With Richard Jones, vice president and service director for the data center strategies service, the Burton Group.

Question: What will we see this year in cloud computing?

Jones: I think during 2008 we will see some continued growth in the experimental phases. [It's reminiscent] of what happened when you had the Arpanet and the government helped it grow to a point where it took off in 1995. You can look at that as a good example of what probably is going to happen with the whole idea of cloud computing.

Question: What must be overcome for cloud to succeed?

Jones: Organizations have to move from traditional client/server and SOA-based applications [that are dependent on static allocation of resources]. What's happening here is virtualization — granted, everyone is pounding on that term — helping to break those barriers down from the technological perspective.

Question: What is the biggest obstacle?

Jones: What is going to hold us back are political issues in the organization. What is the CIO or IT manager who is responsible for mitigating risk thinking? [What may be going through] his mind is that he will have no control. So his [applications] could be running right next to competitors'. All of that infrastructure needs to be put in place first. Amazon made the initial attempt, but it had a lot of problems. People are leery … and more curious right now. In my own area, two of the local television stations as well as one university — Channels 2 and 5 and the University of Utah — got together on a high-performance computing structure. They use it for weather forecasting for the two channels. It's kind of like a proprietary grid or a community grid. It was just for members of that community. The groups had to join the community to manage it. They were crude SLAs.

Question: So a lot of the challenges are not technical.

Jones: There are still a lot of political boundaries that have to be overcome. I was originally of the opinion that we had to go through a generational change of IT managers. IT grew up in the '90s. [I thought] those IT managers would have to retire and a new generation come into the market during this decade [for cloud to catch on]. That's mainly because those who started in the late '80s and early '90s all worried about security and huge virus problems. They are not about to think about putting critical system and processes into the cloud. [The ones to do that] had to grow up with blogging, IM, the next-generation social sites and be the folks who understand how to operate in that model without getting burned.

Question: Do you still feel that there has to be a generational changeover?

Jones: I still do to a degree, but am not as adamant as before. I've seen a lot of IT people accept and understand the new technology. So I think they are subject to change. You've got some of the technological concerns getting hashed out. For example, how to build an SLA. When I used to think of the political side, I originally thought it was a tougher nut to crack. Now I am thinking it is not as tough as SLAs [and similar issues]. The jury is still out on that.

Question: It seems that progress is being made.

Jones: What we've seen happen in the early part of this decade is a lot of different attempts at software-as-a-service on the Internet that people said were failures. Now we are seeing trends — a great example is Salesforce.com — that are very successful. They have only grown in the last few years. There has been a political and attitude change with CIOs. Some was forced on them. The CFO has gained more power and the business metrics were pushed on [IT]. And so some of them have gone to the model grudgingly. They can't argue against numbers. Some see the economies of scale. That’s a good trend. Now instead of static services, you can go out over the Internet, where essentially any service you need to run can be found. You can look at the cloud as a timeshare. Politically, the boundaries have broken down a bit faster. Google, IBM [and others] are experimenting more with business models. Most of these are still in the experimental stage. People on the bleeding edge are saying [that there may be] value to get out of these, so they are trying them out.

Question: The reality is that we won't know if it is a success, because there is not a big launch party with a famous band playing.

Jones: Vista, [for example], had a launch event. It was a new piece of software that is self-contained and static in nature. For the cloud, the analogy is much more like the Internet. There wasn’t a day in 1995 or '96 when they had a launch event. It seeped out. The compute cloud is a piece of infrastructure that people began to use. It is not a self-contained entity like Windows Vista is.

Question: Are SMBs the first to adopt cloud computing?

Jones: Historically, that's the way a lot of the trends matured. The bigger enterprises are slower to move off older stuff. The financial industry, surprisingly, likes to play with the newest technology, but at the same time is risk averse. So you'll see in most of the finance organizations still a lot of mainframes running old-style client/server-type things. Still, there also are a lot of higher-performance Linux-based compute clusters. Especially SMBs need access to inexpensive computing. They can't go out and build a supercomputer. But they may have $15,000 to rent time on a supercomputer cloud out there. A company called Schneider Trucking is using a high-performance computer cluster just to manage trucks to find the most efficient routes and where to fuel up.

Question: There seem to be parallels between open source and cloud computing. Is that so?

Jones: There's a lot of parallels. The reason the cloud likes Linux so much is … that it is in the experimental phases, and you need the source code. With Linux, you don’t have to worry about licenses. If you use Windows, it just becomes very cumbersome. Cloud computing is a serious form of collaboration. The tie is flexibility and the need to make it so much easier to manage. But when you can get beyond the initial experimentation phase, you see stuff from Microsoft. A great example was the Microsoft compute cluster server in 2006.

Question: What will it mean if cloud computing does end up predominating?

Jones: What it means is the way people run the IT side of business definitely will change. They will just have to worry about the business processes necessary to run the business effectively. The performance will be there and it will be an inexpensive way to run things that is highly available.