Most people usually make most of their predications for next year at year’s end. This year, however, there are so many game changers this year I wanted to make my forecasts early in time to be incorporated into our plans and projects scheduled for 2011.
Here, then, are my six game changers for 2011:
In-row cooling (IRC): In the row, overhead, on the rear door, or any combination thereof, IRC options abound. Even if you still want to keep chilled water out of your data center, there are options as all these systems also come in refrigerant-based versions. The economy has held up many major data centers projects, making the low-capital investments like IRC solutions more attractive as blades and virtualized servers migrate into data centers. Data center owners have begun to embrace IRCs as efficient, reliable, and economical solutions for moderate to high density cooling.
Containment systems: After IRCs, containment systems are the next level in the drive for energy efficiency. You can build your own containment system or buy everything from freezer strops to fully engineered roofs and doors. Two vendors have even developed plenums for the rear of the cabinets that direct 100 percent of the hot air directly back to the IRC. Yet another design adds a cold-aisle plenum to create a 100 percent-contained rack row without concerns on lighting or sprinkler coverage.
Direct water-based cooling: It’s back! Trying to reclaim heat from data centers has always frustrated everyone who has tried. Now IBM is introducing a water-cooling system to the chip level that removes a large percentage of the heat from computers just as in the old mainframe days. Companies like Hardcore Computer are going one better and “submersing” the entire set of electronics in a circulating fluid bath that removes 100 percent of the heat. Neither of these technologies is new. They are just new adaptations of proven technologies. IBM started the water to the chip concept in the days of the mainframe, and every electrical utility transformer serving commercial buildings makes use of fluid-based submersion concept. Soon these solutions will be found wherever return temperatures are extremely high, and heat reclamation will begin to make economical sense.
Imagine the day where the heat from a computer is sufficiently hot to drive an absorption chiller that produces the cooling for the computer. We are moving in that direction.
Solid-state drives: Combining fluid- or water-cooled processors with new solid-state data storage will bring a precipitous drop in data center power requirements. But many have said the same about earlier technology breakthroughs only to see the savings consumed by increased demand. Though this cycle may repeat, just think how these technologies will then mean for information processing? The resources and capabilities that can then be applied to increased knowledge will take giant steps forward. Will life-like Androids be far behind?
Cloud computing, Software as a Service (SaaS), and Colocation: Reports of the death of the enterprise data center are quite premature. Cloud, SaaS, and Colocation have merely stepped in to meet demand that the enterprise model could not. Even these powerful new models are coming up short as growing pains hold them back. The cloud continues to have security and integrity issues that are being addressed as the industry rapidly enters its adolescent stage. As it matures it will come into its own and command respect as a valuable alternative for select needs.
SaaS is in much the same state. These services are evolving, and users are getting more familiar with this alternative to individual user license models. Buying what you need when you need it is a powerful motivator so long as the service provider remains competitive and doesn’t offer a low-entry fee just to lock customers into higher fees later. This model requires a lot of trust between provider and user and trust takes time to build.
There are now over a 1,000 colocation firms nationwide, and many are selling out in less than 12 months (see the table). Their services range from ready-to-use space to fully managed services
• Pre-engineered data centers, containers, and modular construction: The full custom build it on-site concept must go. It is inefficient, riddled with quality and integration issues that take years to ferret out. Enter the age of the Henry Ford assembly line for data centers. Like cars of today, there are already numerous models and manufacturers to choose from and there are even more custom shops that can add that extra edge in performance. The need for speed to meet the market delivery demands will continue to push this market forward. With the low cost of entry to this market the growth of suppliers will expand exponentially.
That’s it folks. 2011 will be a banner year for change and with change significant growth in all the new fields. Embrace it or reject it, either way the wave is coming our way. We either catch it for the ride of a lifetime or drown in its undertow.
If you have other game changers let’s hear from you. Email us at: firstname.lastname@example.org.