kentoh - Fotolia

Hyper-converged infrastructure architecture changes IT admin role

IT leaders who have adopted a hyper-converged architecture say it streamlines costs and eliminates silos and finger-pointing, but also brings a fear of the unknown to their staffs.

A selling point of hyper-converged infrastructure architecture is that it alleviates the need for storage, compute...

and hypervisor specialists. CIOs and admins appreciate that, but the specialists aren't always happy about it.

IT staffing received a lot of attention last month during the Nutanix .NEXT user conference that brought together a horde of hyper-converged IT customers. CIOs and admins talked about how adopting a hyper-converged infrastructure architecture reduces the need for specialization, streamlines costs, eliminates blame between teams and allows them to better deploy staff. But while the elimination of siloed roles makes management happy, the change in status quo also makes IT employees nervous about long-term career changes.

A hyper-converged infrastructure architecture combines storage, compute and virtualization into one system and usually one team to manage it.

"Like with the cloud, people get scared," said Lawrence Lozzano, senior database administrator for Los Angeles law firm Sheppard, Mullin, Richter & Hampton. "They think: 'Things will be more simple and easier to manage, so what am I going to do every day if things just work? Will my job go away?' At the same time, it gives them an opportunity to learn a new skill set and be current."

Tim Crawford, a CIO strategic adviser for Avoa, said he senses the same thing from IT staffs with which he works.

"It's scary for a lot of IT people," he said. "The biggest challenge [to hyper-convergence] I see is how it affects the culture and the people."

The biggest change in IT admin roles with hyper-convergence comes down to specialization: There is a lot less of it. A hyper-converged infrastructure architecture often eliminates the need for distinct storage, server, virtualization and, sometimes, network specialists. And that may not help those particular specialists sleep well at night.

"The people change is always the hardest part" of IT change, said Rob Catterton, director of data systems for online coupon company Valpak in St. Petersburg, Fla.

But admins who embrace hyper-convergence said the new architecture moved their IT brainpower from the data center into the boardroom.

"I can now take smart people that were heads-down carving storage and plug them into meetings with other people to share their knowledge," Catterton said. "They can talk about business needs."

Goodbye gurus?

Lozzano said, by eliminating the need for specialists, a hyper-converged IT architecture can save money and streamline management.

"You don't need to hire one guy who's a hypervisor guy, another guy who's a storage guy and another guy who's a server guy," he said. "The old-school model would be, 'One person's a hypervisor person. That person's going to go on vacation, so I need two hypervisor people. And I have to hire two storage people.'"

Now, he said, "You don't need to be an expert on virtualization or storage; you don't need to be a guru. You can be a jack-of-all-trades, because it's simple and easier to manage."

Of course, that doesn't mean there is nothing left to manage.

"Even though we're consolidating and converging silos, there's still plenty for everybody to do," said Joshua Lukes, senior manager of computing services for technology provider Itron, based in Liberty Lake, Wash. "It's just more exciting than the standard mundane tasks they'd been doing in the past."

Besides streamlining IT admin roles, hyper-converged adopters said they also find it easier to deal with a single vendor than separate storage, server and hypervisor companies.

"We're collapsing everything into one architecture," said Alton Levesque, IT strategic director of enterprise platform engineering at healthcare company LabCorp. "Now, instead of needing network and storage engineers and Cisco, VMware and EMC all on the call to figure out why I have this latency issue, or this SCSI problem, I'm able to take a small team and add greater value.

"I can stop focusing on storage upgrades, compatibility matrixes, storage operating systems and drivers, and focus on, 'How do we automate workloads? How do we support our Agile teams and their high turnaround need?' Lifecycle management used to be very time-consuming, and we don't have to worry about that anymore. Let's find the true value we can add to the overall business model."

Hyper-convergence equals fewer fingers pointing

Hyper-converged infrastructure architecture adopters said the elimination of silos means less blame getting thrown around internally, as well as among vendors.

"The challenge I had before was the storage guys are complaining to the processing guys, the processing guys are complaining to the network guys, and anytime there was an issue, there was finger-pointing," said Bob Orkis, CIO of Fairway Independent Mortgage Corp in Madison, Wis.

Orkis said his first step to stop that was to buy an EMC-Cisco Vblock converged infrastructure system for "one throat to choke." Then, in 2014, he switched to Nutanix and found hyper-convergence less expensive than converged infrastructure. He said the converged roles have made it easier to expand to 157 Nutanix nodes.

Actually, there is literally nothing I miss about legacy architecture.
Matthew FrewIT director, Wharton School at the University of Pennsylvania

"We were able to educate folks who were not server folks or storage folks on how to manage it," he said. "In the three-tier environment we had before, we had to have a specialist in each specialty. And they're not cheap. It was amazing what kind of staffing expenses we would have had to support the three-tier architecture."

When asked if he misses anything about traditional three-tier infrastructure, Matthew Frew, IT director for the Wharton School at the University of Pennsylvania, said sarcastically, "I miss the finger-pointing between vendors.

"Actually," he added, "there is literally nothing I miss about legacy architecture."

Frew said finger-pointing at his school came to a head during an issue with Dell EqualLogic SAN arrays, Dell servers and VMware hypervisors before Wharton switched to Dell XC Series appliances running Nutanix software. The trouble came when hypervisors lost connectivity to the storage array.

"EqualLogic said, 'It's your hypervisor.' VMware said, 'It's your storage,'" Frew said. "We got them both on the line, and they were pointing fingers at each other and gathering logs. After two weeks of troubleshooting, they said, 'Oh, it's your networking.' And it was Dell networking. So, we had to start yet another support call with Dell networking."

He said since switching to XC, all support goes through one Dell team, even for Nutanix or VMware issues.

Frew also said using a hyper-converged infrastructure architecture allows him to deploy staff in a smarter fashion.

"We still have same responsibilities, but it allows us to dedicate more time to refreshing services we're already running, getting them up to date," he said. "And more time to deploy more new services."

Next Steps

How storage fits into a data center converged infrastructure

Know the key features of an HCI appliance

SMBs can get an edge with hyper-converged implementation

Dig Deeper on Converged Infrastructure Management