| MaynardKrebs |
to HeadSpinning
said by HeadSpinning:said by freejazz_RdJ:In the end, I'm not sure how much you can force the incumbent (via regs) into building the network that conceptually has the lowest unit cost. If they still want to use 10G sonet instead of ethernet for interfaces for some strange reason, can you really change that? If they were to build with the lowest unit cost technology but that yielded excess capacity, would wholesale share in the initial losses due to the cost of the overprovisioning?
No, but when the commission uses meaningless units of measure like $/GB to measure network costs that goes down as you load it, then convert it to a capacity rate, and the incumbents submit cost studies based on who knows what level of efficiency and network loading factors, you're going to end up getting numbers like $23 per megabit/second. Exactly!! The CRTC may as well have used dimensions like leap furlongs per long hundredweight in the tariff rates they created - it would have been more transparent, and probably more accurately reflected how they arrived at the numbers they did.  |
|
MaynardKrebs
1 edit |
to InvalidError
said by InvalidError:said by HeadSpinning:But what about Bell deploying a VDSL DSLAM that has all sorts of additional features built in such as Multicast that are NOT made available to competitors, but we as competitors are required to pay for? What is the fractional cost associated with the DSLAM's multicast capability? Probably $0.00 since I doubt there is any option to NOT have that and many other features even Bell might not be using. Kinda like Bell getting to charge $2.95/month in perpetuity for touch tone service on your phone, when the switch has had the capability built-in since 1980 and nobody except the 30,000 most rural of users who aren't served by a DMS100 or a similarly featured switch don't have touch-tone now. |
|
| |
to MaynardKrebs
said by MaynardKrebs:Is the economic life of a pad 10 or 20 years? It's probably about the same as a sidewalk Could be less than 10 years since the number of subscribers serviceable from each pad drops by half every few years as remotes get pushed deeper into neighborhoods... or to put it the other way, the number of pads doubles every few years, much like Moore's law. Since each new pad serves about half as many lines as the pad it replaces or supplements, equipments would also tend to get smaller rather than larger even without factoring in integration/miniaturization. You also cannot assume that the "pad cost" for 400 loops is "half for the data side" and then 1/400th of that since not all loops have both services (many may have only one or even neither) and the cost can only be spread across revenue-generating loops, so the cost "per service" is much greater than 1/800th. |
|
jfmezei
Premium Member
join:2007-01-03
Pointe-Claire, QC |
jfmezei
Premium Member
2012-Mar-5 1:22 am
In areas where real FTTN has been deployed, the number of remotes is not likely to change in a long time. The next step is to tack on a GPON splitter at/near the remote and progressively replace copper loops with fibre to the home.
If you consider how a neighbourhood has been wired to a JWI, adding a new JWI to reduce distances in the neighbouhood (essentially splitting the neighbourhood in two) might require considerable rewiring to bring half the homes to that new DSLAM. If you're going to do that amount of work, you might as well bring in FTTH for that new half of neighbourhood so you could compete against cable. |
|
DavesnothereChange is NOT Necessarily Progress
Premium Member
join:2009-06-15
Canada
3 edits |
to MaynardKrebs
said by MaynardKrebs:Kinda like Bell getting to charge $2.95/month in perpetuity for touch tone service on your phone, when the switch has had the capability built-in since 1980.... AND that it costs Bell less money to enable TouchTone than to keep it externally blocked with an analog filter. Yet they have always charged for enabling it, told us that we NEED it, when to this day we do not (as the switches by design still understand pulse), and the CRTC has always rubber-stamped Bell's illusion, collusion, and confusion, and subsequent monthly contusion to our wallets.  |
|
Davesnothere
2 edits |
to jfmezei
said by jfmezei:....If you consider how a neighbourhood has been wired to a JWI, adding a new JWI to reduce distances in the neighbouhood (essentially splitting the neighbourhood in two) might require considerable rewiring to bring half the homes to that new DSLAM. If you're going to do that amount of work, you might as well bring in FTTH for that new half of neighbourhood so you could compete against cable.
Agreed. However I wonder whether Bell sees things that way ? Futhermore, as (real) fibre is virtually lossless, there would be no need to split the JWI and only upgrade half of the 'hood - simply put the GPON and/or whatever other fibre-related gear next to the original JWI, and run longer fibre(s) from there to the WHOLE 'hood. |
|
| |
to jfmezei
said by jfmezei:In areas where real FTTN has been deployed, the number of remotes is not likely to change in a long time. Alcatel's product lineup disagrees. If there was no market for remotes down to 48 ports or less, Alcatel would not be offering 48 ports ruggedized outdoors sealed remote expansion modules and manufacturers would not be talking about going down to 8-12 ports in that market. said by jfmezei:If you consider how a neighbourhood has been wired to a JWI, At the scale I'm talking about, the JWI would be more expensive than the remote so it would be more cost-effective to splice the remote directly in-line... saves the costs of building the JWI and splicing cables in/out of it. If the objective is 100/xx, there wouldn't be enough range to pass much more than 48 homes in the first place anyhow, so the flexibility of a JWI in a deep-FTTN scenario is superfluous. |
|
jfmezei
Premium Member
join:2007-01-03
Pointe-Claire, QC |
jfmezei
Premium Member
2012-Mar-5 3:33 am
My point is that for currently deployed FTTN/VDSL2 areas, the next step isn't to deploy even closer DSLAMs, it would be to goto FTTH.
So any costing which predicts that more dslams will be deployed would be erroneous.
What may happen is that as VDSL2 is deployed to less dense areas, they will need smaller DSLAMs to cover few homes within the distance limits.
But existing deployments will not likely change much for xDSL copper stuff, even if Bell claims its copper is capable of carrying photons. |
|
DavesnothereChange is NOT Necessarily Progress
Premium Member
join:2009-06-15
Canada |
said by jfmezei:....But existing deployments will not likely change much for xDSL copper stuff, even if Bell claims its copper is capable of carrying photons.
Or Tim-Bits, RedBull, Popcorn, or Montreal Bagels. But yes, Bell ought to recognise the most efficient and effective way to spend whatever cash they must spend. |
|
| |
to MaynardKrebs
said by MaynardKrebs:Pads are generally made bigger than the size of the currently installed gear - to accommodate newer bigger boxes. A few new holes drilled, a few new Hilti anchors and some epoxy and the pad is repurposed for new gear.
Typically not. They pour the size of pad required for the existing equipment. You can't really re-use a pad either, as the migrations from one technology to another are rarely "rip and replace" in one night, so both pieces of equipment have to be active for some overlapping period of time. |
|
| HeadSpinning |
to MaynardKrebs
said by MaynardKrebs:Kinda like Bell getting to charge $2.95/month in perpetuity for touch tone service on your phone, when the switch has had the capability built-in since 1980 and nobody except the 30,000 most rural of users who aren't served by a DMS100 or a similarly featured switch don't have touch-tone now.
I don't care if Bell charges their RETAIL customers $2.95/month for touch tone forever. I only care about access services for competitors. |
|
