ARCHIVED - Transcript, Hearing 14 July 2011

This page has been archived on the Web

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

Providing Content in Canada's Official Languages

Please note that the Official Languages Act requires that government publications be available in both official languages.

In order to meet some of the requirements under this Act, the Commission's transcripts will therefore be bilingual as to their covers, the listing of CRTC members and staff attending the hearings, and the table of contents.

However, the aforementioned publication is the recorded verbatim transcript and, as such, is transcribed in either of the official languages, depending on the language spoken by the participant at the hearing.

Volume 4, 14 July 2011



Review of billing practices for wholesale residential high speed access services. Notice of Consultation CRTC 2011-77, 2011-77-1 and 2011-77-2


Outaouais Room

Conference Centre

140 Promenade du Portage

Gatineau, Quebec

14 July 2011


In order to meet the requirements of the Official Languages Act, transcripts of proceedings before the Commission will be bilingual as to their covers, the listing of the CRTC members and staff attending the public hearings, and the Table of Contents.

However, the aforementioned publication is the recorded verbatim transcript and, as such, is taped and transcribed in either of the official languages, depending on the language spoken by the participant at the public hearing.

Canadian Radio-television and Telecommunications Commission


Review of billing practices for wholesale residential high speed access services. Notice of Consultation CRTC 2011-77, 2011-77-1 and 2011-77-2


Konrad von FinckensteinChairperson

Len KatzCommissioner

Tom PentefountasCommissioner

Timothy DentonCommissioner

Candice MolnarCommissioner

Michel MorinCommissioner

Marc PatroneCommissioner


Lynda RoySecretary

Crystal HulleyLegal Counsel

James WilsonLegal Counsel

Tom VilmansenHearing Manager and Manager, Costing Methods and Tariffs


Outaouais Room

Conference Centre

140 Promenade du Portage

Gatineau, Quebec

13 July 2011

- iv -







20. MTS Allstream Inc. 696 / 4249


22. World Broadband Foundation (Munly Leong)756 / 4641

23. Claude Hénault779 / 4769

21. Jatinder Bhullar804 / 4883


24. Vaxination Informatique827 / 5023

- v -



Undertaking740 / 4519

Undertaking741 / 4521

Gatineau, Quebec

--- Upon resuming on Thursday, July 14, 2011 at 0903

4244 LA SECRÉTAIRE : A l'ordre, s'il vous plaît.  Order, please.

4245 LE PRÉSIDENT : Commençons.

4246 LA SECRÉTAIRE : Merci.

4247 We will begin today with the presentation by MTS Allstream Inc.

4248 Please introduce yourselves for the record, after which you have 20 minutes for your presentation.


4249 MS GRIFFIN-MUIR: Thank you.

4250 Good morning, Mr. Chair, Commissioners, Commission staff.  My name is Teresa Griffin-Muir.  I am the Vice-President, Regulatory Affairs at MTS Allstream.

4251 With me today are Mike Strople, Vice-President, Technology Development and Pauline Jessome, Senior Regulatory Analyst.

4252 The Commission has invited comments on how best to implement the following principles:

4253 First, as a general rule, ordinary consumers served by small ISPs should not have to fund the bandwidth used by the heaviest retail Internet service consumers; and

4254 Second, it is in the best interest of consumers that small ISPs, which offer competitive alternatives to the incumbent carriers, should continue to do so.

4255 The fact that the impact on consumer choice is emphasized signifies that the Commission recognizes that when establishing wholesale rules, even though certain competitors may occupy only a small portion of the market, these competitors still play a fundamentally important role.

4256 The Commission's first principle is just another expression of the pricing principle that those who use must pay.

4257 There are numerous examples of usage-based pricing: minute-based compensation plans for long distance and prepaid wireless services.  On the other hand, there are also many examples of flat-rated pricing.  Most notably, local voice services are priced in this manner without consumer backlash.

4258 Consumers are willing to countenance unequal usage for a reasonable fee in return for simplicity and certainty of pricing, and, in mandated wholesale pricing, tariffs must be structured in such a way that competitors are not forced to mirror the incumbent's retail offerings.

4259 Now we will address the questions posed by the Commission.

4260 Mike will speak to the first question.

4261 MR. STROPLE: In MTS Allstream's opinion, our proposed wholesale FTTN tariff reflects the preferred billing model for a wholesale aggregated DSL access service.  Like our legacy tariff, our proposed FTTN tariff is comprised of only two monthly rate elements:

4262 - a flat rate access charge to recover the fixed cost of the last mile access facilities between a competitor's end user and the serving wire centre, as shown in the diagram between points A and B; and

4263 - the AHSSPI or interface charge to recover the aggregated usage cost of transporting the competitor's end users' access traffic across MTS Allstream's shared network right to the interface point where the traffic is handed off to the competitor's own network.  This is shown between points B and C on the diagram.

4264 MTS Allstream's preferred model prices usage on the basis of peak-period capacity utilization on our network.

4265 We acknowledge that the usage-based costs of shared transport could be recovered in at least three ways, on the basis of:

4266 (1) peak-period reserved capacity in terms of bits per second, as measured at the interface point between the incumbent's and the competitor's networks, as we and other ILECs have proposed;

4267 (2) peak-period floating usage as measured on a 95th percentile basis, again at the carrier interface point, as CNOC has suggested; or

4268 (3) on the basis of the total amount of data transmitted over a month in terms of bytes of data transmitted at a given point in the shared network, which forms part of Bell's proposal.

4269 Since networks are provisioned on the basis of overall peak-capacity requirements, we believe that a pricing method that reflects peak usage is most appropriate.  Under our proposed wholesale FTTN tariff, a competitor plans ahead and tells us in advance how much peak-period capacity they wish to reserve on our network.

4270 It is important to note that MTS Allstream does not currently measure a competitor's network utilization in a way that would allow us to bill on either a 95th-percentile or a total-volume basis.  Therefore, we do not believe that it would be appropriate for the Commission to force us, MTS Allstream, to adopt a billing model that is either of these approaches.  To do this would require significant changes to our network operations and billing systems, thereby unnecessarily increasing the cost of the wholesale service.

4271 MS GRIFFIN-MUIR: In terms of question number 2, network usage is inherent in the nature of an aggregated access service, so any billing model must capture and charge for that usage.  In MTS Allstream's opinion, a method that is based on peak capacity more closely matches the way networks are designed and augmented.

4272 When setting the price of the variable transport network component, such as the AHSSPI, MTS Allstream uses the same methodology that all ILECs employ, in accordance with well-established, Commission-approved Phase II costing principles.

4273 First, we estimate overall peak-period demand on our shared transport network, as Mike said, between B and C on the diagram.  In order to do so, we take into account current and future demand on each link generated from all services that share the transport network.  These include:

4274 - first, Internet traffic generated by our own retail Internet customers;

4275 - second, traffic generated by customers of other MTS Allstream retail services;

4276 - third, Internet access traffic generated by end users of competitor ISPs; and

4277 - finally, traffic generated by the retail voice and data services delivered by competitors using other MTS Allstream wholesale services.

4278 Next, taking into account the overall peak-period demand from all services that use our shared transport network, we come up with a per-unit cost for this peak-period demand, including adjustments to reflect the fact that the shared network is not intended to and does not operate at 100 percent capacity.  All of these assumptions -- overall traffic growth over time, increases in peak-period demand over time, fill factors and various other loading factors -- are built into forward-looking Phase II costing and are reflected in the resulting capacity costs.

4279 Third, we then estimate the proportion of overall peak-period demand on the shared transport network for the wholesale aggregated DSL access service.

4280 All of the ILECs should be following the same Commission-approved and established Phase II costing principles to cost out their own proposals.

4281 When this is done correctly, regardless of which of the three billing methods is used, the resulting costs should be essentially the same, i.e., revenue neutral.

4282 That is not to say that the costs between ILECs will necessarily be identical, but for a given ILEC there should not be significant cost differences between a peak-capacity approach, a 95th-percentile approach or a total volume-based approach.

4283 Therefore, there should not be significant differences in terms of the impact on independent ISPs resulting from the method of measuring usage that is applied.

4284 If the choice of model results in a wide discrepancy in the revenues generated, that would suggest that the tariff structure is flawed, that the pricing bears no relation to actual costs, or both.

4285 It concerns us that Bell is attempting to charge more than once and in more than one way for the very same usage-sensitive cost elements under its AVP proposal.  In terms of usage, Bell's FTTN proposal contains the following rate elements:

4286 (1) an AVP usage charge of $200 per terabyte of pre-subscribed monthly usage, which works out to an effective rate of about 19.5 cents per gigabyte;

4287 (2) a charge of 29.5 cents per gigabyte for any usage that exceeds the pre-subscribed level in a given month; and

4288 (3) the same AHSSPI charge that is included in Bell's legacy GAS offering.

4289 Using Commissioner Morin's assumptions for 100 end users, if we restate all the usage components of Bell's proposed FTTN tariff into volume-based terms, a competitor would pay the following per gigabyte of data transferred in a month:

4290 - 6.75 cents per gigabyte, based on the AHSSPI charge;

4291 - plus 19.5 cents per gigabyte, for a total of 26.25 cents for pre-subscribed usage; and

4292 - potentially an additional 29.5 cents per gigabyte for excess volume above the pre-subscribed amount.

4293 For 1000 end users, the Bell cost per gigabyte would be 21.3 cents of pre-subscribed data transfer during the month -- 1.8 cents for the AHSSPI, plus 19.5 cents for the AVP charge.  And again, there could be the additional 29.5-cent charge for any excess volume above the pre-subscribed amount.

4294 In contrast, using the same assumptions to restate the AHSSPI charge under MTS Allstream's proposal, the price a competitor would pay for data transfer over a month would be 6.8 cents per gigabyte based on 100 end users or 5 cents based on 1000 end users.  In both cases, Bell's total pre-subscribed usage rates would be four times higher than MTS Allstream's.

4295 Likely this disparity is caused by the fact that neither Bell's 19.5 cents nor 29.5 cents per gigabyte charges are based on the cost of capacity.  Bell's pricing approach is not consistent with either of the Commission's two principles in this proceeding.

4296 Mike will talk to question 3.

4297 MR. STROPLE: MTS Allstream believes that our wholesale FTTN tariff is the best model to meet all of the Commission's objectives.  It is more transparent, it is administratively the most simple, it is well understood and it reflects the way networks are provisioned to avoid congestion.

4298 Under our model:

4299 - there is no feasible way for our wholesale customers, regardless of the number of heavy users that they may have, to exceed their allocated capacity; and

4300 - network capacity upgrades over time are built into the underlying assumptions through the use of survivor curves, working fill and other loading factors.

4301 Thus far, and despite the explosive growth of the Internet and of broadband services in general, MTS Allstream has been able to continue to make the necessary investments to our network to keep up with increasing demands for bandwidth.

4302 By continually monitoring our network utilization and investing where required on an ongoing basis, MTS Allstream is able to ensure that the potential congestion points, which are generally in the switching and routing equipment located at the links in the transport network, are able to handle retail and wholesale Internet traffic generated by all services that are transported over the network.

4303 There are important differences between our peak-capacity tariff structure and both the CNOC and Bell proposals:

4304 - both require that the ILEC measure a competitor's usage in ways that are administratively complex and would require us to make significant investments;

4305 - CNOC's proposal is a pay-as-you-go model, which could complicate network planning and provisioning since the competitor does not reserve capacity in advance;

4306 - Bell is attempting to charge in multiple ways for usage, which sets up the conditions for double charging competitors for the same network transport component and also results in a significant lack of transparency regarding which cost elements are being recovered through a given charge;

4307 - Bell's AVP charge does not distinguish between usage in peak and non-peak periods, so it does nothing to incent ISPs to manage congestion; and

4308 - Bell's rates are extremely high and do not reflect their costs to provide capacity.

4309 All ILECs should not be mandated to adopt the same billing model.  Regardless of which billing model is chosen by a given ILEC, there should be no significant differences between the billing models in terms of impacts on incentives to build out networks for either incumbents or competitors.

4310 MS GRIFFIN-MUIR: With respect to questions 4 and 5 there is no rationale for applying different mark-ups for different service providers nor for varying the level of mark-up between monthly access and usage rates.

4311 In the Essential Services Decision, the Commission found that services such as last-mile access are not economically duplicable by competitors on a widespread basis.

4312 Although the Commission did not reach the same conclusion with respect to the ILECs' wholesale aggregated ADSL access service, on the basis that these include network transport, the fact remains that only the ILEC providing the last-mile access in this case can provide the necessary network transport. Therefore, it would not be appropriate to incorporate higher mark-ups on the transport elements than is permitted on the access component.

4313 Finally, question number 6.

4314 Similarly, as the Commission found in the Broadband Access Policy, with the exception of the additional mark-up approved to reflect the greater risk associated with fibre-based investments, MTS Allstream does not see any valid rationale for treating legacy services differently from the services which are the subject of this proceeding.

4315 As you noted at the outset of this hearing, Mr. Chairman, maintaining a competitive marketplace in which Canadians have a choice of ISPs is one of the Commission's primary objectives.  This includes independent ISPs as well as the large incumbents.

4316 Ultimately, competition amongst service providers leads to variety, responsiveness and consumer choice, including all-you-can-eat pricing models.  Therefore, it is essential that the Commission set wholesale rates that fairly compensate the incumbents for network usage but at levels that will allow independent ISPs to compete efficiently.

4317 We believe that MTS Allstream's proposed wholesale FTTN tariff best achieves this balance.

4318 Thank you.

4319 THE CHAIRPERSON: Thank you for your presentation.  Now, let me make sure I understand it correctly.

4320 You start off by saying that all three models can work, but then if I understand paragraph 10 correctly, you really are opting for version three of those.  Are you?

4321 MS GRIFFIN-MUIR: No.  Actually we're -- at paragraph 10?


4323 MS GRIFFIN-MUIR: Our model would be peak-period reserved capacity.  So that's the version we're opting for.

4324 So essentially the third one would be total amount of data transmitted over a month measured in terms of bytes and that's Bell's model really.

4325 THE CHAIRPERSON: That's Bell's.  So number one is your model?

4326 MS GRIFFIN-MUIR: That's right.

4327 THE CHAIRPERSON: Now then you say in 11 -- let me just get my head around this:

"Since networks are provisioned on the basis of overall peak capacity requirements, we believe that a pricing method that reflects peak usage is most appropriate.  Under our proposed wholesale FTTN tariff, a competitor plans ahead and tells us in advance how much peak period capacity they wish to reserve on our network."

4328 Now if they do that and they guess wrong, that's their problem, they cannot buy additional capacity from you?

4329 MS GRIFFIN-MUIR: They can buy additional capacity, yes.

4330 How the tariff is structured is we have the access charge, which the rate doesn't fluctuate, that your end user accesses --

4331 THE CHAIRPERSON: Yes, right.

4332 MS GRIFFIN-MUIR: -- at a point between A and B, a single flat-rate charge.

4333 Then we have the interface or AHSSPI charge, which you can buy at transmission speeds of 100 megs, 400 -- or gigs.

4334 And so the competitor would look at what they think they would need in terms of the number of end-user customers they have and purchase on a monthly basis one of those three.  And if in fact they did not use -- they purchased 400 megs and did not actually need that much capacity in the month, they could change it the next month.

4335 THE CHAIRPERSON: What about the opposite, they underestimated and they need more?

4336 MS GRIFFIN-MUIR: Well, I mean they can definitely get more, but as --

4337 THE CHAIRPERSON: I mean what I'm trying to get at is do you have the same sort of penalty that the AVP model has, that if you need additional capacity, you guessed wrong, you pay for it?

4338 MS GRIFFIN-MUIR: No, because actually what would happen is you have the capacity.  If you exceed the capacity, you just -- it won't -- your customer or your service that you're providing would be degraded.

4339 THE CHAIRPERSON: It's their problem, they do not get additional capacity from you, they have to manage it --

4340 MS GRIFFIN-MUIR: That's correct.

4341 THE CHAIRPERSON: -- and it degrades their service or whatever or throttle?

4342 MS GRIFFIN-MUIR: That's right.  So they would have to manage the efficiency on their own network.

4343 THE CHAIRPERSON: Yes.  So basically you estimate how much you get, you get that at this fixed price, everything else if your problem?

4344 MS GRIFFIN-MUIR: That's correct.

4345 THE CHAIRPERSON: That's your basic approach.

4346 Now, you keep saying FTTN.  What about legacy?

4347 Bell made this very level pitch that the legacy models are established, everybody knows how they work, et cetera, and the AVP would only apply to FTTN, leave legacy in place, changing it or adopting a new structure for it is not worth the trouble.  That essentially, as I understood it, was their point.

4348 Do you share that?

4349 MS GRIFFIN-MUIR: Well, our legacy tariff is actually structured very much the same way as our FTTN, subject to a difference of prices.  But the structure is identical.  There's a flat rate for the access charge and then we have different interface speeds that you can buy to manage your capacity as a wholesale customer.

4350 THE CHAIRPERSON: In your case, it doesn't make -- but would -- that takes me to 29 and 30.  I'm not quite sure what you're saying.

4351 On 29, for one thing, you say:

"All ILECs should not be mandated to adopt the same billing model.  What matters is that whichever billing model is chosen by a given ILEC, there should be no significant differences..."

4352 So it seems to imply you envision different models for different ILEC.  But then in 30 you essentially go back the other way and say they should all be the same.  So I'm not quite sure where you stand on this.

4353 MS GRIFFIN-MUIR: Yes.  Actually in 29 and 30 we're only speaking about the actual mark-up over the cost.

4354 However, I would like to say that what concerns us most is we to date don't measure.  We don't measure our retail volumes the way in which Bell says they measure theirs, or our wholesale.  We measure the wholesale volume or the capacity at the interface or A-HSSPI point in our network.

4355 So if the Commission were to find that Bell can implement it's AVP proposal -- which we are not advocating -- then we don't want the Commission to decide: Okay, MTS Allstream, you too implement this proposal.

4356 THE CHAIRPERSON: Well, let me put it the other way around.  Let's assume we fall in love with your model, we think it's great, you should be doing it, and we say to Bell: Okay, you adopt the MTS model for your FTTN, but you can keep your legacy business, which is well established, which has 30 percent of your users already, et cetera, we won't upset that, keep it on your present basis, would that cause you any problem?

4357 MS GRIFFIN-MUIR: Well, actually no.  Actually, Bell's legacy model is very similar to our model here.

4358 THE CHAIRPERSON: That's what I'm getting at.  Bell made an impassioned plea basically don't touch legacy, it works, it's not broken, leave it --

4359 MS GRIFFIN-MUIR: Right.

4360 THE CHAIRPERSON: -- let's concentrate on FTTN, which whatever we adopt for FTTN by implication means, at least for Bell, there would be two separate systems.

4361 Let's assume we adopt your model, that would mean you have the same model all over, Bell has a slightly different legacy than FTTN.  That wouldn't cause you any problem?

4362 MS GRIFFIN-MUIR: The only problem it would cause me is Bell's model is not transparent for the FTTN.  Other than that it doesn't cause us a problem, but in fact Bell could have the same FTTN model as us and it shouldn't cause them a problem either, because their network -- our model is the ILEC network model.

4363 THE CHAIRPERSON: Yes.  Well, we will hear from them undoubtedly on Monday whether your model causes them problems or not.

