Gartner prognosticators looked into their crystal balls this week to point out the top data center technologies
and trends for the coming year. I made my own list based on the most popular stories on SearchDataCenter.com and my own secret sauce and here's what I came up with:
Coming in at number one -- data center pros are looking for metrics
. You can't manage what you can't measure. Systems like the Uptime Institute's Tier Ratings
, Syska Hennessy's Criticality Levels
, server energy efficiency metrics from the EPA
, cooling guidelines from ASHRAE
and a continuing interest in ITIL will be in demand for 2007. The tools to measure data center performance are out there, but it takes digging and some knowledge to use them.
Number two -- people are talking about Big Iron
again. Having trouble with power and cooling issues? Configuration nightmares? Who the hell let the Wintel admins start running the asylum anyway? Mature processes, new open system workloads, cheaper hardware and a smaller power and cooling footprint are all having an effect on mainframe buying decisions -- sales are booming right now. And it's time to take a look at your staffing requirements if you're executing millions of lines of COBOL for the foreseeable future.
Number three -- blades
are hot and heavy. They're new, tiny and look a lot better than that old pizza box. It sounds like a midlife crisis, but data center pros are learning to deal with these high-maintenance lovelies. A few guys got burned or had a rack bust through the raised floor, but people have learned you can't treat them like any old server. Bottom line: isolation is key. Look for new power and cooling infrastructure products that accommodate and isolate blades in the coming year.
Number four -- Configuration management databases (CMDB)
take over the systems management space. Dave Wagner at BMC put it best this week during a conference at the Gartner Data Center expo. He explained why the need for a CMDB has reached a tipping point. "Proprietary legacy platforms all had very rich process disciplines around change. That was because if you brought down an iSeries, you brought down a whole company. Mainframes, AS/400s, VAX systems didn't have the automation because the scale was easy enough to manage manually." So in comes the age of pizza boxes and it's one app to one server and if something goes FUBAR, you just buy another one because it's cheaper to replace it than to pay somebody to figure out where the problem is. But here comes the tipping point -- virtualization. Running multiple apps on the same physical infrastructure, if you break it you affect more people. "Everybody knows 80% of all failure is because you changed something. Virtual environments are dynamically changing by nature." Hence, everybody rolls out a CMDB. According to data from audience participation at Gartner, HP's Mercury line and BMC's Atrium are leading the pack.
Number five -- Everybody makes nice with the utilities
. Want to save hundreds of thousands of dollars on energy costs but don't want to get crushed on the front end cap-ex for the upgrades? Call the electric company. Utilities have historically offered commercial customers rebates for avoiding demand spikes, but now power companies like PG&E and Avista have teamed up with hardware and software vendors like Sun Microsystems, VMware and ISR/SprayCool to offer prescriptive rebates to offset upgrade costs. Execs from IBM and Emerson/Liebert confirmed that both companies (along with many others) are already in talks for more announcements like this. Other utilities across the country are looking to get involved. Get in now while it's still a free-for-all.
Number six -- Liquid Cooling
. You can't fight physics folks. Nobody wanted to bring water back into the data center, especially not the systems vendors, but that's just too damn bad. It's here. SearchDataCenter.com called it back in April 2005
when cooling guru Bob McFarlane said: "While everyone knows servers are one day going to be water-cooled, no one wants to be first, believing that if their competitors still claim they are fine with air cooling, the guy who goes to water cooling will rapidly drop back in sales until others admit it is necessary." Who's in? Companies like Target and Goldman Sachs are preparing facilities for it. Companies like APC, Liebert, ISR/SprayCool, IBM, HP, and finally Sun Microsystems are making water-cooling products. And everybody has a different take on it, which is the main factor holding it back from acceptance. Once ASHRAE's book on liquid cooling
technologies comes out this month, we might start to see some standardization and more acceptance.
Number seven -- Site selection
continues to be big business and the data center construction boom is officially on. Web2.0 Companies are outgrowing their facilities faster than they can build them. The build out of the 1990s is coming to the end of the 10- to 15-year facility turnover cycle. Companies are looking to consolidate multiple outlying, outdated facilities. The price of power matters more and the geography matters less -- look for more companies to move IT operations to the Midwest in '07.
Number eight -- Green computing
continues to sell. The economics of "eco" took the facility management and construction industry by storm back around 2003. Why? Because the U.S. Green Buildings Council
started to prove that companies could save money building high-performance green buildings, and vendors could make money selling greener products. It's taken a while for IT to catch up, but companies HP and Sun are now competing on miles per gallon, as opposed to miles per hour. The IT industry has formed a group called The Green Grid
-- a bunch of vendors hoping to solve the problems they created. But if it can appoint highly respected people into leadership positions, expect the organization to play a dominant role in the definition of green computing in the coming years. Green Grid doesn't need Paul Perez (HP), Dave Douglas (Sun), or Bruce Shaw (AMD)out front, they need a spokesperson like Bill McDonough
. If Green Grid plays out as an industry mouthpiece, expect it to be ignored and go away. But the issues will not -- someone is going to fill this vacuumm in 2007. Also, expect legislation around energy usage and e-waste.
Number nine -- multicore processing
. Here's one where I agree with Gartner's list. The firm predicts the number of cores on a chip will double approximately every 18 months through 2015. But those gains won't mean much unless users address software. "We're entering a software crisis," research VP Carl Claunch said. "If your software runs one big workload, you need to rearchitect it to be parallel. There is a challenge to make our apps multithreaded."
Number ten -- virtualization and disaster recovery
. According to a recent reader survey from TechTarget, most people are using server virtualization today for test and development sandboxes and consolidating servers, but very few are thinking about disaster recovery. The value of the portable virtual machine has yet to be realized. Time to recovery is going to shrink dramatically when companies can start deploying failover virtually. Disaster recovery giant SunGard says it can shrink downtime in a disaster from 48 hours with tape transport and infrastrucutre set up to under four hours by blowing out virtual machine files to already up and running hardware.
For more info, check out our top ten stories of 2006
based on page views.