What do you need to build a telecoms company? Yes, money. A lot of it. Yes, telecoms expertise. A lot of that, too. And, yes, a marketing plan, so that you can actually win customers.
But until recently, you needed something else as well, especially if you planned to build a fixed-line operator: what we in the UK call a telephone exchange, but what speakers of American English call a central office.
That element in the infrastructure of a telecoms network is about to disappear from the scene, along with all the horses and carts, steam railways and arc lamps. To be replaced, of course, by the cloud – mimicking what has happened in enterprise IT over the past few years.
Both terms, “telephone exchange” and “central office”, accurately explain the function. A central office is usually in the middle of town, where the wires meet from all the different customers that your marketing people have cleverly persuaded to sign up.
The term “exchange” tells you what actually goes on. Pretty much like a stock exchange, a coal exchange or a wool exchange brings together members of a community with a common interest, a telephone exchange brings together people who might want to talk to one another.
How was it done, from the earliest days of Alexander Graham Bell until the near future? Simply by connecting the wire from one customer to the wire of another, initially via a voice instruction, but later via a mechanism controlled by the person making the call (the story of Almon B Strowger, the Kansas City undertaker who invented the first automatic telephone exchange, is fascinating).
I have a wonderful book on my shelves, published in the English version in 1890, called Electricity in the Service of Man. In the original German, by Alfred Ritter Von Urbanitzky, it had the less sexist name Die Elektricität im Dienste der Menschheit.
The frontispiece has a picture, made as an engraving on a steel sheet, of the central telephone office in the avenue de l’Opéra in Paris, showing 18 women mainly standing at a switchboard, ready to connect customers’ calls. There’s one man in the background, arms folded. Whoever said the telecoms industry discriminated against women? (Sorry, that’s ironic.)
But that’s how it was done: thick copper cables snaking across town into the exchange where Mr Bell would demand of the woman operator to be connected to Mr Watson.
The IT industry used to be there, too. Big installations of mainframes, mostly bearing the IBM label, in the basement of every company, whether it be a bank, an insurance company or a supermarket chain.
Then the revolution happened. And it was a revolution, too. From the late 1970s personal computers appeared in the shops and software developers produced applications that people wanted to use, but couldn’t get from their IT team: spreadsheets, for example, with VisiCalc, SuperCalc and Lotus 1-2-3. Word processing software, such as WordStar, which took document production away from specialists called typists and removed for ever the embarrassment of coming back again and again with revisions.
And what happened? Executives with a corporate credit card and enough seniority went out to the early computer shops and bought a desktop machine and the software, without telling the IT people.
Parallels are never precise, but for years the telecoms industry has felt captured by the equipment vendors, especially as they went through a series of mergers. (A similar thing happened with computers. Where now are the mainframe makers of yesteryear?)
The answer, it now appears, is pretty similar. First, enterprises moved their IT needs to off-the-shelf systems, often designed by companies such as IBM, which had transformed itself into a systems integrator. And then, over the past few years, it all moved to the cloud providers, with laptops connected via broadband.
Incidentally, that’s why we are still in business. Even a few years ago, every email sent to and from the company that owns Capacity went through a bunch of PCs in, yes, the basement. One day the servers died. Emails were lost. We all resorted to our Gmail and Hotmail, previously banned, until service was restored. Now, we all work through Outlook – as well as Word and the rest of Office – on Microsoft’s cloud service: it works, which is what you want.
Probably the pioneer in moving telecoms to what we would now think of as a cloud-based service was Skype, created in 2003 by a group of Nordic developers, originally just for voice calls. There was no infrastructure as such: calls were handled by a system distributed across the users’ own personal devices. I simplify hugely, of course, but there was no Skype exchange or central office: a nightmare, incidentally, for law enforcement people and the spooks, who had nowhere to connect their crocodile clips.
But the message has spread. A data centre with the appropriate software is what you need to run a cloud-based telecoms system. Yes, you still need to connect the users, via copper if you must, or fibre, or 4G or 5G waves, but you don’t need something in every district of every city that you can label “telephone exchange”. And we have the infrastructure that connects homes and offices to the system – it’s the internet.
This year, we have seen start-ups such as Japan’s Rakuten Mobile launch services in the cloud. It’s even buying a software company, Altiostar, for $1 billion, so it can spread the word. One of the adopters is Dish Network, now being built up as a rival to the three big mobile operators in the US – AT&T, T-Mobile US and Verizon.
It’s harder for companies with existing, traditional infrastructure. But note how AT&T itself agreed a deal in late June to sell the network on which it already runs its 186 million mobile customers to Microsoft.
This is not an exclusive AT&T-Microsoft deal. The company behind Windows, Excel and Word will be using what it has acquired from AT&T to sell cloud telecoms to other operators around the world. It will be part of Microsoft’s cloud offer, as Azure for operators. Everything is going into the cloud.
