Draft Policy Directions - Please provide your feedback (new deadline: 22 Feb)

[dudesweet]

My zwei-cents on the wholesale open-access network.
I don't think it is true that wholesale means no, or even necessarily "low" profits in the traditional view of what that would entail (eg a stagnating and or struggling business). "Wholesale" in this context needs to mean (if it does not already, then this must be addressed) that the entity operating such a network makes at the very least a normal profit, if not a super-normal profit.
In the material world, we have wholesalers, for eg the base material to make LCDs. They sell these to say Sony, Toshiba, Samsung and the like who add value to these materials, and use proprietary technology to differentiate their products. The wholesaler, is by no means struggling. If Toshiba steals Samsung's market share the wholesaler will not be affected. The pie's been cut up a different way, sure - but the wholesaler is still getting to sell the entire pie anyway.
If Sony through innovation manages to increase the size of the market (maybe through lowering their own costs, and thus stoking greater demand) - the wholesaler benefits too.
it is thus not a given that wholesale means poor or low profits - it all depends on our assumptions of the pricing model adopted by the operator will be. If they shoot themselves in the foot by not assessing their costs correctly, then this is merely a question of poor business decision making or risk assessment, rather than a question of retail vs wholesale as competing models.
I support the wholesale concept as long as it is implemented in such a way that this wholesale operator (let's call it Optimus MegaCom) is sufficiently bound by a competent regulator to oblige by evidence and policy based roll-out targets and objectives. (e.g
1) Use it or lose it.
2) Transparent and clear open reporting at regular intervals
3) Regular accountability tribunal meetings where the public get to ask/question/complain direct to management)

[ellipsis]

Have a read: http://dl.dropbox.com/u/24829246/MyB...ns_310112.docx

Comments, criticisms, suggestions, edits, etc all welcomed.

I'll check back in at 15h00, make any changes required and submit by 16h00.

[WireFree]

Have a read: http://dl.dropbox.com/u/24829246/MyB...ns_310112.docx

Comments, criticisms, suggestions, edits, etc all welcomed.

I'll check back in at 15h00, make any changes required and submit by 16h00.

Looks good.

In my opinion spectrum is a national resource. Assignment of high demand spectrum should not be monopolised by a few entities. The spectrum assignment process should also aim to increase competition in the supply of wireless services as competition will help to drive down prices.

The currently assigned GSM and UMTS spectrum is assigned to a few entities who do not provide wholesale access to their radio network to entities who do not have spectrum in these bands. ICASA should be required to investigate other avenues where licensed operators can access spectrum held by the incumbents, at a wholesale level. One example would be MVNO Regulations similar to those in
- Brazil (http://www.internationallawoffice.co...e-6c14d779828f) and
- India (http://www.ictregulationtoolkit.org/...tion.3636.html and http://www.ictregulationtoolkit.org/...ument.3636.pdf)

WireFree

[dominic]

hi WireFree

one of the outcomes of the LLU "process" is that ICASA will be looking directly at how to introduce competition in services over the MNO networks and MVNOs are obviously one of the prime options. Have no idea when this will be done though...

pricing of wholesale products by the MNOs is currently a massive problem waiting for someone with big enough testicles to take it to the Comp Comm (who i think were looking at this towards the end of last year of their own volition but not sure of the outcome)

[WireFree]

one of the outcomes of the LLU "process" is that ICASA will be looking directly at how to introduce competition in services over the MNO networks and MVNOs are obviously one of the prime options. Have no idea when this will be done though...

But I am sure the Minister can still include this in her Policy Directions as the currently assigned GSM and UMTS Spectrum is also "High Demand Spectrum". Then at least this task would get her Stamp of Approval.

WireFree

[ellipsis]

The DoC has extended the deadline for submission (the day after submissions were due...) - see http://dl.dropbox.com/u/24829246/Ext...201_140429.pdf

If you have anything to add, please post by 22 February, and I will update the submission accordingly and re-submit.

[WireFree]

Will Mybroadband be responding to the ICASA Framework for the Licensing of High Demand Spectrum?

[dominic]

But I am sure the Minister can still include this in her Policy Directions as the currently assigned GSM and UMTS Spectrum is also "High Demand Spectrum". Then at least this task would get her Stamp of Approval.

WireFree

agreed - will be included

[dominic]

Will Mybroadband be responding to the ICASA Framework for the Licensing of High Demand Spectrum?

yip - will be a dedicated post once we get the new timelines which ICASA will announce shortly

[WireFree]

yip - will be a dedicated post once we get the new timelines which ICASA will announce shortly

New Timelines? I thought the deadline was extended from end January to end February!

WireFree

[dominic]

New Timelines? I thought the deadline was extended from end January to end February!

