BT Tries To Fix Global Services with Open Source
BT has frequently been in the news here; innovation chief Matt Bross presented about his aim to change the company from one defined by infrastructure to one defined by software at the November 2008 Telco 2.0 event. More recently, its systems-integration and IT consulting wing, BT Global Services, has run into trouble, notably in its involvement with the UK’s controversial giant healthcare IT programme.
A software-defined telco - like a software-defined radio - sounds a great idea. But how will BT execute on it - and how will they stabilise Global Services, which is surely a crucial element in such a strategy? It’s an important question for anyone trying to implement Telco 2.0. It seems that BT is hoping that it can achieve this by embracing open source software (OSS) and the habits and methods that go with it.
An anonymous delegate at the last Telco 2.0 event sent us this Mindshare feedback message:
We need more outside the ecosystem players like Apple coming in to cross pollinate with our gene pool. My guess is that apple doesn’t attend Telco events because they are worried about damaging their own gene pool with our status quo. Give some kids full artistic license at a reference acct/operator to build their playground. Lock them in a room with caffeine and pizza and a big pipe. Carriers have great toys they would like to build with. Output = Telco 2.0
As it happened, we had the opportunity to see BT’s efforts to do just that recently. As well as its well-known investment in Voice 2.0 start-up Ribbit, and its previous Web21C API suite, BT’s efforts to initiate transformation towards Telco 2.0 saw it acquire a small UK software house called Osmosoft, which specialised in open-source development. They’re now installed in the spooky and very, very Bellheaded confines of the Westminster ATE - you know you’re in a telco when half the building is devoted to Ministry of Defence networks.
But what are they doing in there?
The short answer is that they are trying to infect the rest of BT.
J. P. Rangaswami, old friend of Telco 2.0 and MD of BT Design, the cross-cutting engineering function of BT, described the main purpose of Osmosoft as being to inject open source culture and methods into the BT mothership. He’s selected a dozen people as “evangelists” - oddly enough, the same number as the original evangelists. One hopes no-one will have to be crucified. He remarked that BT doesn’t want to be a consumer of open source, but wants to be a contributor, both to support the movement and also to influence the direction of the projects it uses. Further, once you’ve recruited open-source developers, they need to keep contributing to maintain their skills and stay current with progress in the broader community. Open partnering, he said, is vital. (You might almost think ex-Telco 2.0 people were involved).
So, Osmosoft, and to a lesser extent Ribbit’s, real role is threefold - as a brain farm for recruiting talent that wouldn’t necessarily fit into a traditional telco culture like BT, as a terrorist cell spreading subversive ideas inside BT, and as an exercise in costly signalling. Biologists who study signalling in nature have long theorised that the reason why so many natural signals are disadvantageous to the creatures that produce them is that signals have to represent a real investment to be credible, as otherwise they would be universally faked. Osmosoft and Ribbit cost money - so they are a credible demonstration by top management that innovation is not a crime.
Jeremy Ruston, the head of Osmosoft, said that the decision to buy it boils down to the fact that BT knew there was innovation in open-source software and needed help to understand it. He was convinced by the fact that they didn’t mention lower costs; but did mention community. You can’t say anything sensible about OSS without talking about community. This helped overcome his suspicion about moving to a company where the finance department turned down the funds for the deal twice.
BT, he said, seems like a hostile place for OSS; but as soon as he got there he found it was under every stone.
This is an important point for Telco 2.0 implementers - open innovation can come in through the back door as malware and semiofficial, undocumented work-arounds, or it can come in the front door as solutions. But it will get you, one way or the other; so it’s best to accept it on your terms. And there is often a surprising amount of innovation going on from below that might flourish in a more welcoming culture. BT engineers frequently used their own home-made or OSS solutions although something expensive in the way of proprietary software had been purchased - management knew nothing of this until the creation of an OSS business unit helped to make it visible.
A major strength of OSS is that it is a low friction way to solve problems in the enterprise. However, management suffered a widespread belief that OSS was incompatible, illegal, or fattening. BT being an engineering-led culture, full of experts, they felt they needed to have something that they could touch and feel. They decided to use Osmosoft’s TiddlyWiki as an example. This is a set of collaboration tools based on the wiki paradigm, but implemented as a single file and a back-end server that permits multiple users to synchronise their versions of the document as they add to it.
In Osmosoft’s own right, they’ve integrated it with the conference resources website Confabb. As a BT project, it’s being used to visualise and document BT infrastructure. J. P. Rangawami remarked that “as we solve problems with OSS, we find other problems we could solve. There’s much more richness and value in doing it this way as opposed to buying in monolithic systems.”
