The man behind the webtone switch

Published on:

Sun founder Scott McNealy now has a strategic position in the company. He's having what he calls 'big conversations' with leaders of the telecoms industry about Sun's adherence to open standards, and his vision that they should be providing the 'webtone switch' as a way of achieving the company's long-held belief that 'the network is the computer'. Interview by Alan Burkitt-Gray

McNealy: Now I'm the CEO's
airplane crash dummy

Technology, says Scott McNealy bluntly, has the shelf life of a banana. That's why he and the company he founded, Sun Microsystems, believe so fervently in open standards. He doesn't want to lock customers in to proprietary in-house standards, he says, because at some time they will want to move on to the next generation of technology.

He has stopped being CEO of Sun Microsystems, but he's still chairman of the board of directors of the company he co-founded in 1982. He was CEO from 1984 until earlier this year, during which period the company created Java — now used on 3.5 billion devices — and developed the Solaris operating system. And he created the slogan, "the network is the computer", before the internet became all-pervasive: today, that ideal is more realistic.

Now, the CEO is Jonathan Schwartz, a former McKinsey consultant who founded Lighthouse Design, a company taken over by Sun in 1996.

So what's McNealy's role now? "The clever way to describe it is I'm Jonathan's airplane crash dummy," he jokes. He spends his time flying around the world discussing strategy with Sun's key customers — "the telco market is our number one market" and he's "spending an enormous amount of time" with them.

While "Jonathan gets to run staff meetings" and look after Sarbanes-Oxley compliance, "I've got a fun job," says McNealy. "Someone described it as being a grandparent rather than a parent."

There's justification in his remark that "telcos have been our number one market for ever". The Unix operating system on which Solaris is based was developed by the old AT&T at Bell Labs — now owned by Lucent — in 1969.

"Solaris is the operating system of choice for the telco and network equipment provider marketplace," says McNealy. "We do an enormous amount of OEM business with the Alcatels and the Ericssons and the Lucents. We continue to play big in that and that's absorbing an enormous amount of my time as we work with the carriers on how we move to the IP environment."

That OEM business is about to take a further step: Sun is playing a leading role in a new open-standards-based concept for a telecoms platform. It has signed up a number of vendors that it will be working with to develop the specification: see panel.

What are the big issues that CEOs and CTOs of the larger operators should be watching out for?

"They need to get the authentication and identity management right," says McNealy. "That's our most important conversation and we're the number one player in large-scale identity management. If you don't know who's who, what's what and who gets access to what, you get a free-for-all in service and you don't know who to bill for what."

The second big issue is scalability, he continues. "There are economies of scale and I believe our track record with the service providers, in providing carrier grade reliability, is unmatched — especially when you look at some of the newcomers into this marketplace."

Zero barrier to exit

And thirdly, openness, "because all technology has the shelf-life of a banana", says McNealy. "You need to migrate to the next generation or a new supplier, and the zero barrier to exit philosophy we offer provides the most flexibility for the carriers long-term, so they don't get trapped into a mainframe or PC-like, or a Microsoft-like decision."

Sun's open source interface gives "the lowest barrier to exit" for systems developers, he explains. "If they write to a mainframe they're stuck. If they write to Windows, they are stuck. There's no second source. We have open source, open interface. We give them the lowest barrier to exit to the next generation technology."

There are, he says multiple versions of Unix and Solaris "that have been ported to and are certified to run on more than 700 non-Sun machines". That gives companies "an enormous choice in hardware, operating system, middleware, platforms and migration path". But, he huffs, "people are still stuck on mainframes".

As the industry is changing, the really big question for an operator to answer is whether it is focussing on a carrier-based model, "just store and forward packets", or whether it is "going to get into something that's a little stickier", he says.

"They're all struggling with the question: 'Are we a carrier or are we a full service provider of network-based services?'. Everybody's trying to figure out how to be sticky, how to lower churn, how to raise ARPU. There's been a lot of companies who've lost focus on their voice services, specially in the wireless space. Staying focussed on voice is very demanding."

And he's concerned about the changing relationships between operators and the new internet companies. "Does AT&T partner with Yahoo!? Great. But then does Yahoo! announce instant messaging and voice over IP and go after AT&T's revenue base? Oops," says McNealy.

Partner or competitor?

"There are always those kinds of challenges. Is MSN — as Microsoft — a partner, a technology supplier, or is MSN a competitor longer term for subscriber management? These are the questions everybody's trying to figure out. Who's my partner, who's my supplier, who's my competitor, who's trying to get the subscriber dollars and how?"

That's why McNealy is determined to make it clear that Sun isn't competing with the operators. "We're not doing an MSN, a Google or a Yahoo! service. We're not doing a voice over IP service. We're not doing an instant messaging service," he says. "Our strategy is not to compete with the cable companies, the internet service providers or the telcos, but rather provide them with tools to improve churn and ARPU and subscriber acquisition."

How are operators managing the challenge of deciding their future route? "I don't want to do public performance reviews of my major customers but I do believe that Microsoft and Google and Yahoo! are in very interesting and quite strong positions to offer things like voice over IP and IPTV and those sorts of things on their own websites," says McNealy.

"People are probably more affiliated with their internet service provider than they are with their carrier — in terms of stickiness, where their content is, where their affiliation is."

So if customers are drifting to MSN, Google and Yahoo!, how can the carriers pull them back? "There are two ways to do it," says McNealy. "One is to buy your way into it. There's always that acquisition challenge. The other one is to start providing more value-add services — mail, calendaring, instant messaging, VoIP, IPTV, all of those sorts of things."

Where are the new ideas?

