Interview: TM Forum's Chairman, Keith Willetts
It's been trying to simplify the interactions between telcos for years by standardising their OSS-BSS interfaces; if the Telco 2.0 vision comes to pass, what with all those third-party innovators, payments going upstream and downstream, and user data flowing through dozens of APIs, we're really going to need answers to a whole world of questions around OSS-BSS, roaming, interconnection, and settlement. After all, arbitrarily large numbers of small services (either on their own or as part of bigger organisations) will be busy interacting with telco networks, and the complex of telco, third party developer and customers will be interacting with other telcos, other systems, and other businesses; we'd better have these things nailed down.
As a prelude to their important Management World event on 18-22 May in Nice, we spoke to Keith Willetts, their Chairman (and one of the nicest men in telecoms), about convergence, trends in OSS/BSS and data mining - critical issues, as was re-iterated at the CMO Forum at the Mobile World Congress:
Telco 2.0: Convergence is a popular buzzword, but is it more likely that the industry will diverge - become a much more diverse variety of products using a common IP network?
Keith Willetts: Convergence doesn't mean "everyone doing the same thing"; it means a new value chain. Most operators do two things - transport, and an end-to-end service. The future won't be the same; instead, the end-to-end service will be supplied by entirely different players or created from several different products.
Instead of a world in which the operator controls everything it needs, it'll be about gluing together many different things.
Telco 2.0: So we're looking at a standard for exchanging commercial information - a really big XML struct?
KW: We've got to define a standard trading and management interface between companies. Otherwise, it won't work for thousands of participants if you have to get on the phone and talk.
Telco 2.0: Rather like GSM roaming, then.
KW: But for many other services. The challenge is integrating it with the SDP. Once, you knew what the service was, telephony, it came out of the network, and you applied OSS-BSS to administer it and bill it. Now, it comes out of a server, and there will be many more services.
As a result, you'll no longer be able to see the join; some functions are OSS, some are BSS, some belong to other parts of the SOA, some are charging. You won't be able to walk up and plant a sticker on the BSS. Convergence at the software level means the join between these elements is invisible. I expect there will be fewer small ISVs.
Telco 2.0: How can we price complex information based on multiple telco data sources, the sort of thing we'll need for advanced mobile advertising? Will we have to move to a more Amazon-like pricing structure; a flat fee per event?
KW: Provided you can measure it you can bill it; the real problem is responsibility. There will be many parties to every transaction - how do you settle responsibility instead of profit? We need to understand this in advance; what happens if someone downloads a movie and the file is corrupted? Does the telco give them a refund - and if so, does the content provider bill them for another copy? Imagine some sort of tied-in marketing that goes wrong.
The trading relationships dwarf the complexity of actually counting the packets. For things like mobile advertising, the basic stuff is all there in the OSS - calling circles, location, etc - but it's in a gazillion different formats Monetising it upstream is dwarfed by the difficulty of getting it out; in the short term it'll have to be flat-rate. A service provider like O2, for example, has several thousand OSS-BSS instances with information in different formats.
Telco 2.0: How can we make progress towards solving this?
KW: Iron filings on a sheet of paper; you've got to provide a commercial magnet to make the operators go back in there and revalue their data. I was recently in Russia with MegaFon, who are quite good; they've configured things to autopopulate the databases. But others are far worse in terms of fragmentation."
We need some sort of agreement on what data you can supply, and a "service indication" protocol. We also need a OneWorld Alliance of operators to work on a much bigger scale; a lot of operators can offer you 20% of the market, but Google can offer you the world, so we need to start working on a cross-continental scale. None of this will occur until there's real money on the table. I give it 3-5 years to the first green shoots of new thinking.
Telco 2.0: Will the solutions be more of a Web services model, with external applications pulling data from the telco systems and processing it themselves, or a software-agent model where external developers specify logic to be executed within the telco systems that returns an answer? The first is much simpler to implement and secure, but the second may be needed to solve the privacy issues...
KW: It's a huge issue; who owns this information? It won't be long before there's a major lawsuit; I signed up to let this service use my data, and they shared it with my wife! We certainly need a detailed opt-in process. There are other questions as well - if you ask yourself what the top five pieces of information advertisers would want about a user, you realise that you can't get age from the network. The big unanswered question is - what do the advertisers want?
If you put users, operators, developers, vendors and advertisers in a room, you'd probably end up with some very divergent sets of answers. I guess that's what the Telco 2.0 events are all about?!.
[Ed - Telco 2.0 will be presenting at and facilitating the Leadership Summit at Management World on 19 May. Details here.]