BOSTON - The most vexing questions created by the introduction of hyper-converged infrastructure in the enterprise...
may end up being more about IT staffing and less about the technology.
Many in IT realize there could be cost savings by moving to hyper-converged infrastructure, but not without finding a way to work with current IT staff.
"What implications does it have for an enterprise?" said Donald Campbell, infrastructure architect for global banking company National Australia Group, during the Red Hat Summit here this week. "We used to have guys that did everything, then we went to specialists, and now we are going back."
IT is a cyclical industry and Campbell said he understands there will be changes but wants to make sure it won't be just two years before he will again feel pressure to move toward specialists. IT pros were told at a Red Hat Summit session that hyper-converged infrastructure will allow them to form single infrastructure teams.
Several years ago, there were systems administrators who handled networking and storage and like many enterprises, the company created specialist positions.
"What do we do with the specialists?" he said. "My questions are about the implications of the technology on the skillsets, the processes and the politics."
Enterprises that need simplified storage management are drawn to hyper-converged infrastructure to manage it at the VM level. National Australia Group used local disk storage for years, Campbell said.
"Over the years we have moved away from [local disk storage] to the SAN environment, and one of the things I am trying to figure out is if we went away from local disk to SAN, why are we going back?" he said.
Hyper-converged infrastructure is a software-centric design that combines a hypervisor, compute and storage into a small footprint that is scalable and highly automated. Red Hat's system combines Linux, oVirt virtualization and GlusterFS storage.
Some Red Hat customers have put together a "home-grown" hyper-converged setup using open source components, said Sean Murphy, a principal product manager for storage at Red Hat.
"We're seeing that ramping up over the past couple of years," he said.
The hyper-converged market grew 162% in 2014 and is forecasted to grow at about 100% this year, according to IDC. Within two years, it is forecasted that 50% of enterprises will be deploy VMs on a hyper-converged infrastructure.
Paul Cuzner, a senior architect at Red Hat, has been "shaking [Gluster storage] to see what works and what needs a change."
Gluster ensures the storage and compute environments don't "trample each other," Cuzner said, while making security a top priority.
"One of the key things for us is when you open the box you don't end up with a scenario on day one where one VM can basically rules the roost and take all the bandwidth and take away the back end of the environment," he said.
Interest in hyper-convergence is across most market segments, from SMBs to large enterprises, Murphy said.
"The interest is all around the simplicity – procurement, planning, deployment and ease of use," Murphy said. "These things resonate with everybody, regardless of what market you are in."