4364 MS GRIFFIN-MUIR: Undoubtedly.

--- Laughter

4365 THE CHAIRPERSON: So that's the whole purpose of rebuttal.

4366 Okay, Michel, you have some questions?  I believe you like this proposal?

4367 COMMISSIONER MORIN: I have some.  Thanks, Mr. Chair.

4368 First of all, I will ask you some questions this morning, I will ask you all of them in English, but if you are unsure of my meaning, please let me know.

4369 First of all, since the beginning of this week I have provided cost comparisons of the different models and it appears that you are doing very well, much better in fact than other providers to small ISPs.

4370 My first question is: What brought you to this model?

4371 MS GRIFFIN-MUIR: Well, Mike can talk about it from a capacity costing perspective.

4372 COMMISSIONER MORIN: Well, it was a question of principle, cost?  What are your thoughts about this model fundamentally?

4373 MS GRIFFIN-MUIR: Well, fundamentally we were looking at the legacy model that all ILECs had used and we also look at how we provision our retail service -- because really we are just providing wholesale customers with the underlying facilities we would use to provide retail Internet service -- and figured out a way that is cost-effective for us to provide service and would enable them to provide a wholesale service where they manage their own capacity and we recover the cost of our usage on the transport network.

4374 COMMISSIONER MORIN: More specifically, what are the advantages and the disadvantages of your model over those proposed by the other ILECs, the cable carriers or the CNOC model?

4375 I understand many things that you have said this morning, but could you add something about this?

4376 MR. STROPLE: I think the primary advantage of this model over the others -- and this model is very much based on an ILEC network so a comparison to cable is a little more difficult, but on the ILEC basis the primary advantage is predictability and that is, I believe, predictability for both the service provider and the competitive ISP, in that capacity is determined initially and then it is properly costed out, provisioned accordingly.  We know what we have to provide, they know exactly what they are going to get.

4377 With the CNOC model, it is very similar in that it also uses the basis of peak capacity, which is the fundamental building block around which networks get built and costed.  So it's similar in that it uses peak capacity, but its variability provides challenges for both parties, I believe, in trying to get the right level in there and is maybe more appropriate when two parties are acting at arms length in a more competitive environment.

4378 Comparing it finally to the Bell model, it is effectively the same in that the three we outlined the Bell model also uses the same principles of an access charge plus an A-HSSPI.  The difference is the addition of the AVP on top and we believe that also adds the opportunity or the potential to charge twice for the same capacity and introduces the variability or the unpredictability on both sides.

4379 COMMISSIONER MORIN: Yesterday I understood that Shaw, it's a cable operator, said that this model doesn't work for them.  As you know, up to now more than 90 percent of all the clients with the small ISPs are with the telcos.

4380 So do you think, first of all, that this model will be suitable to the cable operators and, if you look at the future, that they can adjust or they can agree with your model, which is the simplest, but it will work eventually also for them?

4381 MS GRIFFIN-MUIR: Mike will talk to the network part, because I do think, as Mike said, the cable network doesn't look like the telephone network.

4382 I think there are a couple of reasons which have been raised throughout the proceeding as to why there are more customers on the ILEC wholesale service than the cable service.

4383 One is obviously the reach of the telco network into res and bus.

4384 In terms of both, I mean a lot of why decisions are made to use one or the other has a lot to do with the price points and the business case, as Distributel showed.

4385 In terms of the network, I will let Mike speak to that because it may not be appropriate to try and impose exactly the same structure.  Conceptually potentially, but not physically.

4386 MR. STROPLE: One of the primary differences between the cable network and a DSL-based network, in the DSL-based network the access charge can be fixed.  In the cable network where the entire -- the access is also shared, you get the variability across that entire network.  So where we can segregate in the diagram "A" and "B" being fixed and "B" to "C" having some level of variability in capacity, the cable network is somewhat different in that regard.

4387 The other important aspect is the difference between using peak capacity as the building block versus total volume.

4388 Total volume can also be used to approximate network capacity and investment, but since I plan networks I need to use peak capacity is how I need to use it.  If I have a volume-based measure of how many gigabytes get transferred across the network, what I immediately do is I translate that into what it probably means to peak.

4389 So from a planning perspective you are most closely aligned with the planning process, which means you are most closely aligned with cost if you work in peak and peak as opposed to trying to translate average volume into what it might mean for peak.

4390 COMMISSIONER MORIN: Now, explain to me how your price model satisfies the principles enunciated in our Notice of Consultation 2011-77 in order to ensure that average Internet users do not subsidize the heaviest users and allow competitors to differential themselves?

4391 MS GRIFFIN-MUIR: Well, first, I think since we -- and hopefully all the other ILECs -- build up from our capacity costs, the base, where we strike it is being able to -- between the areas of "B" and "C" on the diagram, and all those links, we determine what it would actually cost us and from there determine how much we would need to support our network and what is a fair price to charge competitors so that they can offer service in a differentiated way while we can still continue to invest in our networks.

4392 So that by giving them the ability to manage their own capacity while we still recover our cost, we allow them to determine how they want to price in the retail market, thereby offering choice to competitors.  We allow ourselves and them to manage the traffic on their network so that they can make allowances for when the peak is, as well as allowances for how much they want to charge heavy users versus non heavy users.

4393 COMMISSIONER MORIN: So the small ISPs will have the latitude to decide whether to subsidize heavy users through their average users.

4394 Would you say that it allows small ISPs to put more business models in place than with the other models presented by Bell, cable operators and CNOC?

4395 MS GRIFFIN-MUIR: Yes, definitely we think that our model -- Mike can speak to the differences between the two peak capacity models, which would be the CNOC model and our model, but in terms of our model versus Bell's model, the concept of volume-based pricing is really a retail construct and it's more if you want to charge per usage you have to be able to give the retail customers some sort of measure that they can look at.

4396 So if you think of it in terms of long distance pricing or wireless voice pricing, it was a minute and the customer can actually identify how many minutes they are using.

4397 So when you turn to volume-based pricing for wholesale, you are imposing kind of the same kind of retail construct on the wholesale customer.

4398 So rather than looking at it from a peak capacity perspective and pricing the capacity that the wholesale customer buys, and allowing them then to manage that, you are imposing another constraint on them that will definitely limit how much flexibility they have in their retail offers, because they always have to be cognizant of the actual volume each end user is consuming at any point in time, not just the peak point.

4399 So we think as long as the wholesale ISP purchasing service under our tariff manages their network properly and makes sure that they most efficiently utilize the capacity they purchase, they will have the flexibility to offer whatever pricing plans they want.

4400 I will let Mike just talk to the CNOC versus ours in terms of predictability.

4401 MR. STROPLE: With the CNOC model it really comes down to a predictability relationship and setting those right levels and I believe this model that we have proposed provides more predictability for both sides.

4402 COMMISSIONER MORIN: In your filings you have specified three tiers of capacity, 100 megabits per second to megabytes per second; 400 megabits per second, which is to say 50 megabytes per second; and 1,000 megabits per second that can be purchased by an independent ISP.

4403 Explain why these tiers were chosen.  How did you get to these tiers?

4404 MS JESSOME: Because we provide on the access up to 32 megabits and the competitor can choose anything up to the 32, which is our maximum, we just looked at what capacity would be required to serve a reasonable number of customers if they were all at 32 megabits and it seemed that 100 megabits was the lowest A-HSSPI size that would really accommodate that.

4405 On our legacy service we have some lower speeds of A-HSSPI size, but we have, I think, 10, 30, 50 and 70, then 100, 400, 1,000.  So the 100, 400 and 1,000 was just picked to mirror our legacy service.  We didn't go any lower than 100 because we didn't think it would serve a reasonable number of customers.

4406 MR. STROPLE: Those numbers also have the basis on the underlying interface speed, so they are convenient speeds that the two, us and the other carrier, would naturally interconnect at.

4407 I might point out, you could buy multiple 100s.  So if there was a desire for 200, you could buy two 100s.

4408 COMMISSIONER MORIN: So how much more difficult would it be to provide customized bandwidth capacity?

4409 MS GRIFFIN-MUIR: Part of how we costed out the model was based on the fact that we have just one access charge.  That doesn't reflect -- I mean, as Pauline was saying, provides up to the 32 megabits, which is the highest speed we currently offer.  So when the wholesale customer buys an end user access, they don't have to keep coming back to us if they want to change the service speeds they want to offer their customers.

4410 But in order for us to ensure ourselves that we will provide them sufficient capacity through, we estimate the cost between BNC based on how much -- the transmission speed they can actually provide their customer, or the maximum transmission speed they can provide their customer.

4411 So as you introduce different layers or customize it, I mean it would change, first the price of the tariff because there would be just a little more work involved in making sure that we can actually handle at peak capacity the different usages.  The same for them I would think, they would have to think about it more.

4412 So it's a simple model, it gives them a lot of flexibility, the wholesale customer, but it gives us a certain amount of assurance that we are actually recovering our costs.

4413 COMMISSIONER MORIN: How does the company limit the traffic to the capacity specified by the independent ISP?

4414 MR. STROPLE: We initially provision based on the A-HSSPI order and then we continually manage the network that includes our internal traffic and competitors traffic and watch for capacity choke points in the network.  We augment, before we get to capacity, choke points and manage the network to that level to ensure that that capacity is made available.

--- Pause

4415 MR. STROPLE: The question is: How do we ensure that the capacity purchased is not exceeded?  We physically rate limit that port.  So if you buy a 100 meg port, only 100 meg will go across that port.  If there is more traffic coming in from the end user's subscribers from either direction, that port won't pass more than the purchased amount.

4416 COMMISSIONER MORIN: So, for instance, if the threshold limit of 12.5 megabytes per second, or 100 megabits per second, is crossed by ISP and users, what will happen to the customer's service?

4417 MR. STROPLE: As the aggregate group of customers approach that 100 megabits each of them will see traffic degradation, they will see packet loss, they will see traffic slow down and then all of them in aggregate, depending on what traffic management processes the ISP may have implemented, will see their service change.

4418 COMMISSIONER MORIN: Another question, of course if something happened this way, what would happen if an independent ISP required more capacity during a month than it had paid for?

4419 MR. STROPLE: They would need to order a larger A-HSSPI and there would be a timeframe for us to implement that larger A-HSSPI.

4420 So they could place that order and then we would fulfil the increase and it would depend on what it was they were increasing from and to.

4421 COMMISSIONER MORIN: Last question.  Proposed mark-ups.

4422 Wholesale service mark-ups that are too high might result in rates that unduly restrict competition, while mark-ups that are too low might not provide a reasonable contribution to the network providers for the users and growth of their network.

4423 How do the mark-ups that you have proposed, or that are already included in your existing wholesale services, balance these two competitive objectives?

4424 MS GRIFFIN-MUIR: Well, that has been the question I guess through the essential proceeding and subsequent to that.

4425 Generally speaking we have followed the -- as you know, we have been very big proponents of more cost-based wholesale access, however in some instances we have been successful and in most we have not.  We basically follow what we think the rest of the industry is following in terms of the trend on the mark-up.

4426 We recognize there is a need for a balance and certainly to the extent the Commission wants to impose a specific mark-up, as long as it does cover off our costs, we don't have a huge objection.  But right now we are following the essential facilities pricing principles.

4427 COMMISSIONER MORIN: Thanks very much.

4428 Those are my questions, Mr. Chairman.

4429 THE CHAIRPERSON: Thank you, Michel.

4430 Len...?

4431 COMMISSIONER KATZ: Thank you, Mr. Chairman, and good morning.

4432 Is your model different from what Primus advocated yesterday as their alternative?

4433 MS GRIFFIN-MUIR: Just listening to Primus it doesn't sound different.

4434 I mean essentially what the witness for Primus was saying, as their alternative they would buy a peak amount -- so 100 megabits -- and sustain that for the whole month, so pay for it for the whole month, and manage their own traffic, manage their own efficiency.

4435 So I would say it's the same really as what they propose.

4436 COMMISSIONER KATZ: Because it reads very similar.  I have their piece right here.

4437 So you agree it is virtually the same?

4438 MS GRIFFIN-MUIR: Well, not knowing precisely, but based on what they said yesterday, yes, I would say it's the same.


4440 In your current business -- and I guess I will use the word "Allstream" here -- you sell capacity to other carriers; is that correct?

4441 MS GRIFFIN-MUIR: Yes, we sell capacity.  Are you speaking backhaul capacity?


4443 MS GRIFFIN-MUIR: Yes, we sell backhaul.

4444 COMMISSIONER KATZ: How do you sell it?  Do you use peak period capacity costing?

4445 MS GRIFFIN-MUIR: Well, yes.  We would always figure out exactly how much it cost us to provision to the peak, so we would always use that as our base measure.

4446 COMMISSIONER KATZ: So as an example, if Primus bought backhaul facilities from you for other parts of the business, for some of their private customers, do they buy a pipe and do they manage their own network against that pipe the same way as what you are advocating here?

4447 MR. STROPLE: Inter-carrier arrangements are size of the pipe as opposed to aggregate transport through the pipe.

4448 COMMISSIONER KATZ: Yes.  So if we came up with a different way, whether it's 95th percentile, the AVP, or something else, if you were selling wholesale services to a customer, you would have different pricing models, if we instituted a different pricing methodology, than what you are currently doing today for other carriers.

4449 MS GRIFFIN-MUIR: I guess Mike could speak to how you would look at the network.  What we feel is, the way we price is the way we would actually look at costing out how we provision our network and how much capacity we have available.

4450 So it would introduce a difference, but as Mike said, the bigger the customer, you might have a different way of providing them with service than in a small retail -- or wholesale ISP.

4451 But, yes, for us, we look at it -- this is the way, when you are provisioning a network, we would cost it, and this is the way we would manage it, to make sure that we have sufficient capacity to carry our traffic and whatever other traffic we provide our wholesale customers.

4452 COMMISSIONER KATZ: Is your traffic and their traffic distinct?

4453 In other words, is it fungible capacity, or is your capacity dedicated and their capacity dedicated?

4454 MR. STROPLE: It's fungible.

4455 COMMISSIONER KATZ: It's fungible.  So when you provide IPTV services to your customers in Manitoba, it is using the same capacity, the same pipe.

4456 MR. STROPLE: It's using the same sets of routers, the same underlying transport, yes.

4457 COMMISSIONER KATZ: So if you have an issue with capacity, does that domino effect filter down to them as well, or are they held to what they have bought and there is no change to their quality of service, if I can call it that, if your customers are using too much capacity, if I can call it that?

4458 MR. STROPLE: If that were to happen, then, yes, because they use the same aggregate network, they would be subject to that problem.

4459 COMMISSIONER KATZ: In your agreements with your customers, do you have quality of service measurement commitments and obligations?

4460 MR. STROPLE: Yes.

4461 COMMISSIONER KATZ: You do.  So if you get below that level of service to them, there is recourse -- they get a discount, they get a payback, there is something there, or is it a best efforts basis?

4462 MS GRIFFIN-MUIR: It would depend on what was negotiated, but certainly a customer of any size who, obviously, also has made commitments, there would be some recourse, depending on what was negotiated.

4463 But, typically, we would like to get -- and, obviously, people buying from us would like to get -- some assurance of quality of service.

4464 COMMISSIONER KATZ: So your network is configured somewhat similar to cable, because the cable network said that it's one big pipe, as well, and they can't dedicate a certain amount of capacity to a customer the same way that Bell's network is engineered, and they are saying it's dedicated.

4465 MR. STROPLE: There are a number of ways -- there are levels of segregation, but it comes to the question of when you ultimately get congestion or capacity, and exactly what happens.

4466 So we do segregate traffic certain ways inside the network.  We fundamentally carry all the traffic from the Central Office as it makes its way back through the network, and some of the IPTV traffic will get peeled off at one point, the Internet traffic will get peeled off at one point, leaving the AHSSPI traffic.

4467 Because they use the same fundamental and shared network, the question is, where do you take off each of those pieces of traffic.  I saw Bell's diagram, where they take their IPTV traffic off sooner than our traffic.

4468 It's kind of a matter of efficiency, size and scale, and which size router fits better and where, as to where it makes the most sense to take off which piece of traffic from that aggregated and shared network.

4469 COMMISSIONER KATZ: So they are able, in their model, to provide a dedicated, guaranteed pipe, basically, whereas the cable industry is not, and you are saying that you are closer to the cable industry than the phone companies -- than Bell?

4470 MR. STROPLE: I think the difference with the cable industry comes in the access side not being dedicated and that also being shared.

4471 Within that shared network, there are ways to dedicate segments of traffic, so that quality is absolutely guaranteed.

4472 I guess, rather than absolutely guarantee it, we use the traffic management approach to ensure that, for our customers, or for, by extension, wholesale customers, by appropriately managing the network to avoid congestion, we eliminate that concern.

4473 COMMISSIONER KATZ: I think that is one of the issues that we will have to deal with, probably, next week; that is, if someone is buying services in order to resell, to what extent can they have the maximum flexibility in using that service, without having a third party -- the supplier actually capable of influencing their quality of service to their customers.

4474 So think about that for next week.

4475 I have two more questions.  At paragraph 12 of your remarks you talk about the fact that you currently don't use the 95th percentile for volume, and you wouldn't want to, because it would be costly, I guess.

4476 Have you taken a look as to what the cost implications would be if this Commission chose to go with that model?

4477 MS GRIFFIN-MUIR: When you say 95th percentile, we would measure our costs to the 95th percentile.

4478 COMMISSIONER KATZ: It's the CNOC model.

4479 MS GRIFFIN-MUIR: Right, yes.

4480 No, we haven't -- to actually put in place and measure traffic for the competitor, as they are using it throughout the course of the month, no.  I haven't looked at that, no.

4481 COMMISSIONER KATZ: I am sure that will be something that will come up next week, as well, so you may want to think about that, because, obviously, timing, implementation and cost all play a role in whatever decision this Commission will ultimately decide.

4482 My next question is one that I am confused about.  It's in your paragraph 16.  You talk about the peak period capacity costing, and in there somewhere you say, "...including adjustments to reflect the fact that the shared network is not intended to and does not operate at 100 per cent capacity."

4483 I recall Mr. Stein, of Primus, yesterday saying that he believes that the phone company's response about gaming is incorrect, because when you buy the pipe, you buy all the pipe, and the fact that he can move traffic around and change the peaks by managing his customers should not be deemed to be gaming because he has bought the whole pipe.

4484 What you are saying here is, you may have bought the whole pipe, but you are not paying for the whole pipe.  I think that is what you are trying to say.

4485 I am just trying to understand what you mean by that.  Is it the issue that somewhere in these models there is a percent utilization being employed and it's not being costed at 100 percent?

4486 MS GRIFFIN-MUIR: No, actually we are saying -- if you think of it in terms of, if the pipe costs $100, we know that we would never, 100 percent, fill it, because we have to manage our traffic volumes to the peak, and you might have a little burst of traffic over the peak.  So you just want to have that buffer.

4487 It is really how we price the service.

4488 We know what the capacity will fill the pipe to.  It's almost like, how much will fill the pipe is how we decide what we will charge for the pipe.

4489 When we sell Primus 100 megabyte service, he can manage that service, and we would, by and large, assume that he can manage to 100 percent of his capacity.  We are not going to take it from him.

4490 And we are just recovering our costs overall based on some estimate of our fill of the pipe.

4491 COMMISSIONER KATZ: What adjustments are you referring to?  You say, "...including adjustments to reflect..."  What adjustments would you have to make if you pay for the whole pipe?