BT Wholesale’s Web site includes 4000 pages and 10,000 documents - not surprisingly, feedback from customers suggested that distributing documents and collaborating on them was a serious problem. Tracking changes to hundreds of pages of detailed and legally sensitive tariffs is hard. So they introduced tiddlydocs, based on tiddlywiki, to help document it all. Wholesale customers are now using this for their information and also to help sell to their downstream customers. Sky Broadband is the first to trial it.
That sounds fun, but we couldn’t help feeling that it wasn’t particularly telco-specific. However, it’s worth pointing out that some of the use-cases we expect to pay off soonest are ones which are in the carriers’ own businesses. For example. BT is now using Twitter to get real-time feedback from its customers. But Twitter is not very different from SMS; why the fuss?
“Social networks have certain common features - a directory, relationships, communication, scheduling, and a record of changes, which is what the status-updating function is,” Rangaswami said. “Telcos, in the past, had the address book, especially when they were Post and Telegraphs as well as just Telephone. What Google, Microsoft, Yahoo!, and Facebook have done is add the scheduling and the record of changes, which no telco has ever done.”
But it’s not really connected to the infrastructure, it doesn’t touch on voice or VAS. BT has, of course, been trying to do better on these for several years, first with Web21C and then with Ribbit. As J. P. Rangaswami says, the decision to do this was taken as long ago as 2006. But Web21C didn’t take off. He has an interesting take on why.
“Communities don’t do what you expect, though,” he said. A lot of early usage of Web21C’s call control API was simply “virtual calling cards - arbitraging between the voice tariff and the service we made available through Web21C.” It’s a classic Telco 2.0 moment; the primacy of voice demonstrated. People will do almost anything for cheaper calls; the flipside of this is that voice is the only form of interpersonal electronic communication that the public are willing to pay for.
Rangaswami also argues that the Web21C project overstated the relative importance of server-side solutions as opposed to client-side ones. As a result, the decision was taken to revise the business model, accelerate opening the network, and acquire Ribbit and its team of Adobe Flash developers.
Developers are offered a choice of routes to market and business models - you can choose to “pay per drink”, or “eat all you can within limits” - and a range of levels of engagement. “Have you got your own termination, billing, data centre? We can accommodate users with all of these, none of these, or any combination,” says Rangaswami.
He argued that a major role for Osmosoft, and for OSS in general at BT, was recruiting; “It’s a talent war - we’re primarily concerned in acquiring talent. Most businesses aren’t expanding, but we’re always trying to gain new capabilities. Osmosoft is our most effective acquisition tool; much better than M&A.” Jeremy Ruston said that the hires we’ve made from the community are much cheaper - you can see what they’ve done in terms of actual code. Also, as one of the developers pointed out, OSS means that you don’t have to abandon your work on a project when you change jobs.
However, open source is becoming deeply embedded in the infrastructure. According to J.P. Rangaswami, “I can’t think of any BT product that doesn’t include Linux in some way.” Rangaswami argues that there is an interesting dynamic at work, which is likely to transform the IT industry if it keeps up. “If a problem is general, you should go to the OSS community; if it is specific, you should go to commercial, because someone will have the commercial imperative to solve it. This is why OSS moves steadily up the generic IT stack - as soon as it becomes necessary to scale up and provide for diverse user needs, it becomes more efficient to become an open source community.”
Telco 2.0 couldn’t agree more. As Thomas Howe, CEO of Jaduka, said at this spring’s Telco 2.0 event, there are a few basic capabilities in the Voice 2.0 field which are applied to tens of thousands of highly specific business processes. Value is created at the touch-point between the two, which requires very specific knowledge of the target business process. This is why the upstream customers are important!
In the end, he pointed out, it’s the cost of change that becomes ludicrous, not the cost of acquisition. In India, he said, he could ask someone to come to dinner and his mother would fix it - the cost of adding some more vegetables was trivial; it’s different in the UK where you have lamb chops and count out the potatoes. “There is a lot of OSS within our stack, but we want it in our culture. We use software to deliver services - when it gets to the enterprise customers it’s a service. The focus is on moving their processes into our systems.”
Doc Searles described this as the “because” effect. Google makes money because of Linux, not with it; “this transition from the economics of scarcity to those of abundance,” Rangaswami said, “is necessarily destructive of value to at least one actor but creative of it to many others.” Similarly, the migration from Telco 1.0 is painful to the old voice monopoly, but it’s hugely beneficial to the wider economy, which is much more money. The clever bit is to make sure a thin slice of the productivity gain is captured for the telco.
So, Rangaswami was asked, is open source going to save BT Global Services? “Platforms, reuse, and collaboration are going to save BTGS; and OSS underlies that,” he said.