It is still unclear whether Talk Talk is suffering for refusing to pay blackmail and then being painfully honest about the consequences, or whether it is at serious fault for not taking security as seriously as BT – the UK’s largest civilian employer and trainer of information and network security staff. What is certain is that, whatever the impact on its finances, Talk Talk’s ambitions have taken a massive hit . I was, however, struck by some of the more informed comments about what appears to have happened (not an attack on the main systems but on the files used to authenticate those using its customer-facing website) and the advice to those being supposedly contacted by Talk Talk with a request to change their bank details. In effect this can be summarised as: “stop your elderly relatives from using a landline and switch them to VOIP or mobiles so that the more sophisticated scams, including those which use a fake dial tone, no longer have any effect.”
This adds a whole new dimension to debate on a Universal Service Obligation. The proportion of the population which has a landline, other than because it is mandatory for broadband services delivered over Openreach, is falling. It is propped up only by Ofcom requirements to be able to rack emergency calls. If it the landline is now to be viewed as a risk, not an asset ….
The Group CEO of BT told delegates at the Broadband World Forum that “We stand ready to deliver minimum broadband speeds of 5 – 10 Mbps to everyone in the UK“. He questioned whether the current levels of cover would have been achieved had Openreach been made independent when it was created. My own view is that, but for Local Loop Unbundling, BT would have been able to provide 5 – 10 mbps across the UK a decade ago. But is that adequate as a Universal Service Obligation today, let alone by 2020?
But what is meant by a minimum speed of 5 – 10 mbps anyway?
And where does Gfast fit into the panoply of technologies that might be used to deliver fixed and/or mobile broadband that is fit for purpose.
Gfast is a clever technology and is likely to make a lot of money for Alcatel Lucent and Huawei (who are, between them, providing 75% of the equipment for the BT trials). It does, however, require taking fibre ever closer to the premises. Its long term use, beyond, say 2025, is probably to serve small clusters of houses of on the outside wall of blocks of offices or flats that are too small to merit fibre networks of their own, The cost of deploying it as an interim extension of fibre to the cabinet, other than in urban areas, should not be under-estimated.
Depending on the nature of the Universal Service Obligation there are many other technologies, including satellite and terrestrial radio, that could serve all households and almost all above ground parts (what about the London Underground!) of the United Kingdom.
The bigger questions are:
• what should the USO look like?
• who should pay and how?
I would argue that the USO should be based on the quality of service necessary to reliably and securely access and use those public sector services which are intended to be “digital by default”. Performance should then be assessed using tools akin to those provided by Actual Experience (BBFIX) (used earlier this year by Ofcom for some of their analyses of quality of services and recently presented to members of the new All-Party Broadband group).
While it is said only about 40% of delivered quality of service relates to the headline speed of the “circuit”, as opposed to bottlenecks or problems at either end, it does appear that almost all complaints to ISP help desks relate to those which fail to reliably deliver 8 – 10mbps during periods of peak demand: evening, daily, monthly or seasonal. That is also the figure estimated (by various sources) to be needed to reliably use the more complex on-line public services of today. It also happens to be the throughput to which many supposed “superfast” services (i.e. 24 – 30 mbps headline speed) drop when under pressure.
The Universal Service Obligation should therefore be to provide “superfast” services that do not fall below 8 – 10 mbps, whatever the workload or weather. In this context it should be borne in mind that, whether or not they suffer from contention, both fixed and satellite wireless services can degrade in heavy rain and snow, even without lightning strikes on aerials and masts. The actual performance delivered should, in any case, be independently measured and enforcement action taken as necessary.
In order to ensure resilience the obligation should be dual (at least) sourced to suppliers whose networks do not depend on each other – although they should be allowed, perhaps even encouraged) to involve partners in the delivery, provided these do not share single points of vulnerability.
The Talk Talk incident raises, however, the issue of security. Our current telephone system allows numbers to be spoofed as easily as internet addresses and is vulnerable to many shocks, from theft of the copper, to flood, power failure and “inadequately planned excavation” as well as to relatively surveillance. BT is in the process of creating one of the UK’s greatest security assets and should be allowed to get an unregulated return for serving its communications partners, in co-opetition with the teams being built by PWC, KPMG, Deloittes, Airbus and BAe (who are helping Talk Talk find out what happened) or those run globally by players like IBM for its banking customers.
But that is a Universal Service Obligation for today.
By 2020 we are likely to have hundreds of thousands whose lives depend on the reliable functioning of systems remotely monitoring chronic medical conditions while they live sem i-independently at home instead of miserably and expensively blocking hospital beds. Meanwhile central and local government rally will expect most of us to go on line for our dealings with them, because otherwise they will have gone broke. Then we have all those smart buildings and the wet dreams for the Internet of Things that depend on the secure and resilient communications infrastructure that we have not got.