4492 MS JESSOME: When you are doing capacity costing, you augment before you ever reach 100 percent capacity in your network overall.  There are different measures, depending on when you would start to augment your routers and switches, pull in more fibre, that kind of thing.

4493 When you are doing capacity costing, that is what fill factors are meant to represent.  You never let your network get to 100 percent capacity.

4494 So that's the rationale for using fill factors.

4495 It's not that you can't -- if you buy 100 megabytes of capacity, that you can't use --

4496 COMMISSIONER KATZ: I understand what a fill factor is.  I understand that fully.

4497 So what you are saying here is, it's the assumptions you make on the fill factors, and the multiple providers that are out there, the more you can utilize those spare fill factors, because it's an aggregation of all of them.

4498 MS JESSOME: Yes.

4499 COMMISSIONER KATZ: Those are my questions.

4500 THE CHAIRPERSON: Thank you.

4501 Let me just clarify something in your answers to Commissioner Katz.

4502 You saw Figure 3 from Bell, where they show that the IPTV network is totally separate and it only links up at the Central Office.

4503 If I take your drawing today, how would you put in the IPTV network?

4504 Obviously it is different from Bell, if I understood your answers to Commissioner Katz.

4505 MR. STROPLE: In the network shown here, it would come off where it shows "Backbone Network".

4506 THE CHAIRPERSON: Excuse me?

4507 Describe for me where it starts and ends, your IPTV network, so that I can draw it in.

4508 MR. STROPLE: Sure.  The IPTV network, starting at the end user, it traverses through the same DSLAMs, it would traverse through the same aggregation in Metro Network, and would come off somewhere inside what is shown here as the backbone cloud.

4509 For simplicity, we adopted the same diagram that you had seen before to illustrate the A, B and C points, as opposed to attempting to show you a more detailed and maybe more accurate depiction of our specific network.

4510 THE CHAIRPERSON: And why is your network so different from theirs?

4511 You are both ILECs, you are both offering IPTV, so why such a very different configuration?

4512 MR. STROPLE: I don't think it is fundamentally different.  I think the only difference is where you take which piece of traffic off.  In their picture they have taken it off earlier in the network and decided to manage it on a parallel network, and we have taken the approach that we prefer to manage it in one big network, capitalizing on the efficiency of larger routers.

4513 I don't think that either is wrong or fundamentally different.  I consider it a design choice.

4514 THE CHAIRPERSON: But, then, if you are worried, as Bell is, about capacity and congestion, and you want to make sure that everybody pays for their part of the augmentation, how do you segregate out what is driven by IPTV and what is driven by Internet access?

4515 MR. STROPLE: We do have segregation methods that we have employed, and we could put in additional segregation methods, as I mentioned, to ensure -- even on the same router, to ensure that, absolutely, that has been protected, and it has not become a problem in our traffic management to date.

4516 THE CHAIRPERSON: And, also, are the costs allocated separately?

4517 MR. STROPLE: Yes, in our costing model that is accounted for.  By putting them together you have the advantage of managing that capacity as a whole, as opposed to two independents.

4518 THE CHAIRPERSON: Everybody is using the Bell diagram sort of as a starting point.  I would appreciate if, for Monday, you could provide us with an MTS diagram which shows how your IPTV fits in.

4519 MS GRIFFIN-MUIR: Yes, we will do that.

--- Undertaking

4520 THE CHAIRPERSON: Thank you.

4521 Tom...

4522 COMMISSIONER PENTEFOUNTAS: Very briefly, I have two follow-ups.  One is on the cost of implementing the CNOC model.  Would you be in a position to have that for us next week, as a follow-up to Commissioner Katz's question?

4523 MS GRIFFIN-MUIR: We could certainly try to have, at least, a high level estimate next week.

--- Undertaking

4524 COMMISSIONER PENTEFOUNTAS: That would be great.

4525 Secondly, what motivates your disinterest in a surcharge for people that go above and beyond what they predicted their monthly usage to be?

4526 Is it a network capacity issue?

4527 Because Bell offers -- you buy the block, and above and beyond that there is a surcharge, if you need more.

4528 That is also a follow-up to one of the Chair's questions.

4529 MS GRIFFIN-MUIR: What Bell is proposing is actually -- you buy a block, and if you use more, or if your end customers use more, they will penalize you.

4530 That's how I would look at it.

4531 COMMISSIONER PENTEFOUNTAS: We will offer you the chance to buy more, but at a premium.

4532 MS GRIFFIN-MUIR: I think, actually, you have just gone over.  They are not offering you a chance to buy more, you have actually used more, so they are charging you a premium for using more.

4533 In our case, we are selling the block, and they actually can't use more unless they buy more.

4534 MR. STROPLE: Just to add to that, in the Bell model they have the same AHSSPI construct as us, with the same properties about -- you buy it, and you get it, and you can't use more unless you order more, and it gets augmented.

4535 In addition to that, then they have the volume-based charge, based on gigabytes transferred, where you predict and then you pay more or less.  But to the point where, if the independent ISP, regardless of the number of gigabytes transferred and predicted, exceeded that AHSSPI, the end ISP would be in the same position that they would be with us, which says that you can't get more on the AHSSPI, regardless of where you may be on total gigabytes transferred, because those two don't necessarily -- while they are correlated, they don't necessarily connect.

4536 COMMISSIONER PENTEFOUNTAS: And if they wanted to buy a bigger block, would that be available for them the very next month?

4537 Would you need more time than that to make it available is my question.

4538 MS GRIFFIN-MUIR: Yes, it would probably take about six weeks.

4539 As the witness from Primus was saying, they are managing their network.  So, just like us, we couldn't -- like, Mike's people would not be day-to-day being surprised, hopefully, by what is -- you know, suddenly the network is congested, or our customers are using far more than we ever anticipated.

4540 It is over a period of time, we are watching the capacity used on the network, wholesale and retail, and making sure that we are augmenting our network at a pace that won't get us into trouble.

4541 The same is true for the ISP.

4542 So, when Bell is designing this surcharge, especially the way they have designed it -- because they could actually control the volume of the wholesale customer in the same manner as we are.  You purchase, it's very predictable.  We know you have bought 100 megabytes, and that is what you are allowed to use.

4543 Now, if you, as the ISP, sees, "Wow, my volumes are really increasing," it should not come as a huge surprise to you either.  You should plan and have time to manage that piece of it.


4545 THE CHAIRPERSON: Candice...


4547 Just so I understand your access, you are making available, or proposing to make available, a maximum available speed?

4548 MS GRIFFIN-MUIR: Actually, in our legacy tariff we did the same thing.

4549 We don't have any distinction, and several of the other ILECs have distinction, as to transmission speed, what your access price is.

4550 We offer one price for res and bus, or propose to offer in the FTTN.

4551 From a provisioning perspective, when we were actually modelling the legacy tariff, we just tried to balance between what is most cost-effective for us, in terms of provisioning, and at the time you order the service, whatever our maximum speed is.  The port is available at that speed, and you, the wholesale customer, decide whether or not you want to make your service available to all of your customers at that speed, or at some lower increment of that speed.

4552 COMMISSIONER MOLNAR: Fair enough.  And I know that you are not the only ILEC that offers a maximum available speed option.

4553 How do you define the difference?  How do you segregate, going forward, the legacy from what is defined here as FTTN?

4554 MS GRIFFIN-MUIR: In the legacy tariff the speed would go up to -- I think, now, 7.5?

4555 MS JESSOME: It's 7, but it's a different network.  It's over the ATM --

4556 MS GRIFFIN-MUIR: Right, but with the legacy, we had the speed set.  So any customers who are on the ATM network are established at a different access speed than those who are on, let's call it, the FTTN network.

4557 COMMISSIONER MOLNAR: I am trying to understand, as well, as we move forward into your network --

4558 Well, tell me.  In the transport network, is there segregation between legacy and FTTN traffic?

4559 MR. STROPLE: The legacy traffic is separated and comes off the network, effectively, at Point B, at the first point of entry into the network, and goes on a separate, different network than would be shown here.

4560 COMMISSIONER MOLNAR: Are the interface ports separate and distinct?

4561 MR. STROPLE: Yes.


4563 MR. STROPLE: From the first aggregation router that they hit in the Central Office.

4564 COMMISSIONER MOLNAR: So, for an independent ISP seeking your service, it would be clear and distinct for them that they would be provisioning interface ports for legacy services, and provisioning for FTTN separate and distinct?

4565 MR. STROPLE: The AHSSPIs would be separate and distinct.

4566 COMMISSIONER MOLNAR: Oh, they are separate?  The interfaces themselves are separate?

4567 MR. STROPLE: Yes.

4568 COMMISSIONER MOLNAR: For legacy versus FTTN?

4569 MR. STROPLE: That's right.

4570 COMMISSIONER MOLNAR: Okay.  That helps me understand how this would work.

4571 So there is no ability for them to consolidate traffic and use a single interface port.

4572 No, it's all separate.  Thanks.

4573 I want to go to paragraph 10 of your statement.  You say that "we and other ILECs" proposed this model.  What other ILECs proposed this model?

4574 MS GRIFFIN-MUIR: Actually, most ILECs, including Bell, have different -- when we say "this model", we are saying the AHSSPI model.  That is where you are purchasing capacity and reserving capacity.

4575 The difference, probably, in the proposal after that is where the allocation of the capacity cost is.

4576 For example, some of the ILECs actually capture some -- or recover some of the capacity as part of the access charge, and some recover a combination of access and AHSSPI charge.  But in terms of the tariff elements -- and that's what we were referring to -- and structure and how the wholesale customer buys the service, there is an access charge and an AHSSPI, or aggregation, interface charge.

4577 SaskTel has something similar, TELUS -- even Bell has all of those tariff elements, and then their AVP charge.

4578 COMMISSIONER MOLNAR: It was confusing to me because, the way I look at this, I don't see that others have proposed this application.  We have some that have proposed to incorporate the usage into the access charge.  I understand that everybody has proposed a charge for interface, and that existed in legacy, as well, but I think you are unique from what others have proposed.

4579 MS JESSOME: In terms of access, to determine what usage would be included in that access component, you would have to make some estimate of what your peak period demand would be on the service.

4580 MS GRIFFIN-MUIR: I think, though, if you are talking tariff structure --

4581 COMMISSIONER MOLNAR: Perhaps it's not relevant for us to keep going.  I think I understand what you have proposed, and hopefully I understand what others have proposed.

4582 MS GRIFFIN-MUIR: I think, though, that it is really a question of the pricing of each of the components.  In terms of, actually, how you estimate the underlying cost and what the wholesale customer is buying, they are buying an access and they are buying an interface.  That includes capacity.

4583 COMMISSIONER MOLNAR: Right.  The networks are similar, the components that are fixed versus variable are similar --

4584 MS GRIFFIN-MUIR: That's right.

4585 COMMISSIONER MOLNAR: -- so, ultimately, those costs are going to go into one of the different components available.

4586 MS GRIFFIN-MUIR: That's right.

4587 COMMISSIONER MOLNAR: That's what you are saying.

4588 MS GRIFFIN-MUIR: Essentially, yes.

4589 COMMISSIONER MOLNAR: Help me.  I believe you said that your model is similar to the Bell legacy model.

4590 Did I hear you say that?

4591 MS GRIFFIN-MUIR: Yes, but in the same way that I am describing this.  There is an access component and there is an interface component.

4592 In the Bell legacy model, though, you are right, they actually have usage in the price for access as well as in the AHSSPI, and ours is only in the AHSSPI.

4593 That is, frankly, why we think ours is superior, because it is very clear where you are paying for the usage component.

4594 COMMISSIONER MOLNAR: Right.  Okay, thank you.  Those are my questions.

4595 I just want to say that, you know, a couple of years ago I know you came in front of us in a different proceeding and explained your legacy tariff.  I was somewhat confused by it because it was very -- in my view it is not similar but quite different from the Bell tariff and it confused us.

4596 So again, I mean, I'm pleased to have this come forward.  Thank you.


4598 COMMISSIONER DENTON: Good morning.  I am suffering a confusion arising from paragraph 19, and if you would be kind enough to straighten me out I would be pleased.

4599 It said basically about the middle of the paragraph "but for a given ILEC, there should not be significant cost differences" between the various models.

4600 And to my recollection, Mr. Stein came in yesterday and gave us a very elegant demonstration that different patterns of consumption under different rules would generate different consequences.

4601 So I'm confused about something.  This appears to contradict what Mr. Stein said and help me out here.

4602 MS GRIFFIN-MUIR: I think perhaps we are talking about two different things.  In paragraph 19 we are actually saying the cost to provision your network at peak capacity should not -- I mean is whatever that cost is once you take into account all the traffic.

4603 How you decide to charge for it then -- so assume -- if you go to our example where we just broke down our service into a per gigabyte charge to compare with Bell's per gigabyte charge, all we are saying is once we have costed out what peak capacity is, how we choose to express it --

4604 COMMISSIONER DENTON: How you choose to bill for it, perhaps?

4605 MS GRIFFIN-MUIR: That's right.  Sorry, that's actually a good way of putting it.  How we choose to bill for it should not vary widely.

4606 So if we choose to charge a monthly AHSSPI charge versus a per-gigabyte charge it should be roughly -- you should work it out the same.

4607 I think what Mr. Stein was saying, which I agree, depending on how you actually impose or your whole tariff structure for your wholesale service is constructed will make a difference, like how the wholesale customer has to actually buy the capacity and pay you for the capacity will definitely make a difference to the wholesale customer.

4608 COMMISSIONER DENTON: Okay.  Then I don't understand you because, as far as I can tell, what he was demonstrating was that how different patterns of consumption are billed will result in a different size bill.

4609 I understand what you -- I believe you were saying that if it's costed properly under any of the three methods it will have the same effect but how it is billed will have dramatically different effects was how -- what I thought he was saying.

4610 And I remain of that view, that he demonstrated that the billing for it would be different, would have -- how you billed for it would have different consequences on the incentives of the third-party wholesaler.  I believe that was demonstrated and I believe you are contradicting that here.

4611 MS GRIFFIN-MUIR: Well, I think -- I'm not sure.  I don't remember whose example, to be perfectly honest.

4612 THE CHAIRPERSON: When we are going to come up to it at rebuttal that's precisely for the rebuttal to find out whether MTS and PRIMUS are saying something different.

4613 COMMISSIONER DENTON: That would be interesting to know.  So we will just leave it, put an asterisk beside it in saying, "Are you saying something different from Mr. Stein here?  If not, why not?" and be interested.  Inquiring minds want to know.

4614 Thank you.

4615 THE CHAIRPERSON: Okay.  Before I let you go, let me just ask two clean-up questions here.

4616 It strikes me when having listened now to both the cable companies and Bell that there are sort of two fundamentally different approaches.

4617 The cable companies say basically as far as wholesale, come on in.  If you want -- it's great if people want to use the Internet.  Use more -- as much as they want, no problem at all.  Just make sure that they pay for what they use.

4618 Bell's approach seems to be based on we need to sort of put some discipline in the system and some incentives and make sure that users are fully aware of how much they use and adjust -- in effect, discipline use.

4619 Where do you fit in between those two?  Are you right in the middle?

4620 I am not quite sure.  Those are some very different approaches to the same problem.

4621 MS GRIFFIN-MUIR: Well, we -- I guess we --

4622 THE CHAIRPERSON: I mean you welcome wholesale --

4623 MS GRIFFIN-MUIR: -- we are saying the wholesale customer should manage their own capacity.  So they look at what our speeds are; the transmission capacity we are selling.

4624 They purchase what they believe they need and then from that point they would manage their capacity.  And if they made a mistake they will have to get more.

4625 THE CHAIRPERSON: But basically are you welcoming wholesale customers in saying, "Fine, as long as you pay you can have as much as you want" or is this something sort of we have to do because you are an ILEC and we impose this on you?

4626 MS GRIFFIN-MUIR: Well, why we have the tariff.  I mean, we are mandated to have the tariff.

4627 However, certainly we would look at it as if we are going to lose the customer having the customer on our network is better than having the customer on somebody else's network -- the wholesale customer.  So I mean -- and our tariff, we are not actually trying to impose some sort of discipline on the wholesale pricing market which Bell's tariff is obviously trying to do.

4628 THE CHAIRPERSON: But you use more than -- you actually used the words penalizing, et cetera, that you feel that the AVP model, the way it is -- if I understand it correctly, what you object to is the fact that the wholesaler if he misestimates and goes over the ceiling he will have to pay and he will have to pay at a fairly increased or steep price.  That's essentially a penalty.

4629 Under your model, sorry, you are not getting any more and you deal with it.  That's sort of the main different philosophical approach to the issue, if I understand it.

4630 MS GRIFFIN-MUIR: Essentially.  I guess we would look at the Bell model as saying there is an attempt to control what the wholesale service provider can do in the retail market through the way that pricing is imposed.

4631 So instead of just saying, "Here is your capacity and you decide how you want to sell it, but you have to stay, you know, within this construct or this capacity amount that you have purchased", the Bell model adds another layer where it's not that transparent to the wholesale provider and where it probably doesn't have too much impact on the Bell network in the way they perceive it might, but where the wholesale provider has to be cautious about what they do in the retail market.

4632 THE CHAIRPERSON: You make it sound sort of a revenue grab more than anything else.

4633 MS GRIFFIN-MUIR: Yeah, m'hmm.

4634 THE CHAIRPERSON: Okay.  I guess we are going to have an interesting discussion on Monday on this point.

4635 Thank you very much.  Look forward to seeing your revised graph on Monday.  Thank you.

4636 We will take -- I gather we are waiting for some intervenors.  Let's take a 10-minute break before we start.

--- Recessed at 1020

--- Resumed at 1041

4637 THE CHAIRPERSON: Okay, allons-y.

4638 THE SECRETARY: So, Mr. Chairman, we will now proceed with the presentation by Mr. Jatinder Bhullar, World Broadband Foundation, and Mr. Claude Hénault.

4639 However, Mr. Bhullar has not arrived yet.  So we will proceed with the World Broadband Foundation and Mr. Munly Leong representing the World Broadband Foundation.

4640 Mr. Leong, you may now proceed with your presentation.  You have 20 minutes.


4641 MR. LEONG: Okay.  Thank you, Chairman and Commissioners, for the chance to present today.

4642 My name is Munly Leong and I am the founder of the World Broadband Foundation which will be a new party before many hearings because it's actually only formally been established for two months even though there have been lots of prior months in planning conception.

4643 The foundation, already made up of directors across three countries, is in the early stages of creating a world broadband map as a one-stop reference point for diverging broadband standards around the world.  That is the current chief purpose of the foundation.

4644 Not merely intended as a map, eventually the goal is to be a comprehensive rich Internet application with comparative metrics, speed records at the local level with executive summaries of every region and broadband histories for every region.

4645 THE CHAIRPERSON: Can you slow down a bit?  The translators are having problems.

4646 MR. LEONG: This is hard for me.

--- Laughter

4647 MR. LEONG: Okay -- broadband histories for every region covered both positive and negative.

4648 The foundation believes that broadband, first and foremost, should now be considered an economic investment and just as important as your local employment market, language, schools and other economic factors in choosing where you live and work, whether global or local.

4649 The three things that as an individual you cannot change about your location are the weather, the people and the quality of the Internet which you are stuck with depending on interplay of policy and behaviour from the incumbents.

