Here's a sentence that didn't make it into the final draft of my "Red Shift" feature
, which leads off this week's print issue: "Also gaining currency among a wide swathe of CIOs faced with insatiable computing demands, the red shift concept promises (or threatens, depending on our point of view) to join 'tipping point' and 'long tail' in the business-buzzword pantheon." So is the red shift just a deft bit of Sun propaganda, or a deep insight into the future of computing demand?My answer in the feature, which I'm sticking with here, is "A little of both." You can't swing a cat by the tail at a CIO conference these days and not hit somebody who's facing the twin and apparently contradictory problems of low server utilization and skyrocketing demand for data processing and storage from key applications. Commenting on the story, "RET" put it well: "We are absolutely not running out of processing power. We are running out of space, time, energy, and money, of course. A big problem is that we chase Moore's Law for processors, but have not invested as much effort in understanding utilization ratios of various components in the machines in order to truly deliver a processing platform that is evenly utilized on multiple fronts."
At the same time, Sun's CTO Greg Papadopoulos is making an intellectual leap from the "red shift" to Sun's "big iron" theory that soaring demand for CPU cycles, and the attendant costs in real estate, power, and cooling systems for massive data centers, will lead to a surge in utility computing based on big Sun servers. This I, and the sources I talked to in the IT industry, don't necessarily buy. Here's another quote that didn't make the final cut:
"Sun believes in its high-end computers designed to do massive Web-scale applications very efficiently," says Nicholas Carr, former executive editor of the Harvard Business Review and author of Does IT Matter? , "but on the other hand if you look at the current leaders [in utility computing], Google and Amazon, they seem to be going down the path of cheap boxes and Linux open-source software. There's a big question mark as to what the future of the high-end, red-shifted market is going to look like."
So is Papadopoulos really onto something here, or is his confidence misplaced? Post your comments below and I'll tally the results in a subsequent post.