InformationWeek: Is there a limit in terms of the amount of scale that an enterprise customer should expect from a cloud service? We're not going to see FedEx throw its package-tracking system into the cloud yet are we?

Vogels: I would love them to do that. I think that would be great benefit to them. And this is sort of a side comment, one of the drivers for why enterprises are considering clouds is there is a very good financial driver -- less cap ex, more op ex. Shifting investment from up-front to usage-based investment is something that CFOs at this moment clearly love. From an innovation point of view, quite a number of enterprises are considering moving some of their services into the cloud and then opening them up such that they can become part of the cloud ecosystem, making it easy for Company X to access their services in the cloud and third parties to extend the platform they're building. Whether that will be FedEx or telcos, there's a whole range of enterprise services that companies are considering moving into the cloud, opening up their services so that they can become part of the ecosystem and so other people can build applications against that. This goes beyond traditional enterprise services. Salesforce released a toolkit so customers can easily run Salesforce applications using EC2 and S3. Integration between these different platforms will become increasingly important.

InformationWeek: Let's talk about multiple clouds working together, private clouds, and hybrid clouds. Do you see that businesses would want to create private clouds in their own data centers?

Vogels: It's clear that a number of enterprise customers would like to have some of the advantages of easy deployment and a uniform application container, where they can seamlessly migrate between their own data centers and the external cloud. Whether you call that a private cloud, fine. It's mostly important in that context that there are a number of advantages to deploying things in the cloud, the ease of use and flexibility that they would like to have in their own data centers.

And there are shades of gray. I've heard customers talk about partner clouds, where they may have an internal one for ease of development, maybe some shared services with partners and, after that, they move things into the public cloud. For most of these things, it's really day 1. Who will deliver this software, and can this be done, is still under discussion. Who, how, what, and where that will happen. For us, it's important to focus on the kinds of services that Amazon is delivering now, making sure that we execute against performance, security, reliability, scale, and at a cost level that makes sense to our customers.

InformationWeek: As this grows with more customers serving more users using the infrastructure in more ways, does it begin to test what you and your colleagues know about scale in the sense that you're doing things that have never been done before?

Vogels: I think we are doing things that have never been done before. The kinds of things that we needed to do to develop the services as they have been running now for the past two years were already really unique. We haven't only borrowed on principles of distributed systems techniques out of research, we also needed to develop a lot of engineering techniques ourselves. Most of those were proven through time to work really well within the Amazon e-commerce platform before we brought them into the AWS world.

Will this grow beyond that? I think there are a number of standard principles that we can apply in terms of hierarchies, of loose coupling, of probabilistic techniques that I'm confident will serve us for quite a bit of time. When we developed these services, we were looking ahead in terms of what kind of scale we could achieve, and we're not there yet. Even then, I'm confident that the choices we've made were the right ones.

Return to the story:
Chief Of The Year: Amazon CTO Werner Vogels