TelecomRedux: So what is the IMS story from IBM’s perspective?

Sandy Aitken: We looked at IMS about two years ago and realised that it represented a fundamental change for the whole telecom industry. We believed that everyone would evolve their networks to IMS, albeit all for different reasons, so we invested a lot of effort into designing an IMS Service Platform which is independent of the control and transport layers. This allows us to look at the IMS infrastructure from a service control and service creation viewpoint. You could call it a super middle ware which sits on top of the control planes from various network equipment providers. We have trialled this with Lucent and Nortel and are in the process of carrying out trials with two other equipment suppliers. We have also developed a business modelling and planning tool as we realised this was essential to help clients understand the business case for IMS.

TR: IMS represents the movement of traditional telecoms towards the IT environment, an area where IBM is a strong player.

SA: We are looking at a lot of these projects at the moment and I still think you need a network equipment provider and you need an IT provider. I think these two components are required to succeed. I think you are right that this is a migration of the industry towards, almost an enterprise IT architecture. However, there are still things within the network which we do need network providers to work on. A lot of discussions we get into are about how much intelligence you have to keep in the network now. And I think there is a lot of debate about whether you have to keep very much data in the  control plane or whether all data should be held in the IT environment and carriers, depending on whether they are approaching it from a network centric or an IT centric environment, taking different standpoints in this space. I think in the next two or three years they will all converge to a more common view.

Another thing which is becoming very clear, which people did not realise initially, is that everyone is going to have to change other parts of their infrastructure such as billing platforms because the new services are now generating real time services which can be provisioned, employed and used in real time. Current billing platforms can’t cope with these real time services without significant alteration or replacement.

TR: IMS is about more than technology isn’t it?

SA: IMS has gone from being a down the road, sometime in the future solution to being here and now in a very short period of time. This is forcing a change of thinking from an operator’s standpoint, it is not just about IP and bits and bytes, it is actually a whole different way of thinking about how you run your business. And it requires a radical rethink about organisation design. Rethinking business models provides the opportunity for you to do things differently and actually to start making money through these future services. Although I suspect some of them will do very well, I also suspect an awful lot of them will fail to capitalise on what is happening.

TR: Operators, particularly mobile operators, are very conservative in their thinking. There is a tendency to think of IMS as being a magic wand that will cut costs, increase revenues and make the networks easier to run. But IMS is not a magic fix, operators have got to do something for themselves as well, not just install IMS and wait for the money to roll in.

SA: Well, I think you are right there. We have proved that the technology works. There are still some standard issues of when everything will work together in the scale of things expected but these are being solved on a weekly if not daily basis, but then you have some interoperability issues when you are testing between operators. We are currently demonstrating IMS services running across two different mobile operators from two different countries successfully. We really are replaying the move from analogue to digital in the 1980s and 1990s so we are going to have all the same problems people had moving from analogue to digital when they only had limited devices and services at the start. However that does not mean that it won’t happen, we just have to go through some of the pain which is what we went through that time as well. However, IMS won’t make any money on its own unless we think differently about what we are doing and trying to achieve. This is a transformation of the industry.

One issue which I am not sure is being addressed properly is market segmentation. This is an opportunity to actually have some incredibly high value services for your high value customers which they pay for and give them high spec devices on which to use these services.

IMS allow you to have an incredible array of services across a wide range of environments with relative ease and I don’t think many have thought through what that could do for their business. They could offer games, to people who play games and offer them devices that enable them to play games with integrated voice running in the background as a standard service but you don’t need to offer that to everybody. Many customers will only want a basic services, the flexibili8ty offered by IMS is one of the keys to making this a success. Combinational services are what IMS can offer and these can have a significant impact on revenues and costs. You can use the same platforms which provide IMS multi-user games with voice to provide multi-user enterprise services etc. The key here is to provide common capabilities which can be reused over and over again in many different ways, to simplify the service build and improve time to market. This can allow us for the first time to get to the segment of 1.

TR: But historically of course, the creation of new services was expensive and time consuming and I think they haven't realised and I don’t think anyone will realise until they see it working in the real world that IMS gives them this flexibility to rapidly create and deploy new services as you say, very cost effectively, which means that they can produce a lot more services, a wider range of services which they have not been able to do in the past, and that brings them to your very point of segmentation. This has been very much a one size fits all market hasn't it?

SA: Absolutely. In fact we actually showed one operator that we could design and deploy new services within three weeks. They have never had to deal with an environment like that currently they launch new services twice a year. There is no real reason why we can’t do it within a week or a day because we are using a common capability framework. We call it a common capability portal now as I think that is where things will go and it is also a description which people understand.

