The Network Is The Network
Sydney Morning Herald
Monday September 22, 1997
What's been on Eric Schmidt's mind since he left Sun to take over Novell? SPENCER REISS finds out.
ERIC Schmidt (pictured) liked to say that being Sun Microsystems' chief technical officer was "the best job in the world". For a Berkeley PhD in computer science, it probably was. But in April, he traded his double-digit (number 92) Sun employee badge for the spot as chair and CEO of moribund Novell, the Utah-based networking software giant.
More than a few Silicon Valley jaws dropped. But Schmidt, who's spent a lot of his career in research - including a stint in the early '80s at Xerox PARC - says the move isn't just a case of CEO envy: it's his chance to push the limits of his former company's celebrated motto, "the network is the computer".
The following questions on networking and the future of Novell were put to Schmidt at his Silicon Valley office. You've talked about technology "inversions" being the driving force behind the information industry. What do you mean?
By inversion I mean something that was known as a "truth" - mainframes, time-sharing, browsing - but then turns out to be the wrong model once technology proceeds. What's a current example?
Look at what's happening in telephony. All the major equipment suppliers are now producing hybrid switches - analog and digital. So a relatively rigid, circuit-based structure is evolving into a true network. We started out running the Net on top of the phone system, and we'll end up with telephony running on top of the Net. A completely unregulated network is toppling its highly regulated predecessor. That's extraordinary. What does that mean for the Net's overall architecture?
I used to worry that the Net would become another oligopoly, which is how telecommunications has always worked - the big players, the little players, and a rigid structure. There were these main connection points - MAE-East, MAE-West, actual physical places - and everything ran through them.
Now the second-tier players - the regional ISPs, the ANSes, UUNets, and MCIs - are cross-connecting. And companies are setting up their own private networks. Once that's done, boom! There's this explosion. You make it harder for a few companies to monopolise anything. Are you confident the Net can evolve in the right way?
As long as no single company dominates any of the key connection points. As long as the protocols can evolve, as long as the data network can evolve. The beauty of the Internet is that people will have all these debates about mixed-tier pricing, and security, and so forth. My answer is yes to all of them, because the Internet runs every experiment in parallel.
Just when you start to get bored, it throws you a great challenge. It's too big. Some people say the Net should be a little less exciting. I would be careful about any efforts to be more managed. The current structure works remarkably well. Another of the inversions you mentioned is from browsers to push technology. Why is that happening?
Today's Internet is a return to time-sharing - it's faster at night. Is that OK? No! The right answer is for information to come to me automatically, according to established preferences - mine, the network's, or both. That's what push does, from a systems point of view: it smoothes the load, making maximum use of bandwidth. That's also where satellites can play a big role - delivering bulk data to specific time zones. A colleague says that your real talent is coming up with "stories" for a technology, and that Novell has a lot of really interesting technologies in search of a story. I'm working on that. It's a two-part strategy. First, there's going to be this explosion in networks; lots of services. These need to be managed. Second, companies will want one interoperable platform that provides these services very, very well. So you create one big network everyone can access.
There was a huge mistake made in the last 10 years - lots of people tried to build specialised devices, and they ignored the issue of ubiquity. What the Net did was create a lowest common denominator, HTML. It was a bad lowest common denominator, but it created a market for content, which created devices, which created the opportunity for services like @Home. How do you manage a big, established company in that kind of environment?
Traditional management theory failed to anticipate the rise of intellectual capital as the sole competitive weapon. So what you end up with are very fast-moving corporations, the importance of branding, and a lot of topsy-turvy - an ecosystem, in other words. Where does Microsoft fit in?
Microsoft is an enormously powerful force because of the monopolistic control it has over certain platforms. All we're arguing is how to deal with it. The media portray it as "us versus them, anybody but Microsoft". But in pragmatic terms, the customers - those who should be driving this - are saying, "I like this application" and "I like that application". They just want everything to interoperate. Why did you want to be a CEO?
I wanted to work in this area. Network services is the most interesting area for the next 10 years in computing. All the other markets are consolidating very quickly. Intel controls 95 per cent of the computing market. Microsoft controls 90 per cent of the platform market. We could fight over the other 10 per cent, but it's not going to go to 20 per cent any time soon.
The way for Novell to succeed is to define a new market, occupy it and drive it. It's a lot more like Hollywood and Wall Street than it is like car manufacturing.
Basically, Novell needs a hit. Absolutely. It has been defocused by a series of acquisitions and forays that didn't work out. In this collaborative world, it's more important to do a few things well and just go for them like you've never seen. So the network is the computer?
The network is the network.
© 1997 Sydney Morning Herald