4650 I would like to note that ILECs actually means incumbents because we were not clear as to the terminology, assuming that ILECs actually cover the cablecos as well.

4651 Much of what I will be presenting today is drawn from the data we have been collecting from the map so far.

4652 In addition to that, I also stand before the Commission as an Australian, coming from a country which has already seen a decade of the developments just beginning in Canada to now play out.

4653 I am currently living in Toronto, the original plan being to start what would have been a branch of a mobile/video game studio but, like many Canadian businesses, have been negatively impacted by the uncertainty caused by the usage-based billing issue.

4654 Now, instead of my original purpose -- the idea for the foundation was conceived after a gruelling two-month period in trying to obtain usable Internet after moving here which includes trying DSL with horrible results as well as a long and protracted battle trying to switch to TekSavvy cable from Rogers.

4655 In addition to the UBB issue, Toronto is faced with a unique situation of poor and scattered DSL speeds across the city plus an inherent conflict of interest in having incumbent techs service third party installations, who are also their competitors, which we recognize is a sub-issue and not key to the topic of this hearing.

4656 Now, to answer the first question: Which of the billing models proposed by intervenors should the Commission adopt?  The foundation firmly supports that no third party usage-based billing should be required whatsoever and by that we mean the 95th percentile method is more than adequate.

4657 Now, at this point the plan was to bring up the OECD rankings and why not, but the Commissioner has already indicated that he is not exactly a fan of them and, actually, we would partially concur because while the OECD measures a good measure of speed and pricing, they don't take into account negative factors and punitive factors such as caps throttling(ph) and just how long it takes to get service in the first place.

4658 These are all key metrics which which we will be analysing.

4659 Now, I would just like to add that it's been brought up by several other parties that Iceland and Australia are the only other countries that -- I would just like add New Zealand is in there as well and they are doing worse.

4660 Anyhow, for Canada third-party providers are the last line of defence in many years for Canadian businesses which rely on their pricing models to survive and compete at the same or a similar level to the rest of the world.  And by that I mean it's the only way Canadians can pay the same and get the same service similar to other parts of the world.

4661 Now, should this hearing decide that the UBB model be passed on at the wholesale level to third-party Internet providers, Canada will officially become the new laughing stock of the Internet world which it is now only tenuously avoiding today through the continued plans offered by the third parties.

4662 Our position is that the policy continues to support what the parties are doing as they do now, combined with potential and functional separation and/or community or nationalized ownership should be -- that should be the way going forward if wholesale problems continue to persist.

4663 In the meantime, third-party providers can help fill the competitive gap during the transition to more capable networks.

4664 Now, before you are a bunch of figures from sort of a random assortment of countries around the world -- OECD and some not OECD.  We have chosen the figure $40 to $50 Canadian a month because that seems to be the rough international average that you should be paying.

4665 We may have actually overshot that actually based on the results because several of the results are actually $30 Canadian a month and topping out at $40 a month.  And while I go through them briefly some -- and I should note that some of these ISPs are more important specifically to Canada's -- in answer to Canada's situation than others.

4666 But let's start with Australia.  TPG which is sort of like the Australian version of TekSavvy -- they have always higher -- they have always offered more higher tax than the competition and they have only actually just started offering the unlimited plans for the first time in the period since the whole debacle started between that time and basically now.  They have always catered to basically content creators, heavy users and, you know, people who generally like know how to use the Internet.

4667 Now, we look over at Estonia.  These figures are provided to me by a senior software engineer from Skype.  These were actually quite shocking to me: $30 Canadian for generally around 15 -- 12 to 15 megabits.  That's not even fibre.  That's just on cable.

4668 Now, where it says where fibre is available, okay, they can get 100 megabits at $38 Canadian, you know, or even so at 30 Canadian depending.

4669 Now, Estonia has only just recently been admitted to the OECD.  They have not been OECD in the past.

4670 Latvia, which is not OECD -- now I have got one here from Lattlecom and it's -- well, you know, 200 megabits for $39.52 Canadian.  Note that there was no usage variable charged caps of any sort.  They are simply selling it by speed and that's it.

4671 The Netherlands is an interesting example because they more or less already have fibered up the entire country.  They have completed the national strategy.

4672 The interesting thing about them is that I have actually got three providers here, you know, like, slashed because they are all offering the exact same service for the exact same price.  Again, not looking though -- we would have to deep into the analysis there.

4673 In South Korea, South Korea broadband, which is I guess sort of like the most -- one of the large market shares, almost like the national provider if you will -- they sell 100 megabits for $34 Canadian approximately and that is actually the most expensive non-contract price.  If you sign up for a two, three, four-year contract which makes it easier from the gigabit -- it goes out even further.

4674 Megapass which is one of their competitors, $36 Canadian, again most expensive non-contract price.

4675 And there are references to all of these ISPs listed after the fact.

4676 South Africa which is a country with terrible infrastructure, averaging 3Gb cap a month; in other words, they are where Australia was way back in '96 -- are selling 512 kilobits which is exactly where Australia was back in '96, for $40 Canadian a month unlimited which Australia was not able to offer way back in '96.  And they are profitable and they have been profitable for a year.

4677 Okay, of course we have to go to the U.S.; AT&T 18 megabits $40 a month at the moment, 24 megabits $50 a month.

4678 Now, they are implementing a U-verse cap.  However, they haven't used a service which is not going to be capped and actually there is some debate on whether or not they will actually continue this cap at this time.  My guess is their strategy to incentivize through a cap migration to the uncapped U-verse network.

4679 Verizon, 15 megabit $14.99 a month and, you know, after -- you know, that's where available -- sorry -- up to 15 megabits for $39.95 a month under existing infrastructure with no caps.

4680 Cox, 18 megabits, $14.99 a month.

4681 So we are staying well within sort of like the $40 to $50 price range at this time.  So the only way Canada stays within this $40 to $50 price range with the same sort of usage conditions and service offers is through third parties.

4682 - So TekSavvy, up to 15 megabits $54.95 Canadian for unlimited and, if you want to pay less, 300 gigabyte caps.

4683 - Acanac, unlimited except in Quebec, up to $35.95 Canadian.  You know, you pay a little bit more for 15 megabit.

4684 - And Distributel which is hovering $43 to $49.95 a month.

4685 You can't really get this from incumbents.  I mean if you pay at a similar price -- so the next thing I have listed is incumbent prices -- you are going to get hit with 60 to 50 gigabyte caps you know through Rogers, Bell, TELUS.

4686 Shaw, which at least seems to be upgrading the infrastructure in the right direction, you know they are a little bit weird.  They have an existing pricing scheme and they have a migrated pricing scheme.

4687 But in general like, you know, given the unlimited options available through their website, you are going to be paying $134.90 Canadian, you know anywhere between $74.90 to $134 so that's -- you know, so you are paying well above.  This may be a function of their cost recovery structure that they have to charge this.

4688 Now, prior to my speech, in listening to other parties, the Commission has brought up the need to introduce competition and how do we facilitate that and how do we choose to best wholesale the model to do that?

4689 What is even more interesting about the source that I pulled up for TPG, for being sort of the last largest ISP, was that I actually ignored the second-largest ISP because I assumed that the Telstra/Optus -- you know, duopoly cableco/telco duopoly -- was in a similar place.  I was wrong when I re-read the source.

4690 The second largest is actually iiNet which is an independent.  So now we have the second and third largest in Australia being independents and very profitable independents.

4691 Now, to sort of understand how they sort of buy their service, like how the wholesale sort of works, Telstra which is the -- I guess the incumbent government -- originally government-owned monopoly in Australia -- was forced by regulators to sell their wholesale, you know, to competitors.

4692 So both of these, TPG and iiNet, they sell ADSL and DSL too.  But of course iiNet which actually started in a suburban garage, had to achieve revenues somewhere.  They have actually -- more importantly they had to build up enough revenue to commence mergers and acquisitions and start acquiring across the country.

4693 So how did they get to that position relatively quickly and much faster than Canadian independents?

4694 Now, this I don't have as a reference but because the Commission has asked about wholesale in the past and it gives me a better understanding of the scope of this proceeding, I went and actually had a look.  To make it easier for you guys I have posted a Telstra wholesale link on my website that you can pull up.

4695 Now, if you go to the front page -- you know if you guys go to World Broadband Foundation at all and go to the front page under "Recent Content" I posted the references to provide it being the second largest and Telstra's 95th percentile billing that they state they offer through down the wholesale.  It's actually page 6 of their fax sheet.  So despite Australia starting the world up on Internet caps, they have potentially been building 95th percentile the whole time.

4696 Now, they also mention that they sell also a usage rate per megabit of data here.  The customer pays a unit charge for every megabyte downloaded and they make sure a tariff also applies to this option.

4697 And then following that just the 95th, you know, they mention we also sell 95th percentile and they say:

"Both the usage rate and the 95th percentile tariffs comprise a monthly committed volume charge and a variable charge for any usage above this."

4698 So it doesn't sound all that dissimilar to what many of the parties here are proposing but the devil may be in the details as in the amounts offered.  So you know, costing issues and whatnot, but I would encourage closer examination of that.

4699 So, okay, so I have already mentioned that both these independents are widely profitable.  In fact, the CEO of TPG just entered Australia's Rich 100 list.  It's good for him.

4700 But, lastly, I will pull out a quote from an article about South Korea's broadband which has been written about with much praise around the world, and so I'm leaning to repeat it:

"The Seoul government's clearly articulated vision for modernizing the country's infrastructure stands in stark contrast to the regulatory morass of stunted development in U.S. telecommunications after several decades.  South Korea's policy -- the cornerstone of a national technology initiative to help revise a devastated economy -- has created true broadband competition which, in turn, has helped prices fall and speeds rise."

4701 Okay.  Moving on to the second question: If a usage component is adopted, which should the Commission adopt, via peak traffic or via volume?

4702 We again reiterate that we do not support any usage component whatsoever.  There seems to be no need.

4703 Most importantly for the Commission and Canadians to understand, both are detrimental in their own way especially in terms of the documented Canadian productivity cap.

4704 When people spend more time worrying about rationing the Internet or pointing fingers at family members or even having mortgage payments affected by high Internet usage costs, there is a layer of mental stress affecting productivity that many competing countries are unaffected by.

4705 Now, I have -- just for your reference here, I do have a figure which I will call attention to, coming shortly.  That figure is from Optus, the now formerly number two being displaced down to number four and lose market share every day.

4706 "Peak" traffic management is the more onerous of the two.  If the model is implemented in a similar fashion as in Australia, it potentially causes subscribers and families to plan their living habits around Internet scheduling.  It will substantially reduce the overall quota available -- sorry -- sold to any subscriber.

4707 And the reason why -- because either way, the provider is setting aside an allotment which will then be divided into peak and off-peak periods and, in the Australian implementation, exceeding one metre will remove the quota on the other, resulting in even more excess profit towards providers especially if they have charged for a block prior to selling to the subscriber.

4708 In other words, they wouldn't even have -- they wouldn't even be close to actually having to provide the maximum of data allowance you know that they are selling it for and that the subscriber has paid for.

4709 Having to essentially schedule their lives around Internet usage hours was a factor in keeping many Australians still on dial-up for years even after broadband became available because of a greater flexibility and less in the way of surprises in terms of billing.  It also retarded what the local IT industry could target in terms of services.

4710 The economy is still catching up today, so I hear for now, and this decision would reduce Canada's competitive advantage in the same fashion.

4711 The "volume" approach also reduces Canada's competitive advantage.  But at least this gives the subscriber flexibility in when and how they use up the volume.  It also gives them some form of predictability of billing and accountability when the time comes, you know, as a unit of measure.

4712 However, if volume pricing is up for discussion, this really eliminates the argument put forward by providers who justify this form of billing as a way to manage or compensate for peak traffic.

4713 A subscriber could still use a large volume but spread throughout different times not affecting peak times, while someone who is just a normal peak user could be considered a heavy user during peak hours but may have a substantially lower volume overall.  Our volume assessment is particularly concurred with by a third party provider, Distributel.

4714 Now, the remainder of this I would have talked about, you know, how this would impact services like video and anything with really large datasets.  But that's really been brought up enough.  So let's move on.

4715 Are there significant differences between the billing models in terms of impacts on incentives to build up networks and manage traffic?

4716 So this is the implication that, you know, models really are over charging.  And punitive models, such as the ones proposed by Bell and Rogers would actually have an impact on incentives to build up networks.  We don't believe so.  Both models will collect additional fees and overages that are just distributed differently.

4717 The question is whether or not these fees will go towards building and upgrading infrastructure that will be committed to going towards building and upgrading infrastructure so that such pricing schemes are not necessary in the future.

4718 Incentives are not borne by different billing models, but by demand such as new applications, the ability of the network to provide, and the desire to be globally and economically competitive.

4719 The only argument that can be made for these billings models to incentivize build-up is that they create the desire to be free of them entirely.

4720 In Australia, the formerly public, now privatized incumbent monopoly, Telstra, saw record profits while the country saw gross underinvestment in the network, a lost decade in IT growth, punitive caps --

4721 THE SECRETARY: Excuse me, Mr. Leong, you have two minutes to conclude please.

--- Laughter

4722 MR. LEONG: Wow, okay.

4723 THE CHAIRPERSON: Don't speed up because the interpreters cannot catch up.

4724 MR. LEONG: All right, the rest of it is economic anyway.

4725 I would just like to add, you know, like a causality.  So it is a good argument to be made that, you know, this investment was made just because they wanted to improve the infrastructure.  And really, the prior decade of service had nothing to do with it.

4726 Again, on my website, under NBN causation, the labour party or actually the 2007 Rudd Government specifically mentioned nation-building in the 19th Century meant -- sorry -- okay, I have to pick and choose here.  Australia's behind, we are to remain globally competitive...

4727 This technology is a boon for Australia's long-term economic future, with Rudd stressing that current broadband services were retarding development of Australian business.  Australia ranks 31st in the world for average broadband speed.  Among the advanced economies, it is one of the most expensive for household and business services and one of the very few to impose monthly download limits.

4728 Okay, now can I conclude or can I at least answer the sixth question, or do I just conclude now?  I had two minutes, I would shorten --

4729 THE SECRETARY: You will have to conclude, sir.

4730 MR. LEONG: Thank you, all right.

4731 In conclusion, this Commission has a very important decision to make from this hearing that extends far beyond simple wholesale pricing and into the future economic health of Canada.

4732 The third parties are the very lifeblood of the content creators and other contributors to Canada's economy and entire companies may be forced to relocate.  Whatever decision this Commission makes and especially in the coming months of deliberation, my foundation will be on hand to map the impact and to present them easily and accessibly to other Canadians and the world.

4733 Thank you.

4734 THE CHAIRPERSON: Thank you for your presentation.

4735 Now tell me, if I understand it, you are against usage-based billing by which you mean imposing usage requirement on the wholesaler?

4736 MR. LEONG: Usage requirement and metering as a disincentive to use the network further.

4737 THE CHAIRPERSON: Okay.  On page 3, I guess, of your presentation you say:

"It is extremely important to note that of these examples there are several of interest. TPG, essentially the Australian Teksavvy, but they rose to become Australia's third largest ISP, almost able to challenge the existing duopoly of Optus and Telstra, despite similar usage-based billing plans, the underlying cost structure."

4738 Who are you talking about?  Who has usage-based billing in their underlying cost structure?  Optus or TELUS or are you talking about --

4739 MR. LEONG: Well, in general, usage-based billing is a standard, period, even with TPG, so they do offer plans with usage.  You know, there is a cap and, you know, you pay for more if you go over.  And so it has actually been -- yes, it is the Australian standard.  We invented the standard, unfortunately.

4740 THE CHAIRPERSON: So you are slightly contradicting yourself.  You are holding up Australia as a model, but on the other hand you want Canada to adopt a model that doesn't contain usage-based billing?

4741 MR. LEONG: Because, actually, they have risen to the top from higher caps and certain policies.  Like, okay, we may shape you, but we may not allow you and -- you know, like we won't cap you, but we will shape the traffic that you use.  You know, I guess for you guys that would be called IMTP.

4742 Now, the point is that they have risen because even though they started up billing almost identically to the way Canada does, now the underlying -- we are not exactly sure of the wholesale nature of the underlying cost structure but, you know, somehow they are able to offer more.  Somehow they are able to transition to the unlimited.

4743 Somehow they are able to offer more choice now while Canada seems to be, you know, losing choice.  We are having this hearing as a result.

4744 THE CHAIRPERSON: Okay, thank you.

4745 Tim, you have some questions?

4746 COMMISSIONER DENTON: I think Mr. Hénault follows, and then we will have questions.

4747 THE SECRETARY: Yes.  We will here --

4748 THE CHAIRPERSON: We are trying to do them one after the other.  Would you please follow the order?

4749 COMMISSIONER DENTON: Oh, okay, I thought that was the order.  Okay.

4750 THE CHAIRPERSON: Well, it is no longer.  Okay, we are dealing with Mr. Leong now, so you have --

4751 THE SECRETARY: No, sir -- Mr. Chairman --

4752 THE CHAIRPERSON: Would you please do what I tell you?

4753 THE SECRETARY: All right.

4754 THE CHAIRPERSON: Let's do them one after the other.  We have done Mr. Leong.

4755 Have you any questions for him, Tim?

4756 COMMISSIONER DENTON: Mr. Leong, I am just -- the no third-party usage-based billing whatsoever sounds good.  Yet, as we know, the peak billing method and the aggregate volume method, in some measure they all charge for some kind of volume, only differently.

4757 MR. LEONG: Yes.

4758 COMMISSIONER DENTON: And I also heard you say that you were in favour of the 95 per cent peak billing method.

4759 MR. LEONG: Yes.


4761 MR. LEONG: Yes, yes.

4762 COMMISSIONER DENTON: Okay.  So in other words, you might accept a form of usage-based billing as long as it came at 95 percentile --

4763 MR. LEONG: Yes.  As long as it was basically adequate to cover user-need as well to cover costs, which seems to be the case around the world.  And some of the parties here have also stated that.

4764 COMMISSIONER DENTON: Okay.  Because ultimately, usage must be paid for and there is argument going on here between two methods of charging for usage.

4765 MR. LEONG: I think, from the consumer point of view, it is the metering that happens after the fact, you know, the variable cost.

4766 COMMISSIONER DENTON: Okay, I think that is the only point I wanted to clear up.

4767 THE CHAIRPERSON: Does any one of my colleagues have a question?

4768 Okay, then let's go on with Mr. Hénault.


4769 MR. HÉNAULT: My name is Claude Hénault.  In my presentation, which I am about to read, I state initially, early on, why I am here and who I am.  So if everyone is agreed, I will just start in.

4770 Good afternoon, et merci d'être à l'écoute.  We are here to discuss usage-based billing, also known as usage caps.  Some call usage-based billing an Internet service pricing mechanism.  We already have one of those and it is called pay for connection speed and it is in use worldwide.

4771 UBB is double-billing, practiced by only a minority of Internet service providers around the globe.  Some call it a network decongestant, but it is clearly not made to do that job.  Congestion happens at precise points in time and is best dealt with by activating stand-by capacity.

4772 Usage-based billing is 24/7.  Some call it a charge that disadvantages a competitor, there is some evidence.  Its use in Canada has already caused video stream or Netflix to debase its offering to only two-thirds of the picture quality available in the United States.  Some call it a cash cow, I will make the case for that view.

