Those working in IT operations tend to see themselves as “heroes” because they keep enterprise networks and applications going by tackling day-to-day problems, Haight said in his keynote address at the Gartner IT Infrastructure & Operations Summit in Orlando, Florida.
However, those in IT operations continually battle with those on the development side of the house, whom they think of as “the ones creating all the fires they have to put out,” he added.
The more painful reality is that IT operations can be seen as “hostage takers,” said Haight, because they will push off changes they’re asked to consider, redirecting tasks to committees with lengthy lists where new ideas are simply added to the bottom. Thus the term “deployment-prevention committee,” said Haight.
But the old DevOps battles between development and operations need to come to an end through creation of more small teams, says Haight. “Small teams are agile teams,” he said, noting operations has to work with development to shorten release cycles and lower costs.
“The infrastructure is now software. We’re using configuration metadata to match the velocity of our infrastructure customers.” This means configuration data should be treated carefully, like any other kind of important code, and perhaps stored in repositories,” Haight said.
“And get rid of the dogma. Have that sign say, ‘We’re open. We’re open to new ideas,'” Haight told his audience of hundreds of IT operations managers attending the event to hear Gartner analysis related to new technologies.
Some of the newer ideas, such as fabric-based computing for storage, servers and networking, usher in the distinct possibility that no longer will the same number of operations and administrative staff be needed to keep an enterprise network going. That’s because fabric-based computing works to increase automation of functions, relying less on manual changes.
During his presentation, Gartner analyst Carl Clauch acknowledged that if the technology is widely adopted in the enterprise, it should lead to reduced operations staff.
“Yes, that is the trajectory,” he said. He added that many organisations, which are generally hiring staff now, won’t necessarily reduce staff in the absolute because they will be able to redeploy them for other things. “But you could take some low-value labor off your environment,” he said.
Reducing the cost of human labor is certainly one reason for deploying fabric-based computing, which he said allows the enterprise network to be set up not in the old way that’s “pre-configured, fixed and bound” for separate storage and network servers, but instead as pools of processors, memory, I/O connections, adapters, network connections and storage that are bought individually, combined on the fly and managed via a “Fabric Resource Pool Manager.”
The idea is “continually reshaping the data centre” in as automated way as possible to support any private and public cloud that’s demanded by circumstances. Vendors supplying fabric-based computing components include Oracle, Cisco, HP, Microsoft, IBM, Del and VMware.
There are a lot of questions to ask before deploying fabric-based computing, such as how much it’s possible to mix-and-match vendor products versus selecting a strategic vendor, Clauch said.
There are worries about vendor lock-in. But Clauch said Gartner clientele going the fabric route are reporting back that it’s working for them. One of the biggest advantages it brings is “time to provision,” he said. “People are reporting dramatic improvements” with tasks that used to take months now only taking a few hours or days.
If there’s anyone who gets a new job in fabric-based computing, it might be the security analyst. That’s because fabrics will radically alter more traditional notions about IT security.
Ellen Messmer is senior editor at Network World, an IDG publication and website, where she covers news and technology trends related to information security.