As more large data centers move to converged infrastructure — which is offered by Cisco, Hewlett-Packard, Oracle and other vendors — most of the chatter has been about this slick, new hardware and how it melds compute, networking and storage capability into fewer boxes that are more powerful yet more efficient.
Less has been said about what that trend means for enterprise data center staffing, recruiting and hiring. Suffice it to say, it means a lot, both for the employees themselves and the CIOs and managers that must retrain them — or do away with them altogether.
CIOs and other IT managers, of course, like the notion of a converged infrastructure because it saves on power, wiring and even real estate. Cisco claimed 10,000 UCS customers by early this year, up from 4,000 a year ago but there are not a ton of independent statistics comparing sales of converged infrastructure to sales of regular servers, storage or networking products. However, a new survey of Fortune 1000 IT professionals who deal with servers found that one-third (32 percent) of the 169 respondents have such equipment in pilot now or plan to buy it. Most of that interest in converged infrastructure comes from large companies, said analyst Peter Ffoulkes with TheInfoPro, a division of the 451 Group, which conducted the research (see chart below).
Chuck Hollis, the global marketing CTO for EMC, said converged infrastructure isn’t for every account, but there are enough large enterprises in the market for it to keep vendors like EMC — which is a player with the vBlock bundles of Cisco UCS, EMC storage and VMware virtualization — busy.
Converged infrastructure meets IT resistance
For the last decade or so, there have been three key constituencies in the data center: the server person, the networking person and the storage person. While these administrators didn’t necessarily sign the purchase orders, they definitely held power when it came to which brand and model of equipment was bought. And once that happened, the chosen piece of gear was their baby to care for and maintain. In short, they had deep knowledge of their particular piece of the IT infrastructure puzzle but little insight into the other pieces.
Now the need is growing for IT pros who understand more about more things as opposed to a whole ton about one thing. And the need to field data center specialists in the three tech silos is on the wane. If, for example, a company buys a vBlock or a FlexPod (melding NetApp storage with UCS and VMware), someone higher up the IT stack is signing off on that purchase, and the resulting vBlock will certainly not need three IT pros to maintain it.
“As you converge you get rid of silos and that has tech pros worried,” said Sam Barnett, the directing analyst for Infonetics Research. “No one wants to engineer themselves out of a job.”
Higher up the food chain, CIOs have to figure out how this trend is changing their staffing requirements and how to handle retraining for the administrators who are suddenly without a silo to call their own.
“Hiring managers are already looking for more diversified backgrounds. They want more of a jack-of-all trades than a storage or a LAN specialist,” said Bill Kleyman, analyst with IT consultancy MTM Technologies.
To be clear, this need for retraining to understand more about process and workflow as opposed to the niceties of X server, Y router or Z storage array is a large-enterprise phenomenon. IT pros in smaller companies already tend to be cross-trained across the silos. In a 200-person company that runs a couple of servers, a handful of IT people must know enough about servers and storage and networking to keep the whole shmear running; there is just not as much specialization or siloing in those shops.
But in data centers of all sizes going forward, “broad is better than deep,” said EMC’s Hollis, who spends a lot of time talking to customers about this topic. “There will still be a need for specialists, just not as much need.” He says data center people have to find a more holistic view of how things work and how to translate that into IT services for their end users.
Hollis uses EMC’s own experience as a guide for its enterprise customers. As the company redesigned its own IT infrastructure to serve some 50,000 employees, the data center head count stayed stable or rose slightly, but the mix of skills changed drastically as people retrained to take on new responsibilities.
EMC wanted to set up its IT to act more like an internal service provider — more like a vendor selling services into the company rather than a part of the company itself. “We had set up services that are consumed, which means IT has to focus on service and delivery versus handcrafting every piece of IT,” he said.
The result? “We found that 85 percent of the roles in our IT organization changed significantly in the last 18 months. I don’t even want to think of what that other 15 percent is doing.”