4773 I am here to add my view about UBB and warn that its use is destructive to Canada's international Internet competitiveness.  It will, in the end, hurt our economy.

4774 My name is Claude Hénault, I retired nearly two years ago after 46 years of what we call knowledge work; split into journalism and the federal public service.  In journalism I worked in print radio and television with highlights perhaps being the many years I spent writing about politics and economics while based in the press galleries of the Quebec National Assembly and the Parliament of Canada.

4775 In public service I worked in the early days as an economic development policy analyst and ended with more than a decade running the Federal Department of Finance's websites.

4776 The website experience taught me a lot about Internet congestion, fuelled by a million visitor annual budgets surges.  The experience also taught me how technology, advancing year by year, wiped this congestion out and cut the handling costs.  We just had to make the right investments.

4777 In retirement, I am a student of economic events, a non-partisan political junkie, and a bit of a techie.  I was catering to these interests last February 3rd, though some would argue I was just being a bandwidth hog.  I was watching an Internet video stream.

4778 It was a live House of Commons stream of CRTC Chairman Konrad von Finckenstien's appearance before the Industry Committee.  Among other things, he defended usage-based billing decisions.  That defence brought me here because I can't agree with it and sought an opportunity to present a different view.  It is to the credit of the CRTC's consultation process that I am here.

4779 My presentation will build on what I learned on the job, info from my contacts in the tech field, and my experience as a long-time residential user of high-speed Internet access via cable.

4780 I use cable, Vidéotron to be precise, because in my home, 20 minutes by car and 45 by bicycle from Parliament Hill, DSL is not available.  So much for the duopoly in my case.

4781 Let's start by laying down a few basic facts.  The residential ISP is a local connection over what is called the last mile between Internet content on the backbone on one end and the client in the home on the other.

4782 The ISP is responsible for making it possible for patterns of digital information that exist prepaid on the Internet to be copied over to digital devices in the home.  No bandwidth is consumed, no content used up, no heavy lifting performed by the ISP no matter how long this process of copying goes on.

4783 The price of this service is therefore determined by the speed of data transfer made possible by the ISP, not the duration or volume of any transfer.  The speed delivered depends on the price the household contracts to pay.  The household decides how much speed it needs and it pays the ISP to meet its requirements for a month.

4784 With Vidéotron they contract to provide a household with a basic 3 mbps connection 24 hours a day for a month is $28.95.  Good for search and chat, but it won't stream quality video.  You get what you pay for.

4785 Moving up, Vidéotron says you can download video and music with the 8 megabits per second package at $43.95.  I tried that, but it proved inadequate for the four users of our home network.

4786 Further up the scale what the company calls speed-savvy Internet users can pay $82.95 for 60 megabits of connection speed per second, every second of the month.  That is the one I have $94.50 after taxes.

4787 That is the price I pay for an Internet service, and the last-mile local network capacity to support that service.  Same deal for telephone service or a cable, right?  Well, no.

4788 If I contract with the phone company for a landline local connection and service package at a fixed monthly rate, I can rag that phone for 24 hours a day local calls and pay the same as the hermit next door who keeps a phone exclusive for 9-1-1 emergencies.  The network is built to support the demands that will be made on it, both high and low, and charges are the same for different usages.

4789 If I contract for a cablevision package that meets my needs at a specified cost per month and run four television sets 24 hours a day, I pay the same as the people next door with the same package, but only one TV set on which they watch just one program per day.  The network is built to support variable client service demands and charges are all the same.

4790 But if I contract the same cable company to connect me to the Internet via the very same pipeline they use for TV, they won't permit my robust but reasonable use of the service I paid for unless I pay a usage surcharge after reaching some arbitrary volume of usage limit.  Why?  Why am I told I have to pay extra because the network, unlike with telephone or cable TV, is apparently inadequate?

4791 How can it be that the network allegedly is not built to support the service the ISP that owns the network has sold me and it is my fault?

4792 Upgrading the network to support the high-end service they are charging me an arm and a leg for would apparently be unfair to the people who contracted to use it less and, therefore, pay a fraction of my contract costs.  Does this make any sense?

4793 Apparently it does to the Chairman.  He told the Industry Committee:

"I would like to reiterate the Commission's view that usage-based billing is a legitimate principle for pricing Internet services.  We are convinced that Internet services are no different from other public utilities and the vast majority of Internet users should not be asked to subsidize a small minority of heavy users.  For us, it is a question of fundamental fairness."

4794 Does this statement make sense from a network designer's perspective?  Do you validate only the weight of cars if you are building a road that will have to support trucks as well?  Would you want to use that word?

4795 The Chairman's statement certainly succeeds in pitting the majority Internet demographic against the minority.  Good politics maybe, but is it good Internet network management? Not so much.

4796 The Chairman compares Internet services to other public utilities.  Does he mean they are like telephone and cable?  If so, why are they not treated the same?  Those networks certainly did not grow to their current robust shape by privileging the lowest common denominator of users.

4797 If the Chairman meant the Internet is no different than services like electricity, natural gas and water, as he applied elsewhere in his presentation, then his role demands he refocus on understanding the difference in kind, the difference in nature, of digital and physical products.

4798 Digital products can be infinitely copied, distributed cheaply, and used again and again at practically no cost.  Electricity, water, and natural gas have none of these qualities.  The comparison may be politically resonant, but it is hard to see a role for it in network management.

4799 And finally, the Chairman calls usage-based billing a legitimate way to price Internet services.  But these services are already priced according to speed.  Usage-based billing is not a pricing mechanism, it is a surcharge.  Is it a justified surcharge?

4800 Let's look to see what are the data that give any legitimacy to usage-based billing.  Is it an international standard?  No.  Around the world it is used by a minority of ISPs.

4801 The latest Internet study by the 34-nation Organization for Economic Cooperation and Development (OECD) -- and the study is available on their website -- shows that as of September of last year 80 per cent of cable Internet officers, 68 per cent of DSL offers and 73 per cent of fibre-to-the-home offers were entirely free of usage-based charges.

4802 Canada's nearly universal adoption of caps make it an international oddity, along with Australia, Iceland and New Zealand.

4803 Usage-based billing is charged by the gigabyte, but does anyone know the cost to the ISPs of delivering a gig?  Nobody but the ISPs, and they are reluctant to tell.  Despite their silence, however, industry and expert sources are near unanimous in saying their costs are a tiny fraction of what they charge.

4804 Netflix, admittedly with an eye to their own interests, placed the costs to ISPs of delivering a gigabyte of their video to homes at less than one cent.

4805 In a recent financial report Netflix said:

"Wired ISPs have large fixed costs of building and maintaining their last-mile network of residential, cable and fibre.  The ISPs' costs, however, to deliver a marginal gigabyte, which is about an hour of viewing from one of our regional interchange points over their last-mile wired network to the consumer is less than a penny, and falling.  So there is no reason that pay-per-gigabyte is economically necessary.  Moreover, at $1 per gigabyte over wired networks, it would be grossly overpriced."

4806 Vidéotron charges $1.50 for each gig in excess of my monthly cap of 120 gigs.  At the one cent valuation, that is a cost mark-up per gig of 14,900 per cent, on top of the profit from the charge for speed of connection.

4807 Canadian Internet expert Michael Geist once estimated the all-in cost of a gigabyte, including investment for network expansion, at eight cents.  At that cost, the UBB surcharge cost mark-up would be slightly less than 1,800 per cent.

4808 This info is compelling, but still interpretative.  So let me try a thought experiment on you.  Let's look at how much it costs your home as the ISP of the last yard to move a gigabyte of data around and around and around.

4809 The internal costs of the last-mile ISP are greater in magnitude, but identical in nature to those of the last-yard ISP.

4810 In the home you invest in a router connected to the ISP's modem to serve all the devices that need an Internet connection.  Some you connect by laying Ethernet cable, some wirelessly.  You pay for a variety of radio transmitter-receivers.  You pay for the ISP connection speed you need to meet the requirements of all your devices that connect to the Internet; desktops, laptops, tablets, TVs and a gaggle of music and video players.

4811 You upgrade by adding another router as a network extender if the signal attenuates too much on the deck outdoors.  You find your original main router is inadequate and replace it with a much more expensive synchronous multi-band device.  You do your homework, make your investments, and test for usability.  When you have done all this you have a network that fully meets your needs.

4812 Now, take a high-definition movie on one of your desktops, put it on continuous loop, and stream it through your main router down to your last-yard network to the digital TV, run it day and night if you want, gig after gig after gig, hundreds and hundreds of them, and it will have no effect on the operating cost of your last-yard network, other than pennies for a small amount of additional electricity.

4813 And because you have built your internal network to meet demand, all other users remain fully serviced, including via the Internet throughout the experiment.

4814 You have just demonstrated that there is no commercial justification for UBB within a properly built and maintained and serviced network.

4815 usage-based billing is just a mechanism that permits Canada's ISPs to resist change without suffering the economic penalties that such resistance would normally cause.  And the CRTC is, in effect, aiding and abetting that self-destructive course.

4816 The normal course of business growth for an ISP should be based on continual network improvement designed to encourage high-end users, attract as many newbies as possible into the system, and then progress these new entrants up the usage chain towards the higher-cost plans.

4817 Usage-based billing wrecks this model by making it possible, at least temporarily, for ISPs to use the free money it provides to continue making rising profit within a nearly status quo model.

4818 And we are hitting the UBB sugar teat pretty hard in Canada.  The same OECD study I mentioned earlier took a look at the time it took to hit bit caps at advertised speeds.  In Canada, they looked at 18 representative monthly plans.

4819 Only one of the monthly plans took a full day at advertized speed before usage-based billing kicked in.  Fourteen of the 18 plans took less than a half day before UBB was triggered.  The quickest plan to trigger UBB took less than a tenth of a day.

4820 Are we doing this because we charge less for the speed of connection in Canada than in other countries?  Just the opposite.  The OECD looked at three of the most common speed bands: 2.5 to 15 megabits per second; 15 to 30; and, faster than 45.

4821 Of the 34 countries surveyed, we ranked 28th more expensive, then 29th, and 28th again.  So not quite the most expensive, but close.

4822 Why is this happening?  Looks suspiciously like a lack of competition, doesn't it?  But it was to favour at least a semblance of competition that the CRTC endorsed in support of the counter-force of wholesale ISPs.  How is that working?

4823 The Chairman told the Industry Committee:

"It has happened that a lot of the very heavy users have become clients of those small ISPs who don't have a limit.  What we are saying is most of the caps that will be imposed for people are those who are really excessive users."

4824 So we are here to put caps on competition, giving into a system that already lacks that economic lubricant.  That way lies disaster.

4825 The vector of technology is predictable.  Today's light user is going to become the moderate users, while new light users will be seduced into entry by computing devices that are increasingly easy to use for an ever-widening range of gigabyte gulping purposes.

4826 Nielson's Law of Internet Bandwidth postulated in 1984 that a high-end user's connection speed grows by 50 per cent each year.  That projection has proved correct.

4827 The author of this law also noted that the vast masses of users are low-end who will lag 2-3 years behind the high-end users.

4828 And he further commented on why there is such an apparent gap between high-end and average total usage of the Internet.

"The user base is getting base is getting broader all the time as mainstream users get online.  These new users are more likely to be low-end users than high-end users (all the geeks have been online for years), so the average shifts ever lower."

4829 Most of the heavy users today are not the abusers they are depicted to be, but merely the vanguard, the harbingers of what everyone will be wanting to do in a couple-three years.  Are we setting out to discipline the future right out of Canada's Internet?  Maybe.

4830 The Chairman told the Committee that:

"We are going to try at test, namely, we are going to use economic measures in order to discipline Internet usage."

4831 In my view, if this policy succeeds, the Internet fails in Canada and a massive force of social and economic change is delivered into the hands of international competitors who don't make the same mistake.

4832 And how to avoid it?  Gut the 2009 network management policy.

4833 The Chairman summarized the policy as follows before the Industry Committee:

"If you read page 4 of the statement I gave this afternoon, you will see that the first measure taken to deal with congestion was to broaden the network.  If this measure alone is inadequate and we don't have the necessary means, we then have to use economic measures. User-based billing is exactly that.  It is a way to discipline usage.  Finally, if these two approaches do not work, we can turn to technical means."

4834 This is a policy that consists of one worthwhile initial premise and two escape clauses that void the worth of the first premise.

4835 Just as with the telephone and cable networks built in yesteryear, the priority has to be on selling Canadians the Internet services they need and ensuring the network is able to support the services they are sold.  Fix the network, no escape clauses. Especially not escape clauses that reward failure to upgrade networks by providing guaranteed, growing revenue streams unrelated to costs, as usage-based billing does in the short term.

4836 The Internet is an economic and cultural force that already plays a large part in determining the upward or downward mobility of the economies of nations.  Does the Internet future belong to those countries which best discipline the Internet or to those who best enable it?

4837 There is a lot riding on the answer we give to that question.

4838 Thank you.

4839 THE CHAIRPERSON: Thank you for your presentation.

4840 I presume you do not have any problem with the basic premise that the more you use something, if that use causes costs, the more you should pay?

4841 MR. HÉNAULT: If that use causes costs, and that's where I have the --

4842 THE CHAIRPERSON: I understand you.  I listened to you very carefully.  I asked you a specific question.  Do you have any problem with that principle?

4843 MR. HÉNAULT: If your usage causes costs, then you should pay for it.

4844 THE CHAIRPERSON: Fine.  Thank you.  That's exactly -- you say we should gut our 2009 network management policy.  Presumably you are referring our ITMP Policy.  That is precisely based upon that very idea.

4845 First of all, expand, make the network as big as possible, because, like you, I believe the future is in the Internet.  It's the backbone of the modern economy and we should not restrict it but encourage its expansion.

4846 Somebody has to pay for that, and you, whoever you are, whether you're an ISP, whether you're a cable company or a telco, should have a billing method that makes sure that those who cause expansion and therefore cause the cost get billed for it.  That's the sort of basic premise.

4847 Let me just finish what I'm saying.

4848 That was number one.

4849 We said secondly, if you cannot -- if obviously it doesn't work and the need is faster than you can meet, then you may have to use economic ITMPs.  Economic ITMPs are based on the fact that those who cause a cost pay for it.

4850 Now, you say usage-based billing doesn't do the trick.  Everybody else here at this hearing has agreed it doesn't and obviously we're looking for alternatives.

4851 Then thirdly, if you still do not generate -- sometimes you may be driven to technical means of in effect managing the networks, throttling or whatever it is, but if you do that, do it evenly and tell people what you're doing.

4852 That's our policy.  I don't think you basically differ with the policy.  Where you're basically differing is what are usages that cause costs and how the billing system makes sure that it recovers those costs.

4853 MR. HÉNAULT: That's fairly stated.  I think that your first case, your first provision is bang on.  I do think, however, that the other two undermine it by not saying you really have no excuse but to build the network up to meet not only today's demand but tomorrow's demand because it's coming fast.

4854 People are using different devices that are easier to use.  The demands put on the network by ordinary users are going to become the driving force.  They are going to be the cost generators because they are going to come and they are going to use devices like this, which allow them, with no fear and with great simplicity, to have access to whatever they want.

4855 If they're of Japanese origin and they want to listen to NHK, they can just click on this little thing and they're connected to it.

4856 It's a whole order of difference that is coming down on us, and so I'm saying basically the high-end user is already paying very, very high fees and I don't think the high-end users -- because they're a minority, they're a small minority -- are going to be the real drivers of the Internet costs presently and in the future.

4857 It's going to be a whole new granny talking on the videophone, telling bedtime stories to her children, whether they're in Vladivostok or Montreal or Vancouver or Ulan Bator, and that's going to change the whole nature in which families use the Internet in very normal ways, not in any sort of abnormal high-end tech user's way.

4858 THE CHAIRPERSON: You may be perfectly right but none of us know.  You probably are right and we both agree that if you're an ISP or a telco or a cable company you should meet the demands.  You should see it.  It should be in your self-interest to expand rather than trying to discipline use or trying to penalize heavy users.

4859 I just walked you through our ITMP, what was our -- that's how to.  Now, the devil always lies in the detail and the application, and we obviously didn't get it right in UBB, hence this hearing, trying to make sure we get the right hearing.  So I appreciate your presentation.

4860 Michel, I believe you have some questions.

4861 CONSEILLER MORIN : Oui.  Merci, Monsieur le Président.

4862 Évidemment, vous brossez un tableau assez noir de ce qui s'est fait au Canada, mais vous êtes sûrement au courant des statistiques qui veulent que les consommateurs canadiens, pour ce qui est des visites sur les sites Internet ou du streaming de vidéo, et caetera, sont parmi les champions du monde, que la vitesse offerte par les différentes entreprises, finalement, est une des plus élevées au monde.

4863 Alors, comment vous réconciliez les statistiques de l'OCDE au niveau des prix, et caetera, avec la consommation quand même du consommateur canadien, que vous représentez bien comme utilisateur ici dans votre présentation?

4864 M. HÉNAULT : Nous sommes un pays riche, et l'utilisateur canadien est capable de payer des sommes mêmes exorbitantes pour avoir ce dont il a besoin comme accès à l'Internet, et je trouve que c'est probablement ça l'explication.

4865 Il y a eu plusieurs politiques du gouvernement fédéral qui tendaient à favoriser l'adoption de l'Internet.  Il doit en être féliciter parce qu'on l'a dans toutes les écoles du pays.  Ça fait partie des us et coutumes du peuple canadien.

4866 Mais maintenant, j'ai regardé des statistiques tout récemment, et ça indiquait qu'à domicile -- puis on parle ici de l'utilisation à domicile -- que c'est à peu près 25 pour cent de la population canadienne qui utilise Internet par moyen de leur maison.  Ça, ce n'est pas énormément élevé.  Je pense que ça, ça va changer, et ça va changer rapidement dans les années à venir.

4867 Alors, je ne sais pas si j'ai répondu à votre question.

4868 CONSEILLER MORIN : Deuxième question.  Vous n'avez pas...  Évidemment, on peut peut-être deviner la préférence, mais je veux vous poser la question.  Je ne sais pas si vous avez suivi les audiences depuis le début.

4869 M. HÉNAULT : Oui.

4870 CONSEILLER MORIN : Vous savez que j'ai déposé un certain nombre de chiffres en ce qui concerne le modèle de MTS Allstream, qui vend une capacité plutôt qu'une utilisation.

4871 Est-ce que, en ce qui vous concerne, vous... quel serait le modèle que le CRTC, que le Conseil devrait favoriser, selon vous, pour qu'on soit innovateur et qu'il y ait de plus grandes opportunités pour les petits fournisseurs de service Internet d'offrir plus de concurrence, plus de modèles, plus d'offres, finalement, de services?

4872 M. HÉNAULT : Je pense que probablement...  Je dois avouer que je manque d'expertise, mais je vais offrir une opinion quand même.

4873 Je pense que la capacité est un meilleur gage de la performance d'un réseau que l'utilisation pendant une période de temps.  Alors, je pense que peak capacity, c'est peut-être plus logique comme façon de mesurer l'utilisation d'un réseau.

