Our asks from the Narrowband Review

 

By Simon Woodhead

By rights the 2020 Narrowband Market Review should now be under way and rumour was it’d land at the back-end of July. We’ve had a few friendly discussions with Ofcom and BT recently around technical issues that were time-critical matters of dispute. Whether they are addressed in the review, and that accounts for the delay, or whether we’ll need to publish reams of technical data that support our allegations so they are addressed, remains to be seen. It crucially affects our response either way.

However, as the bleating has started from certain quarters we thought it might make sense to take a step back, ignore the specifics of our technical battle for now, and focus on the war. So whilst it might come as a surprise, I’d like to try and be the voice of moderation, cognisant of the jelly Ofcom need to nail to the wall here, and be somewhat more respectful and optimistic than I may have been in the past!

TL;DR Those that are calling for ‘regulated IPX’ are missing the point. It is regulated. As much as TDM. The issue is not the technology, it’s BT’s long standing market power in transit to minor operators and in number portability that needs addressing.

“Regulating IPX”

This call is somewhat naive. IPX is regulated, to the same extent as TDM. The only difference is that the cable that plugs into the TDM switch is charge-controlled, the IP equivalent is not. While there is an IP reference offer, what it isn’t is subject to all the provisions of the Standard Interconnect Agreement we know and dare I say love. 

BT has freedom to contract for and charge for its IP products to the same extent (except for that little cable) on TDM. That means, what I’ve called ‘the Secret Club’ is a fair right for BT to manage within the boundaries of Competition Law. 

Where we start to come across a problem, and what we say should be the first piece of jelly Ofcom attempt to nail to the wall is defining what IPX actually is

We have to remember that IPX resellers (more formally, BT IPX Type A customers) are just that and any delusion otherwise is dangerous. They are generally of a particular type which, without going into too much detail, we call ‘IPX Scrotes’ and the prospect of giving them equivalence to national scale carriers is terrifying. 

Conflating the IPX Type A platform with the more traditional carrier interconnect, such as that available under the SIA, also distorts a market, not just BT’s commercial model, but that of those of us who compete very well with what IPX offers but for (in our opinion) a higher calibre of customer!

Which brings us back to “regulating IPX”; it already is regulated. At no point in any regulation (except that cable) or charge control, is there reference to any TDM specific technology. It transpires, the ability to interconnect to BT’s IP switches is already the right of any carrier. That entitlement exists in several forms, in the General Conditions and in SMP Conditions and is completely technology agnostic. In that sense, “regulating IPX” is a completely unnecessary ask – so I say leave IP interconnection with BT alone as, rather like a colostomy bag, you might not like what comes out of it if you poke it.

But we do need to know what IPX actually is. Is it a reseller platform as BT seem to claim, hosting numbers for managed service customers on the BT network and paying them a commercial proportion of the regulated payment? Or is it transiting to these resellers, as we understand BT also claim, such that BT have no obligation to handle porting and resellers can hide behind the well known ‘Scenario 7 problem’. 

If it is hosting, as we would say it is, given it competes with our own offering and that of other networks, then there should be no Scenario 7 problem – BT should discharge the resellers porting obligations as you’d expect from such a managed product. If it is transit, as BT also claim, then why does the Scenario 7 capability even exist, why do resellers not get the charge-controlled FTR as a termination payment, and why do they not (generally) honour their responsibilities to establish interconnect and porting agreements with other operators like Simwood? For whatever reason, there are inherent issues with the logic here and we hope the regulator and/or BT can choose to ride one horse or the other.

We know Ofcom are looking at this as we have posed some of the questions privately but we strongly suggest they make a Section 135 request to every BT IPX Type A customer, asking whether they consider themselves a network with all these responsibilities or whether they consider they have outsourced this responsibility to BT. We strongly suspect, based on our own experience, that the answer will be the latter. It then follows that IPX isn’t transit and the numbers should be treated as BT’s for both settlement and porting. Resolution of this issue would solve many of the industry’s worst problems in a stroke.

BT’s dominance in Transit

BT’s Transit prices are not subject to price controls and in textbook monopolistic behaviour, as volumes have decreased, prices have risen. Sure, competition to the major terminators is plentiful, which is why I doubt anybody sane uses BT transit to reach any mobile network or increasingly for geographic numbers – the large networks which are fully meshed with each other benefit from charge-controlled rates and dramatically undercut BT. In fact we dramatically undercut BT! Ofcom have done great work here, so why does it matter what BT does?

I answer that with a simple question: does BT (without counting EE) use this plentiful competitive transit?

If so, Simwood can presumably terminate our interconnect with BT as traffic will find its way between us on commercially preferable terms over other operators we’re interconnected with, and BT will then have an incentive to negotiate equivalent terms with us for a new interconnect. In other words, if the transit market were truly competitive, without a dominant player in the middle, transit prices would tend downwards to the cost of provision. Information in the public domain about BT’s transit pricing would suggest this is not the case at all.  