WireFree

for the policy directions...this will have an impact on the ICASA process which will probably be kicked back a month

[RavenN]

We welcome the intention to increase the availability of 800Mhz spectrum by adopting a wholesale open access model. However, we are concerned that the Department, and indeed ICASA, may not fully appreciate the difficulties associated with a Wireless Open Access model, which differ quite fundamentally from a Wired Open Access model.

The Business Model that Telkom uses for third party ADSL services is NOT an Open Access model. It is a reseller model suited to ECS licensees and not ECNS licensees as ECNS licensees cannot do any meaningful self-provisioning or product and service innovation outside of content services as Telkom controls all aspects of the network service.

We must caution against developing a Wireless Open Access model that mimics the Telkom ADSL model as that will be counterproductive and prevent the Department from achieving the stated objectives as per Section 2.1.1

The challenge for Wireless Open Access and the fundamental difference with Wired Open Access is that in the case of Wireless, the Provider Edge and the last mile connection are tied together because of the spectrum and technology used.

In Wired Open Access, the last mile – typically copper or optic fibre – is a passive connection. In the case of fibre, multiple service providers can be accommodated on the provider and the customer side, without interference and largely technology independent.

With wireless, however, the provider edge technology has a fundamental impact on last mile access and multiple service providers cannot share spectrum allocation at the provider edge if there is to be any quality of service provision. Best effort services, in which service providers share provider edge spectrum on an as-available basis, increase the complexity of the Wireless Open Access Operator and reduce customer experience significantly.

Although the Policy Directions do not go into detail, we would like ICASA to consider the option of whether the Wireless Open Access Operator is going to be licensed to have only a Provider Edge network, or if it will be licensed as a Full Service wholesale wireless operator. Although there are technical issues that arise, it is possible to have an Open Access Provider Edge network (Or Radio Access Network), with meet-me or co-location facilities for third parties to interconnect with the Provider Edge network.
The benefits of having only a Provider Edge Open Access network is that a full service network adds costs to the provision of services and often duplicates much of the network infrastructure of 3rd part operators, taking investment away from increasing the number and reach of the provider edge (and hence reducing the availability of broadband for all). Additionally, only ECNS licensees will be able to use the Provider Edge as they will have to self provide to interconnect their networks.

Provider Edge Open Access works best when the billing for services provided is simple, e.g. a flat fee per MHz per base station per month.

The key principle underlying a Provider Edge network is that it addresses directly the reason for a Wireless Open Access system i.e. that last-mile spectrum is scarce and must be shared. Backhaul, long haul and extensive and complex networks that are best covered by other technologies and network infrastructure should not be bundled into the operations of a Wireless Open Access Provider.

A critical aspect of Open Access is the ability of 3rd party operators to use the last mile connection for self-provisioning. If the Wireless Open Access Operator is going to allow self provisioning for the last mile, it would mean allocating spectrum on the last mile to 3rd parties and allowing 3rd parties to provide their own provider edge equipment for the allocated spectrum.
In the above scenario, the critical question is: How is the Open Access Operator going to allocate last mile spectrum? It is crucial that ICASA address this issue, as the licensing of the Open Access Operator must not be seen as an abdication of ICASA’s responsibility for spectrum management to the Open Access Operator. Scarcity of spectrum means that the mechanism for allocation of spectrum to wireless retailers must be well-defined otherwise economic power and profiteering will be the determining factor as to who gets Spectrum in high value- urban areas. As last mile spectrum is scarce, the potential for profiteering and a focus on lucrative urban markets by the Open Access Operator is a distinct possibility, which again runs counter to the objectives of Notice 898.

If the Open Access Operator does not allow self-provisioning on the last mile, then it becomes the gatekeeper of the last mile connection. If this is the case, then we do not have true Open Access, but network sharing on the last mile.

A combinational award also contradicts the requirements of Section 2.1.1.1 “Introduction of new entrants” as combination awards reduce the number of awardees rather than increase the number of new entrants. Although this is a simplified argument - there are options for combining National 2.6GHz with rural 800MHz and refarming the urban 800MHz on a local basis.

We are also concerned that the Policy Directions do not address the issue of the following:
• Low power 2.6GHz or 800MHz licenses for localized use. The campus deployment of low power devices can drive down costs and provide innovative solutions for customers
• Imposing wholesale mobile network resale obligations on mobile communication providers. The large number of innovative virtual mobile Network operator’s in many parts of the world is testimony to the competition and innovation they can bring to the market. Current regulations do not force operators to offer wholesale services.