But that's not easy, he readily admits. What they really need to be doing is come up with ideas such as MySpace, Facebook or YouTube, "start generating some properties of their own", he says. He points to cable operators such as Time Warner and Comcast which have their own content, which they can use to barter with other operators for yet more content.

Where are with in this struggle for the subscriber? What are McNealy's views on where customers' loyalties will go?

"I would hate to guess how it's all going to shake out," he says. "Does MSN win, does Microsoft win because they control the desktop? Does Google win?"

On the other hand, a big carrier can drive up the quality of service — a carrier such as Deutsche Telekom or BT, he says. "Deutsche Telekom has all of the pieces — the systems integration business, they have T-Online, wireline and wireless. They have the pieces to be a full service provider." But as to whether they or the internet companies will win, he admits candidly: "I wish I knew."

McNealy believes operators have a great opportunity to develop new web-based services for both consumer and enterprise customers as a way to win extra revenue. He calls this concept "the webtone switch", and he compares it with the "dialtone switch", the conventional telephone exchange that provides dialtone to every phone.

"The webtone switch is basically a grid infrastructure with storage that would provide the webtone on top of the dialtone," says McNealy. "You can run any supercomputer app you want on the grid as webtone."

There are already some applications that use this: he cites, the online CRM service that companies can use at a rate of $65 a month upwards. "Google is webtone: you don't buy their grid, you just use their grid. And eBay is webtone," he adds.

"Why aren't AT&T offering a webtone grid for any new start-up to use," he asks, "so they can just write their service — like the next Amazon could write their commerce software on that grid, or the next eBay could do their selling application, their bid/ask application. Even the next Google could write their search environment on that grid, and save all of these new start-ups all of the investment in a data centre, and they can just start writing software."

Sun is promoting what he calls "the dollar per CPU hour grid computing model", including computing power, storage and display. "That's kind of telco-meets-IT," he says. "It's the way to drive ARPU — and it's general purpose compute capability so somebody could run their SAP or Oracle application on there. They could run their messaging or their archiving. They could literally run their desktop on the server and just display it on a thin client environment."

Google is already "starting to go in that direction" with its web-based spreadsheets and other applications. "They're building an office grid, if you will."

This gets McNealy — speaking from his home in California — really warmed up to the theme: "We don't have a telephone switch here in this house. We don't have telephone switches in our sales offices. We use the telco provider. So why would we have a little data centre — and by the way a PC is not such a little data centre any more? Why do we all carry little data centres around with us when the network is the computer?"

The big conversations

This is "one of the big conversations" he's having with the telecoms industry: "Why don't you take on the PC administration nightmare that everybody has, as well as the data centre nightmare that everybody has, by putting that in your own switchroom, and marrying it to your high-speed network, and taking advantage of the availability of the network to centralise that stuff and then drive much better economies."

In the old days, he notes, "people used to hug their money, put it in their mattress". And then it became safer to put it in a bank vault. "And I don't want to keep data in my house, so we put it out on the network. Mobile computing should be thin, because the networks are getting very good."

Telcos are starting to think how they can take over opportunities in IT, he says. "I tell you what they like about our story. One, they do like this model because they see a way to go after enterprise dollars and to go after consumer dollars. Two, they see that we're not going to compete with them like Microsoft in the consumer space and IBM in the enterprise space. IBM Global Services is the one that is out there generating all these custom data centres instead of getting people to go use a standard telco-based data centre."

So are operators listening to him? How does he make them focus on these new ideas? "That's my job and that's why I get in front of them and talk to them about the new opportunities and the ways we could work together."

So does he see them converting to the view that this is the way to go? "It depends on who you talk to and what their areas of focus are — and whether they've been burned," says McNealy.

"Some of them are working with Microsoft on IPTV. Now is that letting the shark into your bathtub? Is it going to work? Is it going to scale? Those are all questions and we're having these conversations. We've got a very open, low barrier to exit, standards-based approach. We're doing that with some carriers, and we're insurance policies with others. Those are the kind of conversations we're having." GTB

 The open-standards platform

Sun is recruiting vendors to create a consortium of suppliers of open-standards telecommunications. The idea behind it is to reduce the cost and increase interoperability by developing a standardised platform which individual vendors can turn into systems that are targeted at specialised applications.

The Telecommunications Platform Product consortium — or TPP — is still partly under wraps at the moment, but it is understood that Sun is recruiting a number of equipment vendors to support the scheme.

TPP "is initially targeted at the service layer, plus operations and administration, but eventually it will also operate at the core layer", says Peter Ewens, vice president of Sun's OEM group. "The benefits are lower cost, a greater degree of interoperability and ease of deployment." The individual vendors will add value to the standard platform, a "base-level computing platform" which Sun will supply.

The specification of the platform will be "driven by the consortium", says Ewens, who adds that other vendors than Sun might also deliver platforms to the TPP specification. "We intend to make the definition public, and it will be an open specification," he says.

Sales of the platform will not be restricted to members of the consortium, he adds. "The more customers there are the more everyone benefits — the cost will be amortized over a greater number of customers."

Darrell Jordan-Smith, vice president for telecommunications at Sun, says that the platform will help network equipment providers reduce the cost of systems they supply to the service providers.

"It's a service delivery platform, a carrier-grade platform. Service providers are still trying to figure out what services to offer on their networks and they need to reduce costs. Using open standards will give them flexibility. Operators will be able to roll out new services in only 90 days, when it used to take a year."

Equipment vendors will be able to differentiate their systems by adding value — a process close to systems integration, he suggests. "We at Sun will supply carrier-grade middleware. It's a carrier-grade platform."