Interview: Digex President Mark Shull

Digex President and CEO Mark Shull spoke to Interactive Week Senior Writer Max Smetannikov about what's next for his company.

At one point during WorldComs acquisition of Digex, the managed hoster was valued at $120 per share. Now this multibillion-dollar asset is fully integrated with WorldCom, and its leadership is talking a new game: surrendering ownership of new data centers to WorldCom and enabling its parent to develop Digex capacities in the data centers of WorldCom partner companies. Digex President and CEO Mark Shull spoke to Interactive Week Senior Writer Max Smetannikov about whats next for his company.

In Interactive Weeks Sept. 3 Managed Service Provider special section, we have one of your executives talking about Digex abandoning ownership of its data centers. Could you explain this new corporate strategy?

I think we have always looked at the market and computing moving incrementally from mainframe and server environments to a network-based environment. I think you could put together an adoption model in which applications would move from an enterprise to a network. Thats what we have built our business around. In order to manage these applications, you need to control a certain set of resources. Our view has never been that you need a lot of data centers. The real challenge is not to have data centers and networks and all that - its really the ability to lights-out remotely manage these things in a very reliable, secure, repeatable way. I think we are far enough along so that in the latest generation of data centers we provide an engineering design, and WorldCom goes and gets these facilities built. We bring in our administration systems, so WorldCom is managing the physical layer to our specifications.

Will Digex quit the business of managing data centers and evolve to provide specialized data services, a la Loudcloud?

I think Loudcloud came out from an idealized technology world that proposed to build an end-to-end application platform that customers would use. You cant constrain customers to that extent. I think you will find that while they [Loudcloud] do have some business with some enterprises, generally youll find that business has not expanded and has not grown very much. Loudclouds first business model was that the big business opportunity was dot-coms. And when that went away, Loudcloud and companies like that switched to enterprises. And they got some pretty well-known people over there and they did get some early customers, but they are talking about the same customers they were talking about from the very beginning. The newest thing they are doing is to turn their technology over to their customers and let them use it. This is the reaction to [the idea] that what they created was too constraining. How many reiterations will they do before giving up because the model wasnt right?

The model we always had was more component-oriented, so we work on best-of-breed components to give customers flexibility in terms of which platform they use. Although you dont have diversity for its own sake, just to have just lots of stuff. Like in databases - you will look at DB2 or WebSphere, but because you use some IBM products, you dont go and use IBM servers. Once you pick the platforms, you worry about how they work together and automate that, and customers can select what they need. Our approach is well-engineered and automated. We have focused on not constricting the customers, which is Loudclouds biggest challenge.

The new data centers we have are owned by WorldCom. We dont want a cage; you dont want to be in big shared environment. From a security perspective, it has to be completely separated out, but we dont have to own it.

Do you see WorldCom taking over data centers you owned before the merger?

We have done that in the U.K., but not in the United States. Thats a theoretical possibility, but there is no motivation for it right now.

Having this separation between the physical and the application layer opens the opportunity to go into non-WorldCom facilities?

Thats our "servers anywhere" model. We have another platform in our labs that is designed to go into a "fly-away" data center: You can put together a number of racks and systems that could be deployed anywhere. The reason we built that was to address WorldComs need to have capability in, say, Brazil, because they have sales and network there, but we dont want to put a tier-1 data center there. [In Brazil, WorldCom owns former long-distance incumbent Embratel - M.S.] Depending on their sales needs, this enables them to quickly put up a data center that we can manage remotely. The other theoretical possibility for that is that if we can do it for a specific country, we can do it for a specific company.

What types of applications would you run if you were to deploy this platform in corporate data centers?

Are you going to do real mission-critical, highly secure transaction processing and database management? Probably not. It would be more applicable not for tier-1 distributed apps, but some of the distributed applications that a company might have in different locations. I believe that with the wide area of Gigabit Ethernet and the availability of dark fiber to enterprises, suddenly you can do networking between their location and our location - even if they are fairly far apart - at a very economical rate. In fact, you can do client-server applications.

When would this platform become commercially available?

Its available now. Its a matter of demand. Today, even global enterprises are asking us to support them in three major economic blocks, as opposed to emerging markets.

How do you envision your application service provider (ASP) services?

First, you need to parse ASP. Application software is going to happen. It just makes sense. Software as a service is a long-term phenomenon - Microsoft is moving in that direction, and so is everybody else. I would differentiate that from some business model referred to as ASP. If it aint broke, conservative CIOs [chief information officers] are not going to fix it. As software companies and system integrators deploy these software models, we will run those applications for companies.

Any particular system integrators?

We have worked with the major ones, the ones you would think of. We think these are key partners of ours. We are developing larger and more strategic relationships with system integrators, and there would be certain system integrators that we will have a larger relationship with.