Earlier this week I spoke with Erich Clementi, General Manager, Enterprise Initiatives (otherwise known as the head of IBM’s cloud computing efforts) about Big Blue’s cloud strategy. After we raked the computer and service provider over the coals earlier this year for talking about the cloud without offering substance, in June IBM finally unveiled part of its cloud plans. They revolve around providing workload-specific services via an IBM cloud, as a hosted cloud, or inside a company’s own data center. It kicked off its cloud rollout with a test and development service, and last month it announced an analytics offering. Clementi revealed that IBM won’t stop at workload-specific services, and will build a WebSphere platform-as-a-service offering for clients.
It would go up against other general purpose platforms-as-a-service offerings like Microsoft’s Azure platform, Rackspace’s Mosso effort and to a lesser extent, platforms like Google’s App Engine and Salesforce.com’s Force.com. And earlier this week, VMware purchased SpringSource and said it would build a platform as well. Clementi was cagey about plans to provide an infrastructure-as-a-service offering, however, despite rumors I’ve heard about IBM’s interest in providing compute cycles. Below is an edited version of our conversation:
GigaOM: What Does IBM mean when it talks about cloud computing?
Erich Clementi: We think every 15-20 years there’s a big shift in the computing model. We have been at the top of the hill around the centralized compute model but we overslept on the distributed model. We saw it coming but we didn’t react to client service server. And in 1993 we nearly went belly up when the industry shifted, and that’s not going to happen to us again. So when you ask me what cloud is, it’s the new delivery and consumption model for IT support and service. It’s that broad. The cloud is nothing but the industrialization of any IT supported services. Many people equate cloud computing to virtualization. It is not virtualization. To get the value you need standardization and automation on top of that virtualization.
GigaOM: So clouds don’t need to be virtualized?
Clementi: Google is not virtualized and virtualization is not sufficient to qualify as a cloud — it’s a better use of the physical infrastructure, but the real bang comes from modeling out the whole data center and taking energy, labor, software, and hardware and acting on all those levels. It’s like Google’s idea that the data center has become the computer.
GigaOM: So then will IBM provide Infrastructure-as-a-Service products or PaaS products?
Clementi: We need to understand what the enterprise use cases are. We obviously need to have something for cycles for our CloudBurst services, so there is a compute cloud underneath that. But Amazon is a street corner — an outlet where people can go to get compute cycles. We are going to give our customers choice, but those choices have to be enterprise-grade. That means you have to have SAS 70 certification and an audit trail, not let an employee swipe a credit card and have everyone order machinery.
There’s a lot of specific use that people are making of Amazon and others, but in the enterprise the first step in the short term is deployments of hybrid environments that customers will build internally and compliment that with public cloud services. That’s a very rich space that can deliver business for years.
You should see us play in the platform space. We have the largest base of middleware in the industry, and as the computing moves to a cloud delivery model we are going to support that move. Look out our CloudBurst services and WebSphere; we are clearly going to have our software stacks as a platform.
GigaOM: So what’s the biggest challenge to moving your business — or any business — to the cloud?
Clementi: How do I move from a license model to a recurring revenue steam? That is the big transition that business are going to make. When I sell a license, I monetize that up front and that requires a different investment stream and changes how fast you build up revenue. We’ve gone through a period like this before when you had a rental model for computers to a straightforward pay model, and these transitions are pretty significant.
Share moves happen only when there is disruption. We were pretty good in mainframe, and this time we see a big opportunity because this is centralized computing and it has the potential to change the economics of enterprise computing. It requires a complex set of skills, from service processes to software and hardware experience.