We are saying to operators, "you do know you will have to have a browser in this and you will have to treat it like an internet style portal which you search from because if you don’t people will still recreate more capabilities and they will no longer be common so you will back where you are now if you are not careful. You will a massive number of potential new services based out of maybe 10,000 common ones and what you want to do is try and keep to the lowest number of common capabilities you need to recreate rapidly new services. I also believe that you should also be opening up gateways to allow people with much more imagination in certain fields to develop their own services.

TR: Well, that is going to happen if you are talking about a standardised solution with standardised APIs, but it is not the problem of opening up your network, it’s not technological, it’s cultural. Operators - particularly mobile operators, think that anybody who plugs into their network is going to make it fall over.

SA: Exactly. And another issue that operators should be facing, if they go down the track we are talking about, is how do you charge for services. You can either charge a bundle or you can be quite imaginative and say, well people who want location will pay slightly more. People might want location plus presence. Its not A + B = C, it might be A + B = B1/2. You've got an enormous ability to manipulate your charging to become unique. Instead of what we have at the moment what you have really got are one or two services with hundreds of different tariffs, you can have lots of different combinations all of which add up to a different number. and you can get them to pay by the drink, you can get big multinationals where we all get the same service. Why should we all have the same? Maybe the sales guys get one type of service, maybe the sales directors get a different type of service, maybe the guys on the road doing the actual sales get the best service because actually they really need some sort of sophisticated comms platforms. The constraint here is imagination which is always going to constrain the way we engineer our businesses. However, customers have a greater understanding of their own requirements and hence are much better with some help to design them. In these environments you will get quite sophisticated users who really understand what they need and what they want, who will select from a common capability portal the pieces they will pay for, will improve their lifestyle or make them more efficient and build them into their own services. This will have a significant impact on the relationship a telecoms company has with its customers and this is equally true for enterprises and consumers. You will have other groups of customers who as usual just want to pick up what is provided but the real impact of IMS is the flexibility to cover both types of customer for little or no added cost.

TR: you have been focusing on business models and business strategies and re-engineering the businesses because I think that this is the way that you have to go because I think the technology by and large is done and dusted. It is all 3GPP compliant and everybody's signed up for it and everybody is doing the operability testing and that should all work but it is your business models, the way you run your business is – that has got to change to respond to this technology and this is the area I think where the biggest problems are going to occur.

SA: And that is exactly what we found. People need to recognise that there will be an organisation in the future which manages the future infrastructure and it will not be like the IT or network organisation of today. The skills required for products and service definition and deployment will have to be much further to the front end of the business if you are going to rapidly develop new services, launch them and also retire them. One of the other big issues is how do you retire services? This is where a lot of the cost currently exists. So there has to be this rapid shift in how organisations currently work and a significant change in roles and responsibilities of different groups within the operator. Actually you want to make product development and deployment something which happens at the front end of the business so the sales marketing organisation with some IT support you can do quickly model and utilise the new development capabilities within a framework which ensures you don’t screw the network up. The challenge then is, how do you merge two organisations, which for the last hundred years have actually sat in four different buildings? You have all the classic things of cultural alignment, process redesign and these, you are absolutely correct, are not technological issues.

TR: So where do operators go now on the road to IMS?

SA: The architecture we have been proposing to a number of major mobile and fixed operators is actually a derivation of the whole IMS architecture, and we have said to them, - "You have to start thinking now about extracting the customer and product and service databases out of the applications they currently reside in as well as creating the common capabilities, service creation, service execution and orchestration platforms. Because until you do that you are never going to break loose from the inflexible services which are currently built. To do this you need a high performance enterprise service bus which is the IMS service plane. If you have a service bus you can just layer on top of it more sophisticated processes which can handle the volumes, at the bottom level you can have very simple stuff, which just has to be quick, and on the top layer you can have much more complex services orchestration.

So you should end up with a scenario where you can design, develop and deploy once, globally. Our clients can absolutely see why they should do this and why this makes logical sense because it does away with a lot of the problems they have currently got. But this is a five year programme. You cannot do this overnight, starting to extract your customer database from a legacy billing platform is no mean feat, similarly to try and extract your product database - the first thing they all want to do is replicate everything in the product database to the new one. Why would you do that? You can't go from 30 years of legacy, value added service and then recreate all of them, you have to start making commercial decisions about what ones you do want to recreate. And how do you recreate them in a way that is replicable in terms of common capability? How do you turn off the other ones? How do you manage that process? And all of these things take time. Data migration projects are physically big projects. That is one of the things I think a number of operators are finding daunting.

We've come up with a number of different scenarios and are discussing with operators whether they have considered the different business models. You don’t have to have a one tier biller in the future, this has caused huge ructions, killing off a sacred cow. I think anything that kills off sacred cows has got to be a good thing to consider.

Operators are all looking at trials of IMS components, the business cases and many starting implementing parts of the architectures. What is important for success is making sure the different programs are co-ordinated and under one clear program governance model.

TR: Thank you.
Ian Channing