Interoperability? Meh!

 

By Simon Woodhead

We’re noticing a worrying trend in UK telecoms that is evidenced by a rise in what we call tickets about ‘weird shit’.

The days of a tax-payer funded monopoly are behind us – no, honestly they are, you and I just misunderstand articles like this – but in some way I miss them. That is surprising when I spend a good proportion of my time dealing with attitudes and behaviours that haven’t moved on, but there was one really good aspect to it. Back then you could pick up the phone, dial a number, and get consistent behaviour.

Back then, we had a state monopoly in BT and a single competitive operator in Cable & Wireless. Kingston Communications were there too for those who lived in Hull! The majority of call traffic would be on-net and where those operators did interconnect, they did so under standards that were so strict as to almost be a barrier to entry. BT had their own special variation of ISUP (the layer of the SS7 stack relevant here) but that was generally in use for all SS7 interconnects in the UK. Things worked and the consumer had consistency; they just paid a small fortune for it.

Enter de-regulation, the creation of Oftel (now Ofcom), and the market was opened up to new entrants, including later on upstarts like us. But, SS7 remained the primary means of interconnect and UK ISUP was the agreed standard. In fact, SS7 remains the only protocol in a Regulated Interconnect, but that is a separate topic we’ve mentioned before. Furthermore, TDM networks were big, expensive and hard to operate so tended to be centrally managed.

Enter SIP which, like SS7 and pretty much all internet protocols, is an agreed standard defined in multiple RFCs. Like SS7 it is a standard that presents many many options and variables to ensure maximum compatibility. However, we lack an agreed implementation standard, the equivalent of UK ISUP.

To add to the complexity, any SIP platform that 100% adheres to the RFCs simply will not work in the real world, because equipment vendors and developers have interpreted things slightly differently and have a tendency to say one thing in the signalling, for example, and do something different in the media. Curiously similar to some of our competitors!

Those big operators with expensive TDM networks discovered the allure of magic boxes; or may even have been invited to the Secret Club. They strapped magic boxes around the edge of their network. It was really easy to order a magic box and have some helpful consultancy to install it and tick whatever options that particular engineer thought helpful.

But it was ok, because those boxes were around the edge of a TDM network which was itself conforming to UK ISUP or, even if via the Secret Club, TDM was somewhere in the path of that call. It acted as a mediator to an agreed standard and all was still good.

Fast forward a few years and we’re now reaching a tipping point. It is a tipping point where a reasonable proportion of traffic is transited SIP end-to-end, via a series of magic boxes. Peculiarities of our broken porting system, the decentralisation of control, and the ability for different silos within the same organisation to have its own magic boxes facing other people mean there’s sometimes craziness in the routing. That craziness is transparent when everything works but when it doesn’t, it is quite unreal.

Being able to call selected carrier C’s numbers over every interconnect except the direct interconnect with carrier C might be one example. Another might be being unable to call selected carrier V’s numbers over carrier V because they hairpin them through carrier B over IP.

A ‘failed’ call in signalling is easy to deal with, but often the issue is much more subtle – e.g. no DTMF, or media issues. In a growing number of cases, a combination of arrogance and incompetence has led to some magic boxes being installed with the option ticked to play a Chinglish message. “Sorry, the subscriber you calling is power off” is not the message you’d expect your customers to get when placing a call through Simwood, and certainly not something we’d expect when passing calls directly to a multi-national carrier for completion on their own network.

We won’t tolerate that and have terminated traffic to multi-billion pound operators because of it, but it happens. Add in mobile operators exposing withheld numbers due to half-hearted, inconsistent and often confusing implementations of GC C6 which comply with one version of guidance which is subsequently contradicted by another, and we’re in crazy times.

Of course, none of the above analyses have come in a rapid response to a support ticket as Simwood customers expect from us. Filling in multi-page Word files to send to an Indian NOC, does not yield a rapid or coherent explanation of what is misconfigured on various magic boxes. By definition, they’re magic. Instead they come from exhaustive testing by our team here, and quite often the level of skill in driving a magic box being so abysmal that the operator concerned has shared traces demonstrating their entire supply chain.

I’d say welcome to the Wild West, but that would be unfair to the USA. We have none of these issues over there. But then again they have functional porting and the FCC is competent. If only we had a regulator over here capable of dealing with it, but they seem distracted playing with blockchain with their friendly incumbent.

Whilst it is tempting to end with an apt quote, there is hope.

Now Simwood is a carrier in the USA, we’re swimming in a much bigger pond. Bigger ponds have bigger fish and it is laughable to see how our behemoths rank on the world stage, or don’t. We’re a small fish there too, but our presence is growing, and we’re interconnecting with multiple operators over there and dramatically improving the quality of our voice routing as a result. The USA has regional interests different to the UK, e.g. South America is both local and significant to the populace, but they also have a lot more volume. I’m not sure if they’re behind in implementing magic boxes, or in front in realising the pitfalls, but one way or another we’re managing to put in place routing options that are just not available in Europe.