Courtesy of CRNThe way organizations invest in and use technology is in a state of flux, according to business pundit Nicholas Carr. His 2004 book "Does IT Matter? Information Technology and the Corrosion of Competitive Advantage" set off widespread debate about the forces transforming ITâ€™s role in business. And in a spring 2005 MIT Sloan Management Review article, "The End of Corporate Computing," he argued that the way technology is now supplied to organizations is fragmented and wasteful and will be replaced by a utility model led by a small number of large, centralized IT suppliers. Carr discussed the shift and its implications in an interview with John Roberts, CRN editorial research director. CRN: How far along is the IT industry in progressing toward true utility computing? CARR: I've actually been surprised at how much progress has been made over the last year, particularly in the increasing acceptance of the idea of utility computing. There's still a lot of debate about how broadly it will actually be adopted, but there has been a lot of progress in software, in the recognition of software service as a viable alternative in some areas. There has also been progress on the hardware side from suppliers that are introducing new types of utility services. So even though I think the adoption of utility computing by businesses will go fairly slowly--because so much has been invested in the old "own it ourselves" model--my own sense is that adoption might go faster than I would have thought a year ago when I wrote the article. CRN: So businessesâ€™ attitudes toward utility computing are becoming more favorable? CARR: That's true, at least over the past year. I think more users are seeing the utility model as a buyable alternative--at least in the future, if not right now. CRN: Is this the case throughout the business market, or is it concentrated more among small and midsize companies? CARR: As far as changing attitudes go, I think it's across all companies. But as far as actual adoption of utility computing, it's definitely greater among small and medium-size businesses. I mentioned in my article that this would probably be the case, simply because the old model of "own it yourself" IT requires considerable amounts of capital to invest in hardware and software. Utility computing offers these smaller companies a way to get some of the capabilities they need without having to invest a lot of capital. CRN: Why are larger companies so slow to adopt utility computing? It seems they could take greater advantage of the economies of scale that utility computing offers than smaller firms could. CARR: Because the bigger companies have already invested so much, it really will take a major event to get them to change over [to utility computing], like upgrading their data centers. They are not going to begin to switch over until there is a real need to re-evaluate their future investment. CRN: Is utility computing something solution providers should fear? Will they be disintermediated as businesses deal directly with big suppliers of utility computing services? Or will solution providers play a role in this process? CARR: I think ultimately it will be a threat to resellers, in particular, over the longer term. In the shorter run, I think an important thing to recognize is that for most companies--particularly the larger companies--their first step toward utility computing might not be to go outside and buy actual utility services but to set up their own internal utility-computing platforms. For instance, if a company has many different business units, those business units might have their own IT infrastructures. I think there's a big opportunity for companies to consolidate and standardize IT infrastructure across their entire company, across all their businesses, and run it as an internal utility. Smart resellers can certainly play a role in helping companies create more efficient, more consolidated and more standardized infrastructures. But I think in the longer run, more IT capabilities are going to be purchased as services delivered over the Internet, and that will inevitably take money away from the old model of IT. So in that [way], it will be a threat to the resellers. CRN: Youâ€™ve said some major hurdles for the utility computing model include security, flexibility and risk. Can these issues, notably security, be overcome? CARR: I think it can be overcome and, ultimately, the utility model will be more secure than the current, very fragmented model. Having said that, I think the onus is on the utility providers to earn the trust of clients. They need to show not only that their platforms are secure, but that they're reliable and the integrity of a company's data will be maintained. I think thereâ€™s still a long way to go for the providers to prove that. But itâ€™s important to note that with a utility provider, the entire business hinges on the security and reliability question. So that's at the very center of their business. Itâ€™s interesting that when you look at where utility computing has made its biggest inroads, at least on the software side, is CRM, which is pretty sensitive data. Clearly, some businesses have made the leap of faith to allow a utility provider to store all this information. So I think the case for utility computing is being made and will continue to be made. CRN: So as more and more companies see success, the move to utility computing will snowball. CARR: Right, and the adoption process will also move from the small and midsize companies to larger companies. In some cases, smaller companies really don't have any other option than utility computing because they don't want to build their own data centers. But this also provides a way for the utility providers to make the case for utility computing, and sooner or later other companies will begin to feel comfortable in going that route. CRN: Do you think the rise and fall of ASPs model has inhibited the move toward utility computing? Is there a "once bitten, twice shyâ€ attitude among business executives? CARR: I don't really think so, because I don't think a lot of companies were "bitten,"so to speak. I don't think they bothered with the ASP model, which is one of the reasons it failed. The original ASP model was an example of the first wave of pioneers being killed, which isn't all that unusual in the technology world. A lot of their customers were dot-coms, and they built up a lot of capacity to serve customers that quickly disappeared. So from an economic standpoint, it just didn't work. From what I've seen, most companies are judging the new providers on their own merits rather than saying, "Oh, well, it didn't work before, so it's not going to work again." CRN: It seems like thereâ€™s going to be a tricky balance between competitiveness in utility computing and the need for big providers. CARR: Right, and the good news is that there does seem to be some pressure toward having open standards at the core of the utility computing model. This trend toward some standardized, open data formats could provide a good counterweight toward tight monopoly control. CRN: Youâ€™ve said it would be years before true utility computing is achieved. Is that still the case, and how many years will it be? CARR: It seems to be going a bit more quickly, but it's certainly going to be more than five years--if by â€œtrue utility computingâ€ you mean big companies getting a lot of or most of their IT capabilities from outside suppliers over the network. I think we are looking more at 10 to 15 years. And it is going to be a very evolutionary process from now until then, with a lot of challenges and disruptions throughout the IT industry as more and more pieces of the puzzle are put in place and as companies begin to shift more of their money away from buying and maintaining their own IT to using the utility model. Despite this, I think the worst thing you can do--whether you're a buyer, a supplier or a reseller--is say, "It's going to be a decade or something before this really takes hold, so we don't have to worry about it." CRN: What will some of those disruptions be? CARR: Traditional software pricing models, which take it as a given that every company is going to buy and install its own enterprise applications, will fall apart as you begin to get these applications supplied centrally through utilities. It will no longer make sense to charge companies based on the process or based on the seat. You now have the ability to precisely monitor actual usage, just like you can monitor actual usage of electricity. That's going to put a lot of pressure on the software companies to change their pricing model to reflect the actual usage of their product, rather than the proxies for usage that have been used up until now that have never really reflected true usage of a product.