News Stay informed about the latest enterprise technology news and product updates.

Face-off: IBM's on-demand computing

Utopia or just a computing pipe dream? Our editors debate the pros and cons of IBM's on-demand computing initiative.

Face-off A brilliant marketing strategy by IBM has many IT pros preparing for an on-demand Utopia. Others are more distrustful of IBM and wonder how much this is really going to cost them. Is IBM's on-demand initiative the future of computing or just another way for IBM to dig deeper into your pocket.

Self-fulfilling prophecy

By Mark Brunelli, Site Editor

In a world where IT pros are constantly bombarded with promises of revolution, thanks to the latest doomed-to-fail technology concept, it sure is nice to know there are still some ideas out there that give me a warm feeling inside. It's a feeling that says, "Hey man, this idea really could lead to a brighter future."

I'm here to say that the future of technology is here. Ladies and gentlemen, I give you the Laserdisc. Just kidding. I'm really talking about IBM's on-demand computing initiative. On-demand's value proposition of streamlining budgets and procedures is so undeniable that the concept sells itself, and in time will become the norm for IT organizations everywhere.

For those that don't know, on-demand is a concept that IBM has been throwing a lot of weight behind over the last two years. Under Big Blue's vision for the on-demand world, the formerly separate departments and suppliers of an enterprise come together to share the same computing infrastructure, but more importantly the daily operations of each department become more tightly integrated.

IBM's on-demand vision is often referred to as utility computing, the process by which companies pay as they go for hardware and software as capacity requirements change. But even though utility computing is an integral part of IBM's strategy, the on-demand concept goes much further to look at how organizations can align key business processes internally and externally with partners to eliminate red tape and save time and money.

As IBM CEO Sam Palmisano puts it, an on-demand organization has its business processes integrated end-to-end across the company and with key partners, suppliers and customers, so that it can respond with speed to any customer demand, opportunity or threat.

Despite the obvious benefits of such business and technological alignment, IBM's vision for the future is definitely running into its share of naysayers. A recent IDC report that made predictions for 2004 referred to utility computing as "Futility Computing." According to IDC, excitement over the idea is based on a lot of marketing hype and that significant investments in the technology will remain modest.

Well, the same can be said for all great visions when they are in their infancy. Surely, there wasn't significant investment in the Internet back in 1990. But today, the vast majority of office workers can't imagine a day of work without Internet connectivity. Give people time to learn about the value of on-demand and investments will eventually follow.

Other critics say that IBM has gotten too far ahead of itself by rolling out an all-encompassing and extremely confusing concept for the future that most people have yet to grasp. This is simply another manifestation of impatience. It is the job of a visionary to shake things up, to look at the way things are and to find a better way, to be a few steps ahead of everyone else. It takes time for their vision to trickle down through the rest of society, but it happens.

Heck, just a couple of years ago most people were still vague on the concept of Web services. Today, you'd be hard pressed to find a large company that hasn't made major investments in the technology.

It's true that IBM still has plenty of work to do when it comes to driving home its vision for users. But no one company is better poised to do so.

IBM's competitors are playing around with different facets of on-demand/utility computing, but not even one of them is in more than three of IBM's six technology businesses. Big Blue's wide technological reach makes them the natural leader of the on demand world, and in time we'll see their vision come to fruition.


Not Utopia

By Michelle Davidson, Site Editor

Talk to almost anyone at IBM about on-demand computing, and you'll think IT Utopia is waiting for you once you implement it. The problem is that Utopia comes at a cost and few people are rushing to adopt it.

On-demand computing is a strategy intended to deliver computing resources the way a power utility doles out electricity. When a network has a surge in demand, intuitive architecture triggers other resources into action, including idle servers, applications or pools of network storage. Companies pay only for the amount of time they use the services.

Sounds great, doesn't it? So why aren't users rushing to adopt? Why is it destined to be just an idea and nothing more?

For one reason, users don't understand what it really is. IBM is pushing this new technology without fully explaining what it's all about. Ask any IBMer about on-demand computing, and they'll say it's crucial to users and that they'll reap enormous benefits from it. Then ask a user what he or she thinks, and more than likely you'll get a puzzled look. You explain what IBM has told you, and they say, "Oh, yeah. I heard about that. No, we're not looking into that now."

Even people within IBM seem confused and give conflicting descriptions of what on-demand computing is. They all recite the same mantra of how wonderful an on-demand world will be, but their description of what exactly on-demand computing is varies depending on whom you talk to and their perspective.

Users are also wary of what IBM says and does, especially their pricing structures. iSeries users know this all too well. Allowing IBM to bill for resources automatically triggered based on demand has users comparing it to running over minutes on their cell phone bill.

Not too long ago, IBM decided to separate batch and interactive CPW on the iSeries and then charge a high premium for the interactive feature. But users say IBM put an artificial governor on the processor that essentially brought down the system if the interactive threshold exceeded capacity. At that point they had to rewrite the application to use less interactive processing power or pay big bucks to IBM to let their processor do more interactive work.

A lot of people were burned by that, and they say IBM's processing on-demand resembles the interactive feature. The processors are already in your machine, but you pay IBM to allow them to be activated.

Cost is also an issue when implementing on-demand computing. When the new iSeries line was announced a year ago, the daily rate for capacity on-demand was $1,100 for a Power4 processor in a Model 825, $1,200 for a processor in a Model 870 and $1,300 for a processor in a Model 890. Few users rushed in.

IBM knows price is an issue, and just recently it slashed the daily rate in half to $550 per day in a Model 825, $600 per day in a Model 870 and $650 for the Model 890. Sure, it sounds like a good deal, but it isn't enough. These models are all high-end servers, with prices ranging from $80,000 to $2.5 million. You won't see many small and midsized businesses buying these.

Even if a company does purchase one of those machines, the cost of using their on-demand capabilities probably isn't worth it. One recent report from Midrange Server's Timothy Prickett Morgan said, "Assuming that the cost of a processor is amortized over three years (which is the economic life of a computer), an iSeries processor activation should cost around $50 per day in this power class of machines. IBM is still off by more than an order of magnitude. And even if you assume that an iSeries loses all of its value in one year, processor activation is still, after these price cuts, several times more expensive than you'd expect, based on the actual cost of the machines."

For all of IBM's talk of on-demand computing, that is all it ever will be. It is not an initiative that will take off among its users.

Dig Deeper on Autonomic Computing

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.