The Future of On-Demand Computing

Computerworld - We asked a variety of industry visionaries to look into their crystal balls and predict the future of on-demand/utility computing. Here are the most interesting forecasts:

• Standards needed. Virtual pools of resources will be a failure unless there are standards for cross-vendor management in place by 2006.-- Corey Ferengul, analyst, Meta Group Inc., Stamford, Conn.



• The open-source competitor. Adoption of utility computing on a broad scale has an unlikely competitor: the open-source software movement. Over the next two years, the more the open-source model succeeds, the more likely it is that customers will keep their IT work in-house, to the clear detriment of utility-style providers. -- Doug Tuttle, director of the Global High Technology Practice, Deloitte Consulting, Boston



• The OS winners. By 2008, Linux and Windows will be the only operating systems that matter for utility computing. -- Brian Berliner, chief technology officer, Cassatt Corp., San Jose



• Doubling your utility. Today, utility computing is in its infancy. IT doesn't have an economy-of-scale economy that amortizes fixed costs as volume grows. We need to work on the systemic issues so that when someone asks you to double the size of your utility, you don't have a failure. Within two to three years, new architectures will be in place. By then, you should be able to say, "Hey, I'm twice as big now but delivering incremental cycles that are practically free." When this is achieved, broad adoption should follow. -- Greg Papadopoulos, chief technology officer, Sun Microsystems Inc.



• Built-in manageability. Utility computing assembles and delivers IT services on the fly to improve the scale, flexibility and productivity of business operations. To achieve this goal, IT will need to manage itself on the fly like a service -- in sharp contrast to traditional systems management, with its centralized control of individual IT resources. Manageability will need to be built into IT products and services as an inherent attribute, instead of bolted on as an afterthought. By 2007, enterprises will no longer need to invest in multiyear, multimillion-dollar deployments of systems management products, which will put traditional systems management vendors on the endangered species list. -- Mike Maples, co-founder and chief strategy officer, Motive Inc., Austin



• Back to the future. Utility computing will inspire a return to true capacity planning, and by 2006, there will be a major demand for individuals with capacity planning and improvement knowledge. -- Corey Ferengul, Meta Group



• Uncle Sam's leadership. By 2005, you'll see the federal government investing more in utility computing than the 10 largest private companies (combined) in the world. -- Mark Forman, executive vice president, Cassatt



• The cobbler's son has no shoes. Every other part of a company today is automated, so why not IT? IT must be run like a utility, on tap, if we ever want to free up our most skilled talent to do smarter, more creative things, not just baby-sit servers and applications. The organizations visionary enough to start down this path today will find themselves with a fully automated utility computing model in less than five years. Perhaps they'll even achieve "IT nirvana" -- one man overseeing 1 million machines. -- Tim Howes, chief technology officer and co-founder, Opsware Inc., Sunnyvale, Calif.



• Virtually no difference. By 2007, virtualization software will be a normal part of the operating system, included at no additional charge and providing no differentiation for vendors. -- Corey Ferengul, Meta Group



• The upside. On-demand computing will fundamentally alter the relationship between IT and business operations over the next five to seven years. IT efficiency will be dramatically improved; distributed computer utilization will be above 75%; service levels will improve by a factor of 10; businesses will have visibility into the IT costs of supporting any business process. In short, IT will begin to reach its full potential as an efficient, responsive, strategic asset. -- Loren Weinberg, technology strategist, Computer Associates International Inc.



• Third World benefits. Over the next four years, the adoption of the utility computing model will accelerate the advancement of developing nations faster than offshoring. It's far easier for developing countries to adopt the utility computing model, since they don't have to revamp 20 years of existing technical infrastructure to get the benefits. The analog to this is the rapid adoption rates of cellular telephones in China. -- Doug Tuttle, Deloitte Consulting



• No compelling need. In the next two years, I don't see widespread, sustainable adoption of on-demand computing except in certain vertical markets such as risk management or analytics. Just as we saw with software-as-a-service (now hitting its stride after five years of hype), most corporate users at this time have no definitive business demand for techniques such as on-demand computing. The early entrants will gain some recognition and limited commercial success, but until on-demand computing can demonstrably and significantly reduce costs and deliver definitive ROI, it will have limited user acceptance. For those with the "new new thing" bug, concentrate on either the security or RFID marketplaces, not this. -- Chris Clabaugh, vice president, Progress Software Corp., Bedford, Mass.

Read more about Hardware in Computerworld's Hardware Topic Center.