The danger is that the Open Access Operator has to then make technology and business standards decisions that can determine how 3rd parties provide service, and indeed which 3rd parties use the last mile, as this is not a technology neutral solution. We must also be wary of the Open Access operator replicating the monopolistic ADSL access model (see earlier), thereby creating a new monopoly in wireless provision.

[WireFree]

We must caution against developing a Wireless Open Access model that mimics the Telkom ADSL model as that will be counterproductive and prevent the Department from achieving the stated objectives as per Section 2.1.1

...

If the Open Access Operator does not allow self-provisioning on the last mile, then it becomes the gatekeeper of the last mile connection. If this is the case, then we do not have true Open Access, but network sharing on the last mile.

Good analysis. What is your preferred access model?
Is the network sharing model bad, if it provides more operators the ability to provide services on the last mile?
What models have been successfully implemented else-where in the world?

We are also concerned that the Policy Directions do not address the issue of the following:
• Low power 2.6GHz or 800MHz licenses for localized use. The campus deployment of low power devices can drive down costs and provide innovative solutions for customers
• Imposing wholesale mobile network resale obligations on mobile communication providers. The large number of innovative virtual mobile Network operator’s in many parts of the world is testimony to the competition and innovation they can bring to the market. Current regulations do not force operators to offer wholesale services.

It is also important that you define what type of MVNO you prefer. Virgin Mobile for example is not a licensed operator and distributes rebranded Cell C SIM cards. They are referred to as a MVNO in all the media. A Full-MVNO will have it's own infrastructure.

WireFree

[RavenN]

Good analysis. What is your preferred access model?
Is the network sharing model bad, if it provides more operators the ability to provide services on the last mile?
What models have been successfully implemented else-where in the world?
WireFree

ICASA is trying to be as neutral as possible about 2.6GHz and 800MHz, but the reality is that these bands are for mobile broadband, and LTE in particular, and most European regulators have allocated them as such. I would like to see the following:
- Get Sentech and WBS out of the 2.6GHz band, freeing up 2 allocations. This will allow up to five operators in this band
- Use the Uk model for auction of the 2.6GHz band. The UK auction is meant to create as much competition as possible while preventing monopoly. In order to do this, a full spectrum allocation needs to be considered and operators may have to give up spectrum in other bands to accomodate new operators.
- Allocate a low power range in the 2.6GHz band for campus type operators (WiFi on steroids).
- Combine the National 2.6GHz allocation with a rural 800MHz allocation.
- Allocate 800MHz in Urban areas to "new" local operators
All the allotments are done on auction, so there is NO Open Access Operator

If you forced me to choose an Open Operator model, then i would say the Open Access Operator dividing up the spectrum on the access level would be better as it allows operators to innovate with access technologies.

As to what has worked elsewhere, the wireless open access model is widely touted in East Africa and Russia. The USA looked at it but never implemented it. New Zealand has the Spectrum Park model, which is a complex collaborative model. Wireless Open Access is not easy and hence there is little widespread usage thereof.

It is also important that you define what type of MVNO you prefer. Virgin Mobile for example is not a licensed operator and distributes rebranded Cell C SIM cards. They are referred to as a MVNO in all the media. A Full-MVNO will have it's own infrastructure.

WireFree

The reality is that we do not have MVNOs in the strict sense - A MVNO should be able to buy a bulk package of minutes on a network and repackage these as it sees fit. It can then sell its own SIM cards with these packages and creat arbitrage opportunities based on bulk buying. Obviously the SIM will be registered with the operator. Some MVNOs do have their own infrastructure, but its not a requirement. Wholesale mobile minutes do not exist in this form as the operators have kept a tight grip on wholesale and offer a form of "resale" in which the package sold by the MVNO is largely "approved" and done in conjunction with the Operator

[WireFree]

All the allotments are done on auction, so there is NO Open Access Operator

How do you increase competition with NO Open Access Operator? The market is still limited to a few players that have exclusive access to the access networks.

The reality is that we do not have MVNOs in the strict sense - A MVNO should be able to buy a bulk package of minutes on a network and repackage these as it sees fit. It can then sell its own SIM cards with these packages and creat arbitrage opportunities based on bulk buying. Obviously the SIM will be registered with the operator. Some MVNOs do have their own infrastructure, but its not a requirement. Wholesale mobile minutes do not exist in this form as the operators have kept a tight grip on wholesale and offer a form of "resale" in which the package sold by the MVNO is largely "approved" and done in conjunction with the Operator

The MVNO model you describe above is simply price based competition, no service differentiation - aka the Virgin Mobile type model. An MVNO that owns it's own infrastructure also has the ability to offer converged services across different technologies (such a fibre, WiFi and possibly in the future over the Telkom LL) and not be limited to the vanilla mobile service offered by the host operator.

WireFree