4874 CONSEILLER MORIN : Au niveau des principes simplement, est-ce que vous pensez ou si vous savez que les petits fournisseurs de service Internet -- ce n'est pas la première fois que je pose cette question -- mais c'est une création du CRTC, dans le fond, pour encourager plus de concurrence, et vous savez aussi que cette part de marché est restée plus ou moins la même depuis bon nombre d'années, autour de 6 ou 7 pour cent.

4875 Dans vos lectures, dans votre connaissance des réseaux ailleurs, est-ce que vous pensez que si le Conseil choisissait la capacité, autrement dit, la possibilité que les gros fournisseurs de service Internet offrent simplement une capacité aux petits fournisseurs de service Internet, est-ce que vous pensez que ça offrirait plus de concurrence dans le système, que ça nourrirait plus de concurrence dans le système, que ça leur laisserait plus d'opportunité d'offrir un plus large éventail d'offres?

4876 M. HÉNAULT : A ma connaissance, ça serait apparemment préférable à l'autre alternative, qui a été l'utilisation du réseau pendant une période de temps, parce que, à ce point-là, on leur enlevait à peu près leur... une bonne partie de leur capacité de concurrencer.

4877 CONSEILLER MORIN : Merci pour ces réponses.

4878 Ce sont mes questions, Monsieur le Président.

4879 THE CHAIRPERSON: Okay.  Thank you very much for your appearance.

4880 Let's go on now with the third presenter, Madame la Secrétaire.

4881 THE SECRETARY: Yes, Mr. Chairman.

4882 Mr. Jatinder Bhullar is now here.  Mr. Bhullar, we're glad you made it safely.  The Panel is now ready to hear your presentation.  You have 20 minutes.


4883 MR. BHULLAR: Thank you.

4884 Good morning, Honourable Chairman, Commissioners, Madam Secretary Lynda Roy, and ladies and gentlemen.

4885 I find it a privilege that you have accorded me to do this oral submission regarding "The Billing Practices for High-Speed Access Services."

4886 It is a very important proceeding for the livelihood of many, critical services enabler for shut-in and less mobile Canadians, and not the least as a basis of education, social networking, broadcast content from above and below the 49th parallel and just plain fun and entertainment for millions.

4887 Furthermore, this CRTC proceeding will play a critical role in supporting ingenuity that Canada is known for in the field of telecommunications.

4888 It was almost 32 years ago that I was a young researcher at Bell Northern Research working in my job assignment for specifying iNet-2000.  This later came to be known as the Internet.  This was early 1980s.

4889 My further assignment took me to services for Canada's North for double-hop satellite services, working on the Telidon project, kind of a mini-cloud messaging and interactive system in today's lingo.

4890 Bell Canada, Microtel-Pacific Research and Government Telecommunications Agency wrapped my career in Ottawa when with my young family I moved to the Greater Toronto Area.

4891 I spent eight years at Ontario Hydro and completed numerous projects specifically related to the modernization of telecommunications infrastructure at the three production nuclear power generation stations. These are the Bruce Nuclear Power Development in Kincardine-Port Elgin area, Darlington Nuclear Station just east of Pickering and the Pickering Nuclear Station.

4892 I participated in the mid-90s at the CRTC in the Information Highway Hearings, not to represent the company or any interest, just as a Canadian on seeing what was happening in the world.

4893 I also occasionally have done interventions, including the Extended LD plan that Bell wanted to implement at the early stages of LD competition.

4894 I joined Rogers as a Senior Sales Engineer in Business Solutions and worked on major projects like Air Canada, TD Bank, CANARIE, the PWGSC Major Voice Services and the CBC Next Generation Network based on the optical solutions provided by Rogers.

4895 I also worked as a Sales Engineer selling wholesale services to customers like Goldline, the number one calling card company in Canada; Clearnet, where I met Mr. George Cope before he became CEO of Bell, I was selling to him; Primus Canada, Cable and Wireless and Sprint USA, among many others.

4896 After nearly 11 and a half years of employment, I lost my services at Rogers.  This happened about a year and a half ago.

4897 Having stated my background, I would like to emphasize that my participation, like always, is founded on my passion for this subject matter.  I hope through this involvement to share both my inputs and my thoughts.  I sincerely believe and wish that my efforts contribute towards greater innovation and services that Canadians can enjoy and organizations can profit from.

4898 So the main part of the presentation, Mr. Chairman.

4899 It needs to be noted that we are not here because the duopoly -- I will use this word quite often; duopoly for me is a bunch of telcos as one column and cablecos as another column -- existing across Canada.  You know, they are not uniformly everywhere but there are different telcos and different cablecos and so on and so forth.

4900 We are here because Canadians don't have possible competitive choices in keeping with the ever-growing global Internet and communication-enabled services.

4901 Japan was, for example, introducing and planning 50-megabit services way back, and the optical network in Canada was even in dial-up modems.

4902 Why is the duopoly not bothered about the lack of competition while claiming the existence of fierce competition?

4903 I was listening to some of the audio broadcast yesterday how there is fierce competition between cable companies for Internet and the telcos.

4904 From what I can gather there seems to be a strategy in place.  The phrases uttered were -- this is from my work experience -- "all boats rise," or in this case two boats rise.  Basically it's a wink-wink-nudge-nudge kind of policy, you know.  Like nobody can prove it, but when you shake out all the numbers, they all seem to land up in the same area or same ballpark.  It is not efficient competition.  It is something else.  I cannot prove what it is.

4905 Although I understand this mentality -- sorry.

4906 And the second one that was very common was "I am not spending any money to enable a competitor."  I think this came through Shaw's presentation yesterday as well, where they said, you know, like my first interest is my direct customers, but if there's a case for another business, yes, okay.  And this is uniform across cablecos and telcos.  Like their interest is first and foremost their direct customers.

4907 This is not only a simple statement, you know, from very high-level people, this is actually a fact at the working level, and I was very amazed to see some of the participations as I followed the proceedings.

4908 I know some companies that there are more VPs, directors and others that are attending this hearing than would be in a total wholesale group in that company.  So why is it that important?  There is definitely something else behind it and I think it is our job to find what that is and make sure we don't overlook it.

4909 As was stated by Mr. Shaw on Wednesday, they would much rather have a direct customer, though he flagged that his subject matter experts in wholesale are now chasing TPIA customers too.

4910 Hardly much can be said about cablecos in Toronto, which seems to be more in a shutdown mode.  This is my personal experience.

4911 It is important to visualize the Canadians behind choices and how they come to choose to go with smaller ISPs.

4912 Say a customer leaves and goes to Teksavvy.  Say it's me.  This does not happen because Teksavvy has the duopoly's marketing budget or their name was everywhere plastered on TDC or Metro in Montreal.  This is because they are not happy with the duopoly solutions or they suddenly discovered there was a choice, they suddenly discovered value.

4913 Is the duopoly really expecting such a customer to come back and be a customer again?  Highly unlikely.  Highly unlikely.

4914 So are such customers, Mr. Chairman, a burden or gravy for the duopoly when they indirectly through such ISPs become the underlying wholesaler?

4915 So my service was originally Rogers.  Through Teksavvy, they bought it from Rogers and sold it to me.  So Rogers hasn't lost their investment in that particular whatever infrastructure was there.

4916 It is very, very important to know that it's very rare that telcos or cablecos take hardware from one place and deploy it somewhere else.  They leave it in place for future use or maybe somebody will change their mind and in this case it gets reused.  So the actual costs on how they show up for customers like me are quite different, which needs to be noted.  So for me, I will say it's gravy for such situations.

4917 Now that it is proven that Canadians seek choice, there is no question that wherever Teksavvy and the others with better quality delivery and turnover -- turn-ups are rolling out that the demand is exceeding their capabilities most of the time.  So why is it so?

4918 I would like to focus on what the CRTC can do in this environment.

4919 The remainder of my presentation focuses on and is based on the framework per the communication where the questions are posed.  So I have contributed along the proceedings my thoughts, my data and everything.  What I have done is I have tried to distil that into thoughts to address the six questions that were posed by the communication for a "Structure and approach of submissions."

4920 So the first question that is posed is: Which of the billing models proposed by interveners should the Commission adopt?

4921 I support, like the previous speaker, the non-usage-based rates tied to a suitable mark-up on the costs.  These could then be related indirectly to retail rates from duopoly participants.

4922 What is important is like you've got to have a handle on costs, whether these are incremental capital costs, whether these groups of services are (indiscernible)-driven or whether these are stranded services, like very low speed services, they're just stranded.  So, you know, if they can get 10 cents on that dollar that they spent or whatever is left, it is still gravy.

4923 And the pricing has to be different.  It is not complex, it is not mind-boggling, but it is possible to do.

4924 So I lay out some of the points that should be included in such a model.

4925 The point of access should be as neutral as possible; no type of economic traffic content or other parameters interference; subject to only traffic engineering associated with access rates.

4926 Engineers have known since early days how to engineer traffic and it's not rocket science.  They have done huge networks.  They monitor that if it exceeds certain capacity then we need to augment.  So that is a normal process.  There is nothing strange about Internet, it happens -- in every network you manage capacity.

4927 There should be flexibility to engineer solutions to differentiate in wholesale and beyond speed matching.

4928 Now, this is like a wonderful thing that the CRTC has done by initiating its original ISP encouragement, ISP facilitation to open this field, otherwise we wouldn't be here today, but I believe that in the future, like telcos, Bell will sell their services at 12, 16, 18; Rogers will sell at 10, 15, 25.  Maybe a competitor can ask Bell, "Can you give me this?"  Other than simple provisioning changes there should be ability so that they can really differentiate also.

4929 There should be a minimum discount of 35 percent off the effective rate and actual access charges collected per user for a given access speed versus rack rate or retail rate.

4930 There is no point linking discounts to retail if no one, or very few, pay retail.

4931 This is one of the presentations or contributions that came from MTS Allstream where they said that, you know, if you apply a 35 percent discount, and when they look at the bundled pricing coming out of the competitors, that what you end up with is not a heck of a lot left for the ISPs to play with.

4932 So it is very important to focus on actuals rather than just rack rates which can be played with through marketing campaigns by the telcos and the cablecos.

4933 Additionally, a cost-effective unencumbered bulk billing process with the built-in ability to allow for future innovations in billing.

4934 Like there was a big hue and cry that, you know, we need six months when CRTC wanted an interim assessment.  I worked in wholesale, I know how to bill wholesale customers, it's much different.  It's billed at a very high level and you are playing with basically Excel sheets at most times,  Because they are not directly in the billing.

4935 If the cablecos or the telcos were directly billing the customer on an ISP's behalf, it is complex.  But since the ISPs bill directly the customer, there's nothing.  It's a lot of hot air.  Like it's a standard IT, you know, six months, just because you asked me a question kind of approach.

4936 Access parameters should not be arbitrarily changed by wholesale providers while introducing new services outside of ISP pipe or access.  For example, IPTV on same loop, cableco on-demand services shifting to ISP Internet.

4937 Like it will be very funny that suddenly first you sell an ISP, say Bell sells ISP a pipe and a customer, then tells the customer, saying, you know, "Get IPTV on that pipe."  That would be wonderful because UBB will kick in in no time flat and the ISP will suffer and Bell will say "Wure, that's a good business case."

4938 So we have to watch how these things roll out.

4939 Pricing methodology should pass the test of time and suffice for many years to come.

4940 So this isn't a six-month thing.  This shouldn't be a one-year thing, it should have some kind of a time horizon of two years to five years short-term, that you don't have to come back to the CRTC saying, you know, "This doesn't work, we want this changed, we want that changef", that it should have some strength of design that sustains the changes in the environment, changes in technology, changes in services introduction.

4941 The CRTC should fast-track the stimulation of new services by ISP as incubators or innovation derived for example out of Netflix, Skype, YouTube, et cetera.

4942 I worked for Bell for a while so I'm not picking on or trying to badmouth anybody, these are large organizations, you know, when they have to do market trials, when they have to try new technology there is a process that would fill this table with practices, but if you are an ISP you can sit down with the owner next to you, and you are the chief engineer and here is the marketing guy, you can figure something out.  You know, "The customer will like this, we want to try it.  What is our risk?  Can we do it?  Let's try it."  Much, much simpler.

4943 So I think there is a great value that CRTC could, rather than give a lot of marketing trial initiatives to telcos to actually use and get these people engaged.  This is the future.  Like they are younger companies, they are new blood, so that will bring a lot of good stuff.  I'm not new blood, by the way, so that's not the point.  But there is a lot of possibility for innovation in that area.

4944 One thing I did was, I took the 35 percent thing and I played with the Allstream model, MTS Allstream model, and I realized after that if I were to give an effective 35 percent discount versus the bundled or what the customers would really pay, it actually translated into what Primus was saying, which was 50 percent of retail.

4945 So in my submissions earlier on I was a little aggressive and said, you know, no way Primus -- that shouldn't even be considered.  So I'm not all anti-duopoly so I was actually asking for less than should be actually offered as an incentive -- or suggesting less for trial purposes.

4946 So if you look at page 6 of the handout, it is a very telling picture.  Like this picture kind of speaks to many aspects.  Like it's not written in my notes, but I can take you through it.

4947 If you look at the top column there are two examples.  Rogers.  I went to their website, I took down all their speeds.  Because I didn't have a similar table from MTS Allstream in terms of bundling and stuff because they did it for telcos, so I created a similar table for Rogers.

4948 So they have, sort of said starting from 0.5 megabits per second -- I wonder who takes it, probably some very legacy and traditional supporters of that -- and 3 megabit services, 10 megabit.

4949 Then you have the second column, I kind of labeled it aggressively, I call it "Innovation, Usage Killer Limits", because that's what they are.  Like when you have somebody with a 2 gig limit and a 500K service, what are you telling them?  It is saying, you know, as the previous speaker said, "Hit the limit, don't use the Internet, you know, we are going to nail you", right.  But that is not the purpose.

4950 They have already used technology for so many years to deliver it, there are probably no costs left other than some basic Opex, so we have to watch out.

4951 This is where also on the extreme right I classified these as "Salvage", "Opex", "Capex", another point to note on how these --

4952 THE SECRETARY: Excuse me, Mr. Bhullar, I'm going to have to ask you to conclude.

4953 You have one minute left.

4954 MR. BHULLAR: I'm sorry.

4955 Mr. Chairman, so I had noted down under the traffic things that I believe that there should be no user support.

4956 In terms of peak or average you ask, you know, if one was selected, which would be better.  Peak is always better because that's how networks are engineered.  But the peak is shifting in the whole network that it is becoming more uniform.

4957 Incentive to build.  I think we have to ensure that the wholesalers are in business.

4958 The same, manage traffic, it is standard traffic management.

4959 Can Bell, TELUS and the others deliver?  Of course they can.

4960 For symmetry of mark-up, absolutely not, because what you will insure is that the ISPs look like the telcos or the cablecos, because that is what's going to happen.

4961 The networks are not inherently efficient that they have reached a stage of perfect pricing and costing and competition.

4962 I apologize, I perhaps took a little extra time on my preamble.

4963 THE CHAIRPERSON: Thank you.  Thank you for making the effort to come here.  I understand you just made it barely because of traffic.

4964 Now, you quite rightly mentioned that it's part of our policy to ensure the ISP is there to offer competition to duopoly and of course neither the cable companies nor telcos have any interest in being any more forthcoming than they have to be because they are creating competitors to themselves.  I understand that.

4965 But you are now putting forward this very interesting idea, saying there should be no symmetry.  You are saying in 34 that by doing that then you would actually have them using each other's networks, you know.

4966 Maybe that would be one of the results, but the other one that we have been very conscious of, that most of the ISPs use Bell, that Bell has a vastly disproportionate share of the ISP traffic as opposed to other telcos and also very much as opposed to other telcos and also very much as opposed to the cable companies.  We tried to make the TPIA a more user-friendly, more easy at reducing the point of interconnection and various other things in our last decision.

4967 Do you see us also, by not adopting a symmetry, somehow being able to shift the burden of applying ISPs more evenly between the cable companies and Bell than it is right now?

4968 MR. BHULLAR: Mr. Chairman, that is obviously a very good question.  You guys worked through this for a long time.

4969 There is no question that from my knowledge as engineer and everything, in a part of my presentation at one point I state that Rogers today will deliver to my house 3 terabytes of programming under their high definition TV service for $29.95; 3 terabytes.  Now, they could run that same network much more efficiently, so the question of 2 gigabyte limits, 100 gigabytes or 200 gigabytes is very artificial.

4970 So if symmetry is taken away and real costs are used for providing discounts and pricing, absolutely it will be a marvelous opportunity to get the two towers to lean against each other a bit and start shaking.  Because at the moment, you know, they are just playing, you know, one gives 10, the other makes 12 so that you can't compare them like for like.


4972 MR. BHULLAR: Thank you.

4973 THE CHAIRPERSON: Marc, I believe you have some questions.

4974 COMMISSIONER PATRONE: Thank you, Mr. Chairman.

4975 Thank you, Mr. Bhullar.  You have tabled for us a very detailed and well organized document, of which I think you got through less than half in your oral --

4976 MR. BHULLAR: I apologize.

4977 COMMISSIONER PATRONE: -- but we have it.  We have it and you have given us plenty to consider going forward.

4978 I just have three questions, some of which were already dealt with with the Chairman.

4979 In paragraph 15 of your oral presentation you restate that you oppose any form of usage-based billing as you see it as impeding innovation.

4980 You see the lesser of those evils that have been presented -- I use that term guardedly, but -- as being the peak traffic model, the CNOC model --

4981 MR. BHULLAR: Yes.

4982 COMMISSIONER PATRONE: -- which, as you know, has a usage component.

4983 Could you speak a little bit more about why you think that model should be the one that we -- assuming that we are going to proceed with any of them -- is your choice?

4984 MR. BHULLAR: Thank you, Mr. Patrone.

4985 The question is all, as I said before, no networks are engineered or costed based on average traffic.  The real costs per driven by peak traffic, like a percentage of.  You may say, "Okay, 95 percent of the time I'm going to ensure that if so many users put on so much load and so many people are active that this will be the traffic, so I'm going to engineer it that way", right.

4986 So that is -- then you will translate that into modules, hardware, like pipes, interfaces, you know, gig interfaces, e100 interfaces and optical interfaces.  So everything is driven by peak.  It was the same for voice, it was the same for data, it will perhaps continue for a long time.

4987 What I did want to say is that I have suggested under paragraph 17 an alternate view of this.


4989 MR. BHULLAR: I think too much is being made in terms of peak versus average or even why, you know, it seems like peak is less of a devil and it's more palatable, but what I'm suggesting is that let it be in a trial phase.  Like let there be no UBB.  Let's watch how much deviation for the ISP users exist for a given interface speed versus the other, like global community for that interface, from that provider.

4990 COMMISSIONER PATRONE: You worked for Rogers, and part of the testimony that we heard back from the cable carriers was that in fact some of their costs are not driven by, as you said, peaks, but during volume increases over the course of the day.

4991 MR. BHULLAR: Yes.

4992 COMMISSIONER PATRONE: They provided one example where, you know, a business may generate a higher rate of volume from 9:00 in the morning until 5:00 in the afternoon and under the CNOC model there was an inability on the part of the cablecos to monetize or to pass on costs associated with augmentation during that stretch of time.

4993 Do you have any thoughts on that?

4994 MR. BHULLAR: As far as I know like Rogers is the last company to sell to businesses, so they are not a big layer in the business solutions or business delivery.  So I do not know where that came from.