The people who were configuring servers for several hours per week are no longer doing that but are building or specifying the IT services needed by various departments. Someone who might have installed or maintained servers, now, with additional training, is building services needed by people in the business units. There is definitely a cloud inflection to all of this as companies move forward. Nearly everyone needs to gain a working knowledge of virtualization at the very least, if they haven’t already.
It is all about going from reactive to proactive mode for IT. EMC IT people who used to configure and provision servers now perform capacity planning and IT process engineering to support their users, for example. Those who responded to help-desk tickets have morphed into IT pros who develop and deliver actual services to users. The difference is that the old-school IT people were always in catch-up mode, dealing with issues (breakdowns, snafus), while their new roles are about creating services and planning for new capacity needs in advance. The assumption here is that the new converged, heavily virtualized infrastructure just works better and is more self-tuning and self-healing than its predecessor — and that means more free time for the people who used to tune and fix it. Whether that means they are out of a job or need to be retooled for a new job is up to them and their IT management.
The more things change . . .
People forget that even in the staid world of data centers, things continually evolve: There is nothing new about requiring IT guys to change.
“Ten or so years ago, when you wanted to talk network, you had to talk to the telephone PBX guys. Then the PBX guys didn’t know IP, so we had to start a network group, then the phone became IP so we didn’t have PBX guys anymore,” said Paul Santinelli, who once ran data centers for Red Hat and now invests in data center technologies as a general partner in North Bridge Venture Partners. “[IT] people have to change as the workloads and data center technology changes.”
Santinelli doesn’t think jobs will be lost, provided the people involved are prepared to learn new skills. “There’s this misperception out there that when you virtualize things [as happens in the converged infrastructure world] you cut people because there are less machines, but in fact you often increase people because there are more VMs atop any given piece of hardware.”
Santinelli’s argument is that as the base level of technology rises, IT folks who learn skills in the “white space” above that tech level will be just fine.
Word to the wise: Train up
But to be sure, not everyone agrees that these jobs are safe.
Greg Shields, a partner and technologist for Concentrated Technology, an IT consulting firm, thinks the mid-range data center admins are in the bull’s-eye.
“When you move to converged infrastructure, you still need to have your hardware monkeys — low-level people who rack-and-stack boxes and you still need the high-end architects,” he said. But all those mid-level sysadmins? The guys who manage servers and storage and routers? Not so much. Most of this converged infrastructure comes with automation that takes on a lot of the work these sysadmins do, Shields said.
Jobs may or may not be lost in this transition, but IT professionals who are not willing to learn the new skills required are definitely in danger of losing their jobs.
The net takeaway is: If you are a server admin, a storage admin or a LAN admin, it is time to bone up on how your piece of the IT puzzle works with the others and how the entire system can deliver IT services that end users want.
Vendors that hope to cash in on this converged infrastructure boom are offering training in new skills to help these administrators learn new skills to move up the food chain.
A big part of this is the need to reshape the mindset of data center professionals. In the past, IT departments at big companies were often more a logjam than an asset. Business users requested things to help them do their jobs better and then waited for IT to get around to doing them.
What converged infrastructure — and its automation — facilitates is the notion of IT as a service organization. It is there to do what needs to be done to help the business units, and long waits are not an option. The smart IT shop will pitch services to the business leaders before being asked. In this new model, for example, an IT shop would have noted the appearance of iPads in the building and rushed to accommodate their use on corporate networks rather than waited for C-level execs to mandate they do so.
As more of the data center is automated and self-managed, IT management guys will be freed up to perform service functions like the above example. That change of mindset might be harder than any technical retraining that must be done.
It also means that the folks who knew about servers and wiring need to know a lot more about the businesses they are serving. Smart CIOs will encourage the sort of education and training — both on the technology and business sides — to ease their path.
“If the sum total of your contribution to your business IT infrastructure is clicking the ‘next’ button, then it’s time for you to move on anyway,” Shields said. “There will be job loss. When we had the dot.com bust, a lot of IT pros lost their jobs, but many of them probably shouldn’t have been in IT in the first place.”
Available to GigaOm Research Subscribers