How should Operators integrate third parties into their network

For the second time in two weeks, I found myselves ‘on stage’ at a major conference and as part of an unplanned event in additional to my original event.

I was just getting down to enjoying IMS 2007 after my panel on Mobile Web 2.0 and IMS and my role as a judge in the IMS developer shootout, when Mark Newman(Informa global head of telecoms and media research and chief research officer) asked me if I wanted to chair an additional panel.

I have known Mark for some time now from my Korea days – a great guy, very clued on and also has long supported me and my work.

When he told me the topic ‘How should Operators integrate third parties into their network’ – I could not refuse! I have been historically ‘pro developer’ and it was great to question two Operators (Mobilkom Austria and Telecom Italia Mobile ) and also Microsoft about developers.

Our panel comprised of Mobilkon Austria head of services/innovation Reinhard Wilfinger, Mario Bonnet from Telecom Italia Labs and Sam Christie(Microsoft IMS – but also ex chief architect Nortel).

Firstly, you have to understand my perspective – I believe in co-operation and with various parts of the ecosystem working together by understanding each other’s view points. Thus, while it would be easy to create controversy – and believe me, it does get you some coverage if you end up with a panel ‘throwing chairs’ at each other – in my view, that’s the wrong way to go.

For instance, for all my talk of opening up the Walled gardens, I presented a contra point of view to Tim Wu’s article on net neutrality – and Tim Wu himself commented on it with my document being included in subsequent versions of Tim’s paper.

So, my overall goal was not go about criticising the Operator – but rather on ‘what can we learn from each other’ and ‘what would the most useful things for developers to know’.

To clarify my personal bias, my view is: in an IP world, as the Mobile Internet mirrors the Internet, the Operator should focus on the core of the network and leave the edge of the network to third parties. Specifically, this means – identify the elements that can be performed ONLY in the core and then abstract them through APIs. This approach gets us away from the dichotomy of the ‘pipe’ vs. ‘no pipe’. It also means that the Operator retains control.

So, my questions were

1) As a Gedankenxperiment (I love that word as you can see if you read my blog) if a developer knocked at an Operator’s door, what would be the advice from the Operator?

2) How is the Operator enabling Long tail applications?

3) What are the key success stories of working with third parties?

4) In a world, where companies are now fighting for developer support(think beyond telecoms here to also include Google et al but also Nokia) – why should the developer go to the Operator in the first place?

5) And the final question was: What specific elements does the Operator see themselves abstracting through APIs(if any)?

My inital observation was: there are not enough ‘Pony tails’ and women in the audience.

In other words, we are still an incestuous community – mainly Telecoms engineers – comfortable with an accepted view point. Although that is changing – too rapidly for some and too slowly for others – the room still reflected people who are at consensus with each other.

Here are the interesting insights that came out from the panel

a) TIM is doing some very interesting work on abstracting the network. This is a topic for a separate blog. I am always on the lookout for Operators who are changing and TIM seems to be one of them

b) Reinhard mentioned that in one case, they went live in about one week with a developer. That’s very interesting (even if with caveats)

c) Widgets are the way to go to hit long tail applications(again Reihard). Spot on! I could not have put it better myselves – and I have been saying that for more than a year now

In a nutshell

As more mobile devices support the full web(CSS, JavaScript etc etc), they support Widgets(through Mobile Ajax)

Widgets span the Web and the Mobile Web – thereby using the same code base.

They are simple and monolithic.

The data source can be on the Web(same as for the Web widget) and RSS could be used to feed the same widgets(web and mobile)

Widgets can be distributed over the Web – so you are no longer confined to being on portal (or not as the case more often is)

d) Mario and Reinhard – partners are not a single dimension i.e. some are closer others are not. Thus, Microsoft is an important partner but Microsoft is also a partner through which they can accept a smaller company. Actually, it may be a better idea for a smaller company to go through a larger partner – specially if the application hits the network elements.

e) Mario and Reinhard again .. What’s the business case? Everyone wants 100,000 euro. But what’s the upside for me? More importantly, what’s the downside(risk) for me?

f) Mario and Reinhard yet again .. Most people don’t consider the impact of customer service. That’s a huge cost to us and a major factor in rejection

g) Sam – Long tail and customer support don’t go together as in the case of other (non long tail) apps. It’s a matter of better options(web) and customer education

h) Sam – (and this is brilliant!) – A list of elements that could be potentially abstracted .. Location, billing(micro payments), Messaging, Call placement and control, Presence, Bandwidth policy(on demand/peak etc), media processing/IVR and of course Identity

i) Reinhard – countering call placement and control as an API level abstraction – but we use SIP – SIP is open. (Skype by contrast is not)

j) Sam – as a counterpoint – it should be as easy as BT 21CN

Update from Sam:

When I referred to the efforts of BT as noted in item J above, I was using their SDK as an example of telecom making its services available for use by “regular programmers”. See for more information about their offer. As I said in the session, my teenage kids can use this SDK to make calls. Until telecom can be this easy to leverage, others will have the mindshare leadership.

k) 21CN is a standard to aspire for – everyone agrees.

l) Sam – yes, we can partner – but we have to also face reality – our competitors are much more nimble and aggressive(Google and Yahoo). We will partner with Operators where we can – but where we can’t – the market will dictate that we seek other alternatives

Update from Sam:

My comments regarding partnering with carriers are captured in a more negative tone than is appropriate. Microsoft strategy is to partner with carriers to deliver services, much as we have partnered with Dell and HP in the past to deliver software. We will work with the best willing partners. As the point captured above makes clear, we have competitive threats which we must meet. This is a call to step up to the challenge, not the threat which is captured in the original post.

m) Reinhard – Would developers come to me(Operator) as first port of call? No! (A fantastically frank answer!).

We control ‘x’ percent of the market(and I forget the percentage i.e. mobilkom’s market share). But that still leaves the rest of the market unaddressed. So, no .. Developers should think of addressing the whole market because any Operator will control only a portion of the whole. This mirrors my thinking as well. A carrier deal may not be the best option in all cases.

Update from Reinhard: I stick with my frank answer, but please add one more point: although we only have a market share of overall 40 percent, per segment this may be quite different, e.g. we have a 70% market share in the business segment. Given that, it may still be a good option for a 3rd party developer to go to mobilkom, especially if he is not going for the same market share that we have. If his service may hit the ceiling (in his niche) at 20% market share, it may be the best option to partner with us and get a penetration of 50% of mobilkom’s 40% market share.

So, that’s it

If you have questions, please comment and I will try to get answers

Many thanks to Sam, Reinhard and Mario

And of course, thanks to Mark Newman. Its great to be able to speak three times at a conference like this.

Update from Reinhard: Operators have to assure privacy of their customers, there are a lot of legal restrictions/regulations, e.g. we are by law obliged to prevent SMS-Spam.

Any external service provider violating has to be put offline immediately, otherwise mobilkom austria as operator is hold liable. So we have always to question: “opening the network means exactly WHAT? for the privacy of our customers ?” This is different to the Internet world, and the problems there with e.g. Spam, Spyware, Popups, … must not swap over on personal devices like a mobile.


  1. Danny says:

    From your perspective, what elements or aspect of the teleco should we open to help developers and consumers?