4996 MR. BHULLAR: In terms of like what I see, I tried to search for traffic movement, traffic profiles which you have access to, perhaps from the telcos and cablecos.

4997 What is happening is, like with the new services, say if it's Netflix, you are going to not create a 9:00 to 11:00 peak that the kids play games or they download or they do peer-to-peer, you are going to start shifting that the 5:00 to 7:00.  You are going to start shifting that to the TV watching hours, right.  So during that day it's the business traffic.

4998 That by itself actually creates a very, very efficient use of infrastructure.  So it is actually creating -- naturally, like more services that network carries it distributes those services across 24 hours and it actually makes it more efficient, not the other way around.

4999 COMMISSIONER PATRONE: You spoke a little bit about economics and economic cases in your presentation and you spoke about artificial caps are what you see as artificial impediments to usage.

5000 As you know, telecommunications networks don't build themselves, it cost billions of dollars --

5001 MR. BHULLAR: Sure.

5002 COMMISSIONER PATRONE: -- to build them and then to maintain them and I'm just wondering where the concepts of risk, incentive to roll out future networks and getting a decent return on investment, where that fits into your arguments regarding billing methods and your view that we shouldn't be going down the route of usage-based billing.

5003 MR. BHULLAR: I have seen, in making proposals, you know, somebody says, "You know, I want to pay upfront on a monthly basis.  I don't want to be nickeled and dimed on this and that". right.

5004 So one can create a model for simplicity, plus a differentiation in the marketplace.  Considering the right costs.  Like there should be no plan to make the telcos lose on selling wholesale.  That's a given, right, otherwise the marketplace will die, right?


5006 MR. BHULLAR: So that is a given.  But the question is how you try to create the ISPs not in the same mold as the telcos, because what you will have is instead of a telco and cableco, you will have to ISPs who will look -- one, if he sells more cableco, will look like a small cableco Internet and the other one will look like a small telco Internet.

5007 Well, that shouldn't be the plan.  That would not create innovation, that will create the same constraints, it will create the same kind of environment that exists today.


5009 MR. BHULLAR: So the object is to change that environment, create a disturbance so that innovation comes in, choices come and people fight for the customer a little bit.

5010 COMMISSIONER PATRONE: Thank you very much for your presentation today, Mr. Bhullar.

5011 Those are my questions, Mr. Chair.

5012 MR.  BHULLAR: Thank you.

5013 THE CHAIRPERSON: Okay.  Thank you.

5014 I think those are all our questions for you.

5015 Let's take a five-minute break before we deal with the last intervenor.

5016 Thank you.

5017 MR. BHULLAR: Thank you.

--- Recessed at 1208

--- Resumed at 1217

5018 LE PRÉSIDENT : O.K.  Commençons.

5019 LA SECRÉTAIRE : D'accord.

5020 Mr. Chairman, we will now proceed with the presentation by Vaxination Informatique.

5021 Appearing for Vaxination is Mr. Jean-François Mezei.

5022 Mr. Mezei, you have 20 minutes, go ahead.


5023 M. MEZEI : Merci, Monsieur le Chairman.  Mon nom est Jean-François Mezei.  Je suis un travailleur autonome sous le nom de Vaxination Informatique.

5024 Je vais faire ma présentation en anglais, mais je peux répondre aux questions dans les deux langues.

5025 I want to thank the Commission for giving me the final word on this issue.

--- Laughter

5026 THE CHAIRPERSON: The definitive word.

5027 MR. MEZEI: Definitive, yes.

5028 I will start with just a general philosophy.  Really, it must be cost-based, what we are looking at, because whenever you don't have cost-based, you allow the incumbents to introduce their own retail philosophies, in subtle ways sometimes, and sometimes in very overt ways.

5029 In terms of the delay of implementing, I think it is very important to get it right, even if it takes a little longer.  If you get it right, you get it once, and it will last a long time.

5030 If you get it wrong, you are going to see me here again next year.  So that should be an incentive to get it right.

--- Laughter

5031 MR. MEZEI: I think one of the issues that we have looked at in the last years was the argument that the current tariff structure doesn't adapt to changing patterns.  I think that what we have looked at so far, either the AVP or the 95th percentile, or even the MTS one, they do adapt to this, and once this happens, I think, it will unleash a lot more potential for the services.

5032 During the last six months there has been a lot of talk in the media, and one of the aspects that has irritated me a lot were some of the incumbents going to the media and saying that ISPs are getting a free ride.

5033 I know about Phase II.  I don't know completely, but I know they are not getting a free ride.

5034 And I think it should be very important that when the Commission issues its decision, it makes it very clear that ISPs become partners who pay in full for the services they use, so that we don't have this issue of incumbents saying, "They are just getting a free ride."  That is not productive at all.

5035 Here I will be a little more controversial.

5036 ITMPs were introduced because of the original tariffs being fixed.  Once you get into tariffs that cover the variable costs, and then cover changes in user patterns, once the ISP pays for what he uses, ITMPs should be reviewed in terms of the wholesale sector.  Retail is okay.

5037 In terms of legacy versus FTTN, in terms of the Bell rates, I think they should be unified, because all we have now is the 5 megabytes that is left of use, plus a few slower speeds.  It would be a lot simpler for everyone, for ISPs and for Bell, if they were all the same.

5038 And reduce the base cost for the 5 megabytes from $20 down to, I think, $12.  I think they factored in about $8 of usage.

5039 And then put the AVP -- if we go AVP, just put it in there.

5040 It is just simpler for everyone.  If you try to explain to a retail customer that if you take 5 megabytes it's completely different than if you take 6 megabytes, at one point it becomes a little too complex.

5041 In 632, you guys decided to give a 10 percent premium on FTTN because of higher risk.  With a new rate structure that follows usage, and covers the cost for sure, then I feel that the risk is no longer there and the 10 percent premium may not be warranted any more.

5042 It was warranted last year, with the current tariffs, but with the new ones, it shouldn't be.

5043 This is a bit outside the scope, but I still have to say it.

5044 Because business GAS and HSA are exactly the same and travel on the same pipes, they really should be using the same pricing paradigm or philosophy.

5045 I know it's a separate process, but we should be having the freedom to choose without being interfered with by the others.  This is perhaps more important, more volume, for residential, but I think this is an issue and you don't want to have conflicts between the two.

5046 In the interim rates decision it was mentioned that the Commission wanted to hear a few comments about retroactive.  Nobody else has done it, but I just had to do it.

5047 First of all, AHSSPI is needed because, when you increase the speeds -- if you have a customer going from 5 to 25 megabytes -- it increases the load on the network, and the ISPs will have to order additional AHSSPIs.

5048 It was announced that the new speeds would be available this week.  It takes, from what I have been told, in practice, about four months for an ISP to get an additional link.

5049 Maybe Bell could comment on that in rebuttal, whether they could shorten that, because if you start taking new customers now, and your AHSSPI links become congested, and it takes four months to get it, then during those four months all of your customers suffer.  So this is an issue to look at.

5050 Retroactive billing -- if you are looking at, for instance, a touchtone fee going from $15 to $16, and you go back three months, it is sort of acceptable because the reseller, or the ISP, will know, basically.  Here we are looking at a completely different billing philosophy, and we don't even know which ones you guys will choose.

5051 It is very hard for an ISP.  The ISP cannot re-bill retail customers, at least not in Quebec, because I have already paid for my service for this month.  You can't go back six months from now and say, "By the way, you owe us another $5 for six months ago."

5052 So the ISPs will have to absorb that, and they can't plan -- they don't know the amounts and they don't know what the philosophy is.  So that is an aspect that needs to be done.

5053 Now I am getting to AVP.  AVP uses what Bell built for its UBB, the TN 7181.  Basically, they calculate it per user.  At the base level they pick up the per-user stuff, and it's only at the billing where they put it all together into one number, as opposed to 5,000 numbers.

5054 Errors for users get aggregated into one number, which the ISP cannot verify and cannot debate with Bell, because you get 5 terabytes and you can't say, "Well, no, that user didn't use that much," there is no way to verify it.  So that is, I think, a big disadvantage.

5055 More importantly, Bell Canada, on many occasions, and even a few days ago, admitted that its AVP is not cost-based.  And I go back to my first page, it has to be cost-based.

5056 You have heard many numbers.  I think that Michael Geist did a study saying that it's about 8 cents per gig.  Bell says it's 19 cents, but Bell says it's an ITMP, not cost-based.

5057 I think that Bell should come out next week -- I challenge them to come out and say how much it costs, because, at the end of the day, we need cost-based for that.

5058 The other thing, too, is that, by giving Bell non-cost-based AVP, it would allow Bell to set a price below which ISPs will not be able to go, and, essentially, that limits competition.  Essentially, Bell could price it high enough so that ISPs would be close to Bell's own retail price, reducing competition.

5059 Again, this is why we need to -- I am going to repeat cost-based, because I think that is really, really important.

5060 It has been said that AVP, because it's a flat rate all day, there is no difference if you consume your gig at night, overnight, or during the day.  There is no incentive to the ISP, because the ISP gets billed the same amount.  So there is no incentive at all for the ISP to shift usage to off-peak hours and reduce the peak.

5061 As was said in other proceedings, having dual -- peak and off-peak -- rates is very difficult for ISPs.  The technical issues make it very difficult, if not impossible.

5062 I will get back to this.

5063 MTS Allstream, for instance, mentioned having a fixed price for the DSLAM, no matter what speed you give.  In the case of Bell, with AVP, that is not possible.  If you give someone a 25-megabyte profile, the load on the network is higher, so each gig that you download at 25 megabytes costs more.

5064 So there is a reason for charging more for the 25 megabytes, when the AVP is the same, no matter what speed you have.

5065 That is why Bell can argue that, with AVP, having a fixed fee for the DSLAM is not workable.

5066 But you will see that, in a couple of other pages, I do have an answer for that.

5067 THE CHAIRPERSON: Could you repeat what you just said?  I don't think that any of us understood it.

5068 MR. MEZEI: All right.  At the DSLAM level, Bell's cost, whether they give you 25 megabytes or 5 megabytes, there is no difference, it is just a profile change, someone keying a number in a terminal.

5069 On the network level, when you download 1 gigabyte, at 25 megabytes it uses more bandwidth on the network itself, so there is a higher cost.

5070 Now, with AVP, it does not factor the speed at which you download.  You download a gigabyte and you get charged 19.5 cents.

5071 So the AVP rates do not reflect the load on the network depending on what speed you have, which is why Bell wants to charge more for higher speeds at the DSLAM level.  They have to compensate for that, which is understandable.

5072 But if you go into capacity-based, as you will see later, that solves the whole problem.

5073 Another aspect is, when you have congestion in a network, you may not be able to download at 10 megabytes, you will download at 7 megabytes or 5 megabytes, depending on the amount of congestion, but once you start to download your movie, you are going to finish it.  It is going to take half an hour instead of 20 minutes, or whatever.  So your actual consumption, because of congestion, may not be reduced at all.

5074 In terms of reducing the -- I was going to use the word "throttling", but that's not it -- in terms of reducing the person's desire for more data, it may not actually work that much, unless you charge, like the cable companies, an extremely high price.  That really puts the fear into people of using it.

5075 Again I mention that even Bell admits that its AVP pricing is designed as ITMP and not cost-based.

5076 Another aspect is, the argument was made that higher speeds will lead to higher consumption.  It is not proportional, and we will eventually come in a few years to a plateau, where people -- you only have so many hours at night to watch entertainment, whether it's on TV or Internet.  Eventually it will plateau, and the growth rate should stabilize.

5077 So a lot of assumptions about proportional growth of speed and consumption may not actually be true, although, statistically, right now they are.

5078 I am coming back now to capacity-based solutions.  We have heard peak values -- the word "peak" used a lot.  I use "capacity-based", because there are multiple methods of doing this.

5079 You have heard the CNOC one, which is the 95th percentile, and even within the 95th percentile there are variations to it.

5080 Capacity-based billing is the method which gives the ISPs the most freedom to shape their retail offering.  The AVP one, because it is charged per gigabyte, provides a strong incentive for ISPs to also charge per gigabyte, because they have to cover their cost.

5081 They can make assumptions, but eventually, if the gigabyte costs are high, they have to charge per gigabyte.

5082 Capacity-based gives total freedom as to how they charge.  They can charge UBB, they can charge unlimited, they can charge by speed, they can charge by whatever they want.

5083 All of the capacity-based solutions, whether MTS Allstream or CNOC, or the others that I will talk about, have the attribute of giving the ISPs the most freedom to shape their offering.  So, in terms of competition, this has the best potential.

5084 It is also the solution which, first of all, matches the ISP's own capacity planning, and also matches -- the ISP's buying transit on the Internet is bought by capacity-based.

5085 So if you buy both input and output at capacity-based, it's a lot easier to make sure that they are all the same and the pipe is all the same.

5086 The ISPs are really used to capacity-based.  It is the best solution, really.

5087 I guess that answers your Question No. 1.

5088 When you buy capacity-based -- right now GAS service has a big capacity-based component.  It is called AHSSPI.  When you buy this, you are limited by the physical speed of the link.

5089 Now, you can also limit within -- if you buy a 1 gigabyte link, you install the modems at 1 gigabyte, but you can ask to pay for only 200 megabytes, and then they will rate limit at the software level your rate, and as you grow, you can just call the incumbent and say, "I want to raise this to 400, or 500, or 600 and 1 gigabyte."

5090 For small ISPs this is important, and I will give you an example.  My ISP electronic box is a small one.  They have a 1 gigabyte link with Bell, and right now they are using only about 400 megabytes.  So they are actually subsidizing, because they are paying for 1 gigabyte, but only using 400 megabytes.

5091 In terms of the 95th percentile, there has been a lot of talk about the five-minute stuff.  There are two aspects of the five minutes, which can be a little confusing.

5092 Routers themselves will analyze your throughput for a five-minute interval.  So when you go into the management of the router and you say, "Give me this interface," it will show you that in the last five minutes the guy has transferred so much, in terms of 5 megabytes or 10 megabytes or gigabytes or whatever.

5093 The router constantly keeps track, so when you poll the router every five minutes to find out what the rate was, it will give you what the average rate was for the last five minutes.

5094 This isn't just a momentary peek into the speed being used.  If you poll every five minutes, and the router keeps track of the last five minutes, you end up having the total speed, or average speed, for the whole month.  Every second of the month is being polled, including those statistics.

5095 What you do is, you have -- I think I have the number -- over 8,000 samples in a month, and you discard about 400 of the highest ones.  So the next one, basically, gives you the highest speed that you have used.

5096 Now, that speed -- let's say that you have a 1 gigabyte physical link, and the highest 95th percentile sample is 700 megabytes.  Then you get billed for 700 megabytes; not for that five minutes, but for the whole month.

5097 So, essentially, what the telephone company is telling you is: While your peak was 700 megabytes, we will assume that you used 700 megabytes for the whole month, and we will bill you as such.

5098 The 95th percentile allows, for instance, my ISP, which only consumes about 400 megabytes, to pay for 400 megabytes, even though the link is 1 gigabyte.

5099 When he grows, it reduces the need to call Bell and add lines, and it also reduces the need -- it's automatic.

5100 A couple of years ago, when Michael Jackson died, there was a big brouhaha about all of the congestion that created.  If my ISP, for instance, has a link for 1 gigabyte, and all of a sudden there is a big show, or the wedding of Kate and Will, then, momentarily, you can use way above 400, and that month you will pay more.

5101 So, in terms of management, from the telephone company's perspective, it is a lot easier.  You don't have to make calls and change the bills, it's done automatically.

5102 One aspect that hasn't been covered so far -- and there have been allusions to this -- is commit rates.  This is applicable to 95th percentile especially.

5103 I will give you an example.  If my ISP is fairly steady every month, 400 megabytes demand on a 1 gigabyte physical link, he can call and say, "Look, I am going to commit to 400 megabytes," and Bell will quote him a rate for 400 megabytes, and he will have to pay that.

5104 So even if one month it goes below 400, it goes to 300 or 200, he still pays the 400.  So it's a commitment a bit like the AVP pre-purchase block, but it allows you to go above for the Kate and Will wedding or Michael Jackson's death.  It allows you to go above, and that month you actually pay for overuse or overage on this.

5105 On the Internet it's not regulated, but because of competition, ISPs tend to negotiate these rates, and, of course, the higher the commit, the lower the per-gigabyte-per-second rate will be, so the less you pay per gigabyte per second.

5106 Generally speaking, if you, for instance, commit to 500 megabytes on a 1 gigabyte link and you exceed that, the overage fee, if you go from 500 to 600, for instance, would be the same as from zero to 500.

5107 So it's basically a package deal that you have, where the overage is not more expensive, but you still pay for that use.  So it's not a disincentive, as AVP is, because with AVP, when you go above, you get a disincentive.

5108 In this case it is not a disincentive, but you still end up paying for what you use.

5109 The advantage with the commit rates, also, is that it helps the network operators to plan capacity.  In the case of my ISP, for instance, they buy 1 gigabyte, but Bell doesn't really need to plan for 1 gigabyte because they commit, let's say, to 400.  So they have a better idea of how much of that link is going to be used.

5110 I have already sort of said that because the ISPs already buy everything else in capacity based, moving GAS and TPIA to capacity based really simplifies everything from their point of view.

5111 It should also be said that, you know, that the five-minute samples that the routers take is also what the networks use.  They poll every five minutes.

5112 And I think, in a previous presentation and in a different life, Bell Canada mentioned they get little red lights you know whenever they get a congested link.  They get a red light showing up so they fix it.  That's because they poll the same number from their routers which they already have.  They poll the five minute sampling interval for capacity and when it goes above, you know, certain thresholds a red light appears.

5113 So they already have the software to poll these routers to get that stuff.  It's just a question of sending it to the billing people.

5114 I have already said that.

5115 Another important -- and this is probably significant to the CRTC, is that all the network equipment is sold by capacity.  You buy a router and it has capacity in gigabits per second either per line card or the backbone itself will have, you know, maybe 100 gigabits a second that you can put the cards in.

5116 So when you get costs from Bell we only see the hash marks but you guys supposedly get costs.  They are costs by capacity.

5117 If you want to verify that the gas rates match or are in line with Bell costs, it's a lot easier when both of them are in capacity because you can really compare.  When you go into a conversion between users and capacity it's an art.  And this changes over time.  I will give you an example.

5118 I download -- I listen to music on the radio at 125 kilobits per second.  It's very, very small.  You can almost do this on dialup.

5119 Or I can listen to that all day long and download 2 terabytes at the end of the month.  I haven't calculated how much it is, but I am giving you an example.  Or I can download 2 terabytes in two hours -- again, that doesn't work but you see the point.  The second one will have a huge load on the network whereas the first one has no load on the network.

5120 And so when you are dealing -- trying to convert capacity to usage, it's never perfect and it's never an exact science.  And so -- and this will change as usage patterns change.  If people start listening to radio a lot you will see their byte volume go up but the network loads won't.

5121 And for you guys, when you do the costing and stuff it becomes -- you will have to update this fudge factor all the time.  Lots are included in capacity base for the CRTC itself in terms of the rates.

5122 Usage numbers that we have in UBB, they are a marketing thing for the retail arm of the -- for the incumbents and ISPs.  One of the reasons, it's a lot easier to explain to someone, "You have consumed 25 gigabytes this month" as opposed to telling him, "Well, you have 25 megabits per second link but because you exceeded 10 megabits per second for more than 10 hours a month we are going to charge you more".  It gets a little complex for the normal person.

