All-time traffic records and BTZero


By Simon Woodhead

It was 11 and a bit months ago the UK went into the first COVID-19 lockdown. We saw traffic drop off a cliff from April 2020, having hit an all-time high in March 2020. Whilst the next couple of months are going to be interesting against 2020 comparatives, it seems opportune to look back to February 2020 – arguably the last ‘normal’ month.

We saw UK wholesale voice traffic bottom in May and by July, thankfully, we were back into growth territory. I’m delighted to say that the trend has continued and February 2021 shows a pleasing 34% growth over February 2020. That isn’t quite the 53% rate of year-on-year growth we were seeing in pre-COVID, but in absolutes, it builds upon it. That’s a pretty tough comparison to beat so I’m really pleased.

I say this not to gloat, as the last year has been tough for everyone, and exceptionally tough for those in industries that haven’t been able to carry on. We’re very lucky to be in this business, but have seen mixed performance across customers, depending on their own sector exposure. Previously our traffic seemed to serve as an indicator for the wider economy so I’d like to celebrate what this could indicate for all of us.

Where I’m less humble is about our progress with BTZero! 

Regular readers will know our frustrations with the regulatory regime in the UK, especially around number portability, which gives the former monopolist the ability to earn super-normal profits from customers it has long lost. Those profits are, in the main, not price-controlled; one of many wrongs we sought to correct in our response to Ofcom’s Wholesale Voice Markets Review 2021 consultation.

BTZero is our effort to expand our voice-peering such that traffic between Simwood and others flows directly over IP. This improves security, resilience and quality because calls avoid that metaphorical long dark alley that is BT – always recommended but nowadays you just know you’re going to get mugged! 

We still have more work to do for other operators to route calls to ported numbers directly, especially where they are third-party ported, e.g. a major network calling a BT number that happens to be ported to Simwood now will invariably transit BT.  We expose those numbers that are on-net via an API which other operators can consume to know what numbers are on-net here. Given the relatively small group of operators that are relevant, we’d willingly consume theirs too if they were to have one. We think it is preferable to the official Yahoo Google Group but is still imperfect. One issue is them having the confidence that we haven’t just made something up to man-in-the-middle a third-party number, but the other one is economic. 

Yes, they may get mugged for transit or additional conveyance fees by sending a geographic call to the range holder but some of the mugging is also on us as the recipient of the call – we’re substantially a net gaining provider. (Not sure if I’ve mentioned lately that we lose money on every minute of every geographic call ported to or from BT and have since 2014?!). The answer here, of course, is a system like developed countries have, such that everyone could avoid that dark alley unless there was genuine commercial value to them venturing into the blackness. 

One could argue that loan sharks add genuine value to some people and at some level are necessary in an economy; indeed, legitimised ‘loan sharks’ such as the pay-day loan companies ran that very argument. Where it crosses into ‘completely wrong’ territory is when the regulatory environment is so derelict that a little old granny can only get some of her pension from said loan shark, after deduction of an uncontrolled handling fee of course, because there is no alternative. Alas I digress! 

So back to BTZero progress… I’m immensely pleased to say that the volume of traffic coming into the Simwood network directly from voice-peers is now around 40%; only 60% comes over BT. Considering, we estimate, 70% of calls have a BT number involved, arguably we’ve over-performed by 10% here. We think we know where the distortive flows come from and it relates to us having high volume users using Simwood or Simwood-hosted numbers . 

We’re at the point where each new direct interconnect will be provisioned for incrementally smaller volumes, as the counterparties get smaller and smaller in scale – which is economically challenging, or we need change to the porting regime. The former is a very good reason to celebrate; the latter, alas, I fear hell will freeze over before Ofcom grasp the nettle.