We’d love to operate our network without having to have BT in the fray. It’d complete BT Zero in a heartbeat, and result with all our traffic ingressing and egressing over modern IP interconnects that are on fair and reciprocal commercial terms, despite some being with companies many times BT’s size globally, and life would be easier. For one, we wouldn’t receive an 80 page invoice every month, 79.5 pages of which represent charges reflecting BT monetising other people’s network choices.

As far as I can tell, while there are many, many other paths into the Simwood network, BT do not use them. Is it truly a competitive market place if one can only exist as a network operator if interconnected with BT? Even worse, given how the SIA works, is it truly competitive if BT has to give their permission for an operator to build a range, or use a price point, or make a change?

This thought experiment would suggest that it is not and a regulatory approach that treats BT as “another carrier in the mix” is premature. That, certainly, is the goal. We yearn for the day that BT is just another network operator, but today it isn’t, through regulatory construct, deliberate act or chance, BT still remains at the centre of the UK telecommunications universe. 

BT’s dominance in Porting

BT started life as the state-owned monopolist. In 1984, when privatisation started, by definition it had 100% of all of the allocated telephone numbers. Through the duopoly and cable years, a couple more players were added to the mix, but today, we see 70% of calls to geographic numbers either made to, or made presenting, a BT number, despite BT’s own retail market share being less than 40%. 

Ofcom claims that BT have Significant Market Power in termination, as it does Simwood, and every range holder, including many BT IPX Type A customers. Our porting regime works on an onward routing basis, i.e. calls go to the Original Range Holder (ORH) and are then forwarded on to the Recipient Communications Provider (RCP). There is sometimes a Transit element here too but let’s consider that dealt with above.

We’ll look initially at ports from BT as, after all, BT is still the biggest losing provider in the market.

In 2013, Simwood highlighted that the proposed 86% reduction in FTR (which the RCP can bill BT, and BT bills the calling network) would result in it being substantially below the APCC (the amount the ORH can charge the RCP for onward routing the call). Ofcom considered it out of scope for the Narrowband Market Review but this change made numbers ported from BT loss-making for the industry and generated a windfall for BT. Subsequent action by Vodafone and Gamma changed this with BT obliged to calculate an APCC based on LRIC (Long Run Incremental Cost) – the same economic model as the FTR.

We can’t quote our specific BT rates as BT have used the confidentiality clause to terminate someone’s SIA before, but suffice to say our APCC from BT remains higher than the FTR and we lose money on every minute of every call to every number imported from BT. We have done so since 2014 and will not be alone I’m sure. Ofcom may wish to procure the extent of this harm and the rates involved from either BT or the marketplace to fully appreciate the position here. They may want to say that the APCCs are out of scope of the Narrowband Market Review, however, in doing so, they need to answer this simple question – if someone has SMP termination, in an onward routing regime, how can they not have SMP in number portability? The two go hand in hand, like Thelma and Louise, or, given the pain that is our porting regime, Bonnie and Clyde. 

While when Simwood terminates a call to an imported BT number, the loss is limited to the FTR minus the APCC, the effect to the industry is wider. The originator, unless interconnected at the DLE, will have paid BT the FTR plus conveyance. Conveyance being the term BT uses to describe transiting around its own network. 

So, for all those calls to customers BT has long lost, BT are earning revenue at both the originating and the terminating end. 

Add it all together, using figures we obtained in the public domain, you see that BT can make upwards of 20x FTR for a call to a customer they have lost. 

That wouldn’t be so bad if there was reciprocity. But as you quickly realise in this market, what is good for the former incumbent is not necessarily good for the rest of us. For numbers that have ported to BT, and as far as I can determine just ‘because’, the APCC some operators are able to charge BT is half of that they charge them. This situation is unique to BT, all our other interconnects for porting are on reciprocal terms – just like they are for transit, and engineered to ensure neither party has the opportunity to profit from losing a customer. 

In this, we also have to consider that TDM switching and capacity is the most expensive in our network, which means each call to/from BT has the highest network cost of any we switch. Add it all together, we lose money on every minute of every call to every number which is ported out to BT too. 

If it walks like a duck, quacks like a duck, and tastes good with plum sauce, it’s a duck. If you can’t enter the market without dealing with BT, and BT is able to leverage its position to influence how you can enter the market, and enjoy the ongoing benefit of being the former number range monopolist, and impose terms which are not reciprocal, then, the chances are, it’s a dominant player with SMP. 

Narrowband Interconnection

Ofcom has done a great thing in mandating operators to provide wholesale access to each other at a charge-controlled call price and other regulation affecting the ancillary components. This gives operators like Simwood the opportunity to avoid transit by exchanging traffic directly with other operators, improving quality, security, resilience and economics. There are two issues though:

  • Due to the ongoing benefit of being the former number range monopolist we discuss above, most geographic calls are still to numbers in BT ranges and in an onward porting world they have to go to the ORH, i.e. BT. This limits the effective volume that can be exchanged over direct interconnects, and thus caps the effect of competition and therefore the benefits to be derived for consumers.
  • Rules without enforcement are pointless. Simwood would willingly enter into a bilateral with everyone in the marketplace but we know two things:

    • IPX Type A Customers view BT as having this responsibility and generally won’t. That could be easily dealt with as described above.
    • Other operators are hard to deal with because getting the request (keeping in mind this is an entitlement) to someone that is suitably knowledgeable and empowered is difficult.

Ofcom could, and in our opinion should, maintain a register of nominated contacts across operators to facilitate interconnection and keep an eye on the situation – after all, they say that a free market in interconnection is required to safeguard competition. Given the interrelationship in dominance between termination and porting, arguably this should be extended to porting establishment too. We know that Ofcom has a responsibility to collect fees from the vast majority of the operators in question, and that they maintain a Number Management System, so we do not see this as overly burdensome either. We just hope they don’t outsource it to the OTA if we want it to be done in any reasonable way or timescale. 

Climate Change

Oh yes, I’ve gone there.

Whether you’re a believer or not, I doubt anybody will deny that waste is bad and sustainability is good. Indeed, a multitude of statutes and government policies say so.

With that in mind, we’d be interested in Ofcom making a Section 135 request to BT to determine the energy costs of both its TDM network and its IP network, on a per call basis – we’d be surprised if this would be at all burdensome given that much of this information we’d expect to be collected for Ofcom to build its economic models when setting the FTR. 

Our expectation is that IP is cheaper, moreso by the avoidance of multiple hops through high power switches around the country. Should that be so, is it legal, let alone morally defensible, for Ofcom to impose a regime where BT is incentivised to maintain a TDM edge even when it migrates itself to an IP core, as has been proposed by BT in the past? BT maintains that other operators can enjoy IP too and similarly invest in a TDM edge to meet BT’s TDM edge. Double TDM in addition to IP where a simple fibre connection between IP networks is not in our opinion justifiable ecologically, morally, let alone from a fairness and engineering sanity perspective. 

Solutions

The facts underlying the above opinions and experience can easily be determined with some targeted information requests from the regulator, if time permits ahead of the consultation on the Narrowband Market Review being published. We think the answers are very important to the industry and should critically shape any action taken – especially as we may have to live with the outcome for 5 years.

Presuming that those answers are as we expect them to be, this really boils down to recognising BT’s market power in transit (to non-major operators) and in number portability. One classical remedy to that is a charge control, but another, and one we favour long-term is promoting an alternative. 

We have suggested a contact register for interconnect above, but that is only part of the solution and more is required, perhaps:

  • A repository for current CP lookup for a given phone number. We have not shared Ofcom’s optimism around the blockchain solution but are willing to be proved wrong if it is ever successful. Of course, commercial solutions have long existed in the market and are used in every other developed economy too.
  • A shared protocol which enables operators to announce on-net numbers to each other, with a safeguard of routing to the ORH in the event of any conflict or doubt (e.g. two operators announcing the same number).

Taking the second option as more viable, Simwood already has APIs in place which are available to other operators to consume and know which ranges are reachable via Simwood, and indeed which ported numbers are reachable via Simwood. Without boring you all today with the detail, Simwood has a high level design that could conceivably run on a Raspberry Pi and service the entire industry; that isn’t the final solution we propose but it serves to contrast with the deployment overhead and expense of the alternatives, as well as illustrate the fundamental simplicity of the problem. 

We actually see little difference in the routing of telephone numbers to the routing of IP addresses, with operators able to announce those they can terminate (or indeed transit) to others. This could be unilaterally or multi-laterally, over bilateral interconnects or through a shared fabric. Recipient operators would route to the most specific matching prefix that is available over a party they have a relationship with, just as in IP. The agreed protocol and potentially the fabric would facilitate this and encourage adoption.

Whilst we are willing to contribute further to this, past efforts seem to have been thwarted by the glacial pace of large operators, intellectual masturbation and a lack of regulatory engagement. Any effort needs to remain technically simple and focus instead on adoption, with minimal central logic to handle routing conflicts (by rejecting conflicted specific numbers). The prize is a free market where BT has the opportunity to be one of many transit providers on an equal footing, and the above issues fall away due to the competitive constraint on its market power.

We would advocate that, especially given the parallels with IP routing, Ofcom engage with subject matter experts such as the London Internet Exchange to explore such a solution rather than spending further public money with conflicted market participants. To repeat, Simwood is willing and able to build a solution too, or at least contribute, but a neutral third party would be better. 

To the question of why Simwood does not just do this anyway, we’d say that we have tried but these initiatives struggle to gain traction. Our heads are bruised from the constant banging into the desk in various forums. 

It needs to be an industry initiative and absent a large operator being willing to engage and work with us, Ofcom are logically best placed to lead. We of course have the APIs and will willingly work with other operators who wish to.

Overall we’re optimistic and excited about this forthcoming Narrowband Market Review as it feels like the stars are aligned for us to make significant progress as an industry. Fingers crossed! We will of course be saying more, and indeed releasing more research as necessary as we go along – but for now, we hope that the regulator digs in the right place.