5123 For network operators, though, the capacity-based, even 95th percentile, is natural.  They do this every day.  They eat this you know every day.

5124 But for the consumer, UBB is simpler to explain.  But it's a purely retail aspect.  The usage is purely retail.  It is not used in networking except with Bell or the cable companies who want to put that in.

5125 Okay.  And there I come back to my first one.  The DSLAM costs when you fix a port at 25 megabits or at 5 megabits you have -- with capacity-based billing the 25 megabit user will consume more bandwidth on the network.  So the ISP will need to have more AHSSPI or more capacity base, a higher speed.

5126 So Bell Canada gets compensated for 25 megabits per second local user because the ISP buys more capacity.  So, in that case, because the capacity is paid for at the capacity level, there is no need to have different prices for different speeds at the DSLAM level because the incumbent is compensated at the capacity level.

5127 Very quickly, in terms of mark-ups, all I can say on this is, because 632 basically declared the last mile to be non-reproducible or very hard to reproduce, perhaps the last mile aspect should have a lower mark-up than the rest.

5128 That's -- you know I wrote a text in it but that's that.

5129 Getting to the end, I have got one here about grandmothers and bandwidth hogs.  Networks will generally look at averages and they say in the neighbourhood we have so many grandmothers and so many bandwidth hogs and they have averages and they plan for that.  If in a neighbourhood you have more bandwidth hogs than there are averaged then you get congestion because they consume more than the incumbent has planned for.

5130 There is no billing paradigm at this stage here that can fix that.  Bell and the others called about localized congestion.  None will fix that, whether AVP or capacity based, this is the job of the incumbent to manage their own network.

5131 The ISPs buy capacity and they tell the incumbent "Give me the service that I paid for" and it's up to the incumbent to manage this.  And if they see, for instance, someone mentioned -- I think it was Shaw mentioned university -- lots of bandwidth hogs there, they will put more stuff in there.  But if in a neighbourhood with retired people and nothing but grandmothers, they will put less capacity.

5132 So, at the end of the day, the capacity that you purchase at the port gets split evenly or unevenly but it still gets paid for and the ISP, the incumbent will actually allocate the money where it needs to be.

5133 And I will conclude now because -- just this.

5134 Just like Doctor Who can regenerate himself every couple of years as a new actor with a different face but still remains basically the same, I think the CRTC has an opportunity here to regenerate GAS and TPIA with better tariffs that will let these small ISPs move forward.

5135 It's important to have something that adapts growing and changing usage patterns which both AVP and capacity based, do.

5136 Give incumbents enough money to invest.  Both of these do because both of these adapt to changing usage patterns.  So if there is growth the incumbents get more money.

5137 We have to keep the margins low enough to give the ISP, the small ISPs, the chance to compete as opposed to just sort of try to match the price.

5138 And if the small ISPs can be more profitable then they can invest more.  So they can build more facilities and stuff.

5139 There should be no incumbent interference at the retail level.  As I said, sometimes it can be subtle.  AVP, on the surface, appears not to have it but there is a very subtle suggestion to small ISPs that you have to go this way.

5140 And again, the right billing philosophy will foster competition and the right billing philosophy will last a long time unless ISPs do long term planning and you get to see less of me.

5141 Thank you.

5142 THE CHAIRPERSON: You are using the expression capacity tariff rather -- capacity-based billing rather than 95th percentile and you make the point that 95th percentile is only one way of measuring capacity.

5143 MR. MEZEI: Correct.

5144 THE CHAIRPERSON: If you were in our chairs, or my chair, what would you recommend the industry should do?  When you say capacity what do you have in mind?

5145 MR. MEZEI: MTS who was before me saying, "Please don't change our method of doing" because their method works.  CNOC wants 95th percentile.

5146 Personally, I think a good compromise based on what the incumbents have said, would be 95th percentile with commit.  So the more you commit the lower the rate you pay, the rate per unit of speed.  And it also -- by having a commit, it also gives the incumbents assured revenues which is something I think they want.

5147 The 95th percentile, as proposed by CNOC, is interesting because what happens if a small ISP buys a 1 gigabit link and only consumes 100 megabits?  You know, they will only pay for 100 megabits and the incumbent, you know, doesn't get much money.  And if the guy goes on vacation for a month he will pay nothing.

5148 Whereas, a commit rate gives you a floor on how much the ISP has to pay.  And ISPs do this when they commit for Internet transit.  So it's fairly easy for them to do.

5149 THE CHAIRPERSON: I am so sorry.  I am not following you.

5150 Are you basically saying you are endorsing the MTS model?

5151 MR. MEZEI: Actually, what I am endorsing is for you to endorse perhaps any capacity-based model.  So you could let MTS keep their own because it is capacity based and you could get Bell to adopt 95th percentile because it is capacity based and you can get Shaw or the cable companies to adopt perhaps a different variation of capacity based -- as long as it is capacity based.

5152 And that would give perhaps a bit of competition between the different incumbents because of different -- the same billing philosophy but different implementations might -- some ISPs prefer one to the other; maybe choose and maybe the market -- the market of ISPs could choose which is best or --

5153 THE CHAIRPERSON: Okay.  Assuming we don't follow that route and we feel it has to be the same methodology for all, which one would you then endorse?

5154 MR. MEZEI: I would say the best one and the most flexible one would be 95th percentile either with or without commit.  I think that's -- you know, it's sort of -- the commit is really just a minimum floor of what you pay and perhaps changing the rate structure to be cheaper when you buy more.

5155 But 95th percentile would be the better one because it's the most flexible one and for smaller ISPs who don't need the full gig like my ISP, it also saves them money because they can buy the physical link now and not upgrade it every year and grow into it.

5156 THE CHAIRPERSON: Thank you.

5157 Candice, you have some questions?

5158 COMMISSIONER MOLNAR: I have a couple of questions that I am just going to follow on with what you have just been talking about.

5159 MR. MEZEI: Okay.

5160 COMMISSIONER MOLNAR: And I heard you say capacity based is what is important; not necessarily how it is structured as long as it follows the principles of cost-based capacity.

5161 MR. MEZEI: Yes.

5162 COMMISSIONER MOLNAR: But then you said, but the 95th percentile might be best.

5163 And your concern, I think I just heard you say, was with the smaller ISPs and they are -- potentially the cost of them committing to the capacity requirements of a model such as proposed by PRIMUS or MTS.

5164 This comment you put on page 6 of your slide where you talked about managing at the rate limit using the software versus the technical limits of the interface --

5165 MR. MEZEI: Yes.

5166 COMMISSIONER MOLNAR: -- would that address your concern?

5167 MR. MEZEI: Yes.  Actually, this is what Rogers does at the moment.

5168 ISPs, for instance, will buy a 1 gigabit link, a physical interface to the Rogers' POI and initially Rogers will give them maybe 100 megabits or 200 megabits.  And as their customer -- as they sign up customers to cable they will get congestion and then they call up Rogers and say, "Please upgrade this to 200 or 300 or 400".

5169 There have been stories of Rogers dragging their feet quite a bit.  This is the advantage of 95th percentile because Rogers doesn't need to get involved.  It's automated.

5170 If you want to use more up to the line speed you can use anything within it and you get charged for it.  Whereas, when you have the rate limit, then you have to call the incumbent and say, "Please raise it for me".  But it allows you to have fixed prices.

5171 For instance, Rogers can say a price for 100 and 200 and 300 and 400 and at 1 gig, you know, as opposed to something that's variable per gigabit per seconds.

5172 COMMISSIONER MOLNAR: Okay, fair enough.  And as you know, because I know you have listened to everything throughout this proceeding, some of the benefits -- and you talked about using commit rates.  I mean, that's some of the benefits that have been put forward for an MTS/PRIMUS type of approach.  It's essentially a commit.

5173 And the other significant benefit that has been put forward is it basically is simple to administer and, you know, simple and cheap and with less potential for billing disputes than something like a 95th percentile.

5174 MR. MEZEI: That's correct.

5175 COMMISSIONER MOLNAR: So let me go back to if we wanted to look at that kind of approach but address some of your concerns that it be -- accommodate smaller ISPs, perhaps more easily than what had been proposed by MTS and PRIMUS -- if we were to look at this rate limit as the point at which you can -- you know that you can commit to the interface at lower limits than the full interface capacity.

5176 MR. MEZEI: Correct.

5177 COMMISSIONER MOLNAR: And we were to put in place perhaps some requirements around the speed at which provisioning should took place, should there be a request to change that capacity, do you think that those two things together could help make the simpler model work?

5178 MR. MEZEI: Perhaps.  I will just note one of the major differences between what you are proposing and the 95th percentile -- and I will come back to Michael Jackson -- in such an event, when you have a rate limit let's say of 200 megabits on a gig link and Michael Jackson dies, you cannot physically go over that 200 megabits.  The router just will not allow more traffic than what you have paid for.  Now, whereas, 95th percentile automatically it goes up.  It allows you to go up.

5179 And that's the major difference between the two.

5180 COMMISSIONER MOLNAR: Everyone is going to be talking about Michael Jackson and not simply the wholesaler's traffic so that you know, the strains on the network --

5181 MR. MEZEI: Exactly.

5182 COMMISSIONER MOLNAR: -- exist overall.

5183 MR. MEZEI: Yes, yes.

5184 And in a case of the incumbents, by having a hard limit of 200 megabits on the link it means that when Michael Jackson dies -- again I use -- I should have used Kate and Will, sorry -- the ISP cannot use more than 200 megabits.

5185 So the congestion will happen at the ISP level and not at the incumbent level because the ISP cannot use more than 200.  Whereas, when you have a physical 1 gigabit link and you have a commit of 200 megabits, when Michael Jackson dies then you can go up to 1 gigabit and that could cause congestion on the incumbent's side.

5186 And I'm sure the ISPs will hate me for saying but that's the reality of it.  Whereas, when you have a fixed limit on the rate limit, you can't go over it so that really shifts the congestion onto the ISP.

5187 So if you have a rate limit like this and the ISP has, let's say, a 24-hour turnaround or 48-hour turnaround that you can ask to CNOC or whatever what it would need to be for it to rise, then that might be acceptable.  I think, you know, that would palliate the effect of that.

5188 COMMISSIONER MOLNAR: Okay.  And we will ask how reasonable it is to use the rate limit versus the actual throughput limit of the link next week, and how reasonable or in what time intervals such changes could be made.

5189 And we will ask that to the ILECs, I think --

5190 MR. MEZEI: Okay.  But I mean, I can tell you the big advantage of this is that with technology today, 1 gigabit links are pretty standard in terms of the technology, Internet cables and the fibre and stuff.

5191 So it's a lot easier for you, especially in the case of the cable where you have to pay to install the ports, it's easier to install a 1 gigabit port and that will last you a few years, than having to install a new port every year as you grow.

5192 Logistically, it's easier to buy a bigger pipe and then have a tap on it to limit it and then you just open the tap as you need it.

5193 COMMISSIONER MOLNAR: I need to ask.  I know that in the comments that you provided you gave comments on cable as well as the ILEC as the underlying carrier.

5194 MR. MEZEI: Yes.

5195 COMMISSIONER MOLNAR: Yet I feel that your conversation today is very driven towards you know the ILEC's incumbent network.  So can you tell me, based upon your knowledge, whether what we have been discussing here is something that can be appropriately applied within the cable network?

5196 MR. MEZEI: I have to cut it a bit down.  I did have a lot to say about cable that I was going to say on Monday but what surprised me and what surprises me is how cable says how different they are from ILECs.

5197 I will be honest with you.  They are not that different.  The last mile is different, you know, between the CMTS and the cable modem is very different.

5198 But between the CMTS and the cable modem it's essentially glorified Ethernet which is the same as what they use between the CMTS and their routers.  So the same sort of philosophy of packets that can collide and you have you know, contention between packets and stuff -- very similar principles.

5199 And Bell says it has it has its congestion not in a DSL side but between its routers within its own environment.  The cable company never mentioned that but I'm sure they have problems there too, you know, because they have to upgrade those paths as well as they up the speeds at the end-users.

5200 So that part of the network is, I would say, nearly identical.  Yes, nearly identical because at the end of the day they have got capacity and they have got traffic going -- flowing through switches and routers all the way to where they can connect to an Internet transit provider, you know.  They are identical for that.

5201 I think all the concepts that I have mentioned do apply to both.

5202 COMMISSIONER MOLNAR: So this interfaced link -- and you were actually commenting on Rogers, so clearly this is appropriate -- can be applied at that point?

5203 MR. MEZEI: Yes, yes.

5204 COMMISSIONER MOLNAR: Okay, thanks.

5205 I want to ask you one more question --

5206 MR. MEZEI: Yes.

5207 COMMISSIONER MOLNAR: -- before they shut me down here.

5208 In your final reply that you submitted -- I'm not quite sure of the date anymore, but you made a comment that I thought was interesting under "General".  You said:

"Tariffs must be structured to be simple, fair, competitive..."

5209 And then you said:

"...and more importantly, independent ISPs must be considered to be partners of incumbents."

5210 And I wonder when you say that is there something that you would suggest that we, as the regulator, could do to have them be considered partners?

5211 MR. MEZEI: That was on my first page, I think.

5212 Make sure that when you submit -- when you publish your tariffs that you tell the world that these tariffs fully cover the costs of not only operation but also the capital costs for the incumbent.  Because at the end of the day, when a customer pays, buys in bulk -- because these ISPs are big customers for -- I mean they are not like me or you.  They are a big corporate company.  They are corporate customers.

5213 They become partners, like banks are partners with the telephone companies because they buy so much.  And when you buy so much a big part of your money goes towards the investment that Bell makes.  Bell doesn't complain about banks getting a free ride.  Bell sees the banks as a partner, you know.

5214 So that was the spirit of the partnering aspect of it because, at the end of the day, once the incumbents are satisfied they get enough money to pay for both operation and investment, they shouldn't poo-poo the ISPs, you know.  They should see them as important customers or partners.

5215 I think there is a term called "co-opetition" I think.  I don't know if I pronounced it properly.  So that is that aspect that I was focusing on.

5216 COMMISSIONER MOLNAR: Okay.  Well, those are all my questions.

5217 I would just note that while we can make the statements and you know the underlying carriers will understand what kind of margins are built in between that, we can't make them treat people or consider people differently.  So that's a challenge, I think, that resides in the industry.

5218 Those are my questions.

5219 THE CHAIRPERSON: Thank you.

5220 Michel?

5221 CONSEILLER MORIN : Merci, Monsieur le Président.

5222 Ce que je comprends du modèle de MTS, c'est que, avec le modèle de MTS, il y a vraiment un mur de Chine entre les petits et le gros fournisseur.  Il me semble que ce modèle, ce mur de Chine n'existe plus comme tel avec le modèle de CNOC.

5223 Et, au fond, est-ce que... avec le modèle de MTS, comme vous l'avez dit, est-ce qu'on ne favorise pas, au fond, beaucoup plus d'investissements pour les petits parce qu'ils achètent la capacité, et, au fond, est-ce que le compromis ne serait pas, même si vous venez de dire que les mêmes principes s'appliquent, pour ce qui est des telcos, qui représentent 90 pour cent, actuellement en tout cas, des consommateurs qui s'approvisionnent auprès des petits fournisseurs d'Internet, est-ce qu'on ne pourrait pas avoir un compromis où le modèle MTS s'applique aux telcos et le modèle CNOC aux câblos?  Est-ce que ça ne serait pas le compromis ou est-ce que MTS, selon vous, est jouable dans les deux cas?

5224 M. MEZEI : Quand vous parlez de MTS jouable, vous parlez de modèle?

5225 CONSEILLER MORIN : De modèle, oui.

5226 M. MEZEI : O.K.  Écoutez, le modèle de MTS ou le modèle de CNOC sont applicables sur le câble ou sur le telco sans aucun problème.  Dans le cas du câble, ils vont se plaindre à cause du last mile, mais essentiellement, quand t'achètes un certain montant de capacité, tu t'attends à ce que l'incumbent, que ce soit téléphone ou câble, te la donne.

5227 Quand tu vas dans un avion, tu prends ton siège et tu sais que la compagnie d'aviation a ses pilotes et paie tous les frais d'aéroport, et caetera, et que l'avion va se rendre à destination.

5228 La même chose quand un ISP achète de la capacité chez un incumbent, il s'attend à ce que l'incumbent fasse le management de son réseau pour lui fournir ce qu'il a acheté, et que ça soit le 95e pourcentile ou le modèle MTS où c'est fixe, il n'y a pas grande différence.  T'achètes quelque chose, il te le fournit.

5229 Alors, on arrive vraiment sur des points juste de facturation et de complexité de logiciel de facturation, et caetera, où le modèle MTS est plus simple à implanter.

5230 Et, d'ailleurs, Bell, c'est déjà implanté parce que, aujourd'hui, le tarif GAS est comme ça.  C'est juste une question de moduler le prix du AHSSPI pour que Bell soit content des revenus.  Ce modèle-là fonctionne et a fonctionné chez Bell pendant très longtemps, parce que c'est ça qu'on a en ce moment.

5231 Et 95e pourcentile s'adonne à donner une plus grande flexibilité pour l'acheteur, le ISP.

5232 CONSEILLER MORIN : Pour les petits?

5233 M. MEZEI : Pour les petits et les gros, parce qu'un gros, écoutez, on parle comme Teksavvy qui a...  Je ne sais pas combien cette semaine qu'ils ont, mais ils grandissent beaucoup.  Là, ils sont rendus, je pense, autour de 20 gigabits de liens.

5234 Bien, pour eux aussi, quand il y a des peaks et des pas-peaks, le 95e pourcentile, ça marche très bien pour l'Internet, l'Internet en tant que tel, les liens avec l'Internet, et ça marcherait très bien avec les liens avec Bell.

5235 Donc, si les ISP préfèrent le 95e pourcentile -- ils ont déjà exprimé cette préférence-là -- peut-être que vous pouvez leur demander la semaine prochaine quand ils vont passer de commenter sur le modèle fixe de MTS versus 95e, et leur demander d'expliquer pour eux c'est quoi l'avantage réel versus celui de MTS.

5236 CONSEILLER MORIN : Merci beaucoup.

5237 Ce sont mes questions, Monsieur le Président.

5238 THE CHAIRPERSON: Thank you.

5239 So you will be first up Monday morning.

5240 MR. MEZEI: Yes.

5241 THE CHAIRPERSON: Maybe you can elaborate on two things that you answered but not -- number one is the question of Commissioner Molnar of how to make the telcos and ILECs regard the ISPs as partners.

5242 MR. MEZEI: Okay.

5243 THE CHAIRPERSON: And secondly, you said the cable companies exaggerate the difference of their network, et cetera.

5244 If you could give us some more detail on that that would be appreciated.

5245 MR. MEZEI: Okay.

5246 THE CHAIRPERSON: Okay, I think that's it for today,

5247 Madame la Sécretaire, over to you.

5248 THE SECRETARY: It is, Mr. Chairman.  We will reconvene Monday morning at nine a.m. for the rebuttal phase.

--- Whereupon the hearing adjourned at 1303, to resume on Monday, July 18, 2011 at 0900


Johanne Morin

Carmen Delisle

Monique Mahoney

Jean Desaulniers

Susan Villeneuve

Karen Paré

Date modified: