If you’ve been busy virtualizing your environment, or migrating to an internal cloud, chances are your software licensing and maintenance costs are spiraling out of control.
Every CIO wants to balance costs and maximize efficiency, but the shift to virtual and cloud hosting models has changed the game when it comes to software licensing, and many IT organizations are playing under old rules that no longer apply.
“Traditionally, asset management has been responsible for a lot of these costs and the operational groups were less concerned about it,” says Andrew Hillier, CTO and co-founder of Richmond Hill, Ontario-based CiRBA, a specialist in capacity transformation and control systems for virtual and cloud infrastructure. “But if you’re building an internal cloud, all of a sudden this becomes your problem.”
Hillier says CIOs need to be aware of five common misconceptions of software licensing that can prevent significant savings in the data center.
Misconception 1: ‘I License Each Software Instance, So Virtualization Doesn’t Impact This’
“One misconception is that if you license per instance, when you move to the cloud it doesn’t change,” Hillier says. “They think it’s going to be one for one.”
But moving to a virtual or cloud environment opens the possibility of moving to a per-host model, where licensing an entire physical host server allows an unlimited number of instances to be run. That seems good on paper, but the truth, Hillier says, is that it translates into savings only if you can control virtual machine (VM) placement and maximize VM infrastructure. Essentially, he says, if you can defrag the environment to fit the expensive components on a subset of your infrastructure, you can realize huge savings.
“The licensing doesn’t change unless you go to a per-host model,” Hillier says. “Then you can radically change it.”
Misconception 2: ‘I’m Not Near a Renewal, So There Is No Point in Looking Into It’
“A lot of people are on a three-year contract or multi-year contract,” Hillier says. “If you’re locked in, you might say ‘what’s the point of looking into this?’ But the fact is, even if you’re not near renewal, it’s still advantageous to save on the maintenance. Maintenance is extremely expensive; just on the maintenance the savings can be huge in these environments.”
Just carving down the maintenance on databases can save millions of dollars, he says.
Misconception 3: ‘I Have to License the Entire Virtual Cluster, So It Is Difficult to Cut Costs’
“People think that if you’re going to run software, you have to license the whole cluster,” Hillier says. “If you don’t have any controls in place and can’t control where it runs, then yes, you have to license the whole cluster. But you only have to do that if you have no way of controlling where the VMs go. If you do, you can license just the piece it’s going to hit.”
It really comes down to having the proper management tools, he says. There are many constraints that govern which VMs can and should go together. These constraints include utilization , type of workload, SLAs, compliance, technical compatibility and security. Even in static situations, figuring out how to maximize density can prove difficult. And virtual environments add the complexity of workload mobility and rapid change.
Given these challenges, it’s no surprise that organizations often throw up their hands, Hillier says. But the right management tools can make all the difference.
Misconception 4: ‘I Get Such Big Discounts on Software That Optimizing It Won’t Save Much’
Hillier says this is a big mistake. He says CiRBA looked at 18 virtual environments of organizations with more than 1,100 physical servers. Simply by optimizing VM placement and VM density, those organizations were able to achieve cost savings of 55.6 percent on average (the majority of environments experienced savings between 40 percent and 70 percent, he says).
“If you take an environment that’s basically random and then concentrate your databases and OSes, you can save 56 percent over no optimization,” Hillier says. “If you paid list price, just on eight servers worth of VMs you could save almost a half million dollars. If you paid a tenth of that cost, you can get the same savings off of 80 servers instead of eight. You’re still going to save a lot by looking at this. Even if you get a 99 percent discount, it’s something you need to look at.”
“If I’m the consumer and I go to external cloud, obviously that simplifies things,” Hillier says. “I just pay a single fee to say, IBM, for the software. But on internal cloud, if you’re the one that runs the infrastructure, it doesn’t necessarily simplify things. You need analytics to figure it out. We find that some of the groups building out the cloud often inadvertently become the asset managers. People can accidentally take on the role of asset management without intending to.”
That’s not a reason to avoid a migration to a virtual or internal cloud environment, he says. After all, there’s a great deal of agility and flexibility to be gained. But, he says, IT organizations taking the leap should be aware that these environments do create higher complexity than the older environments they’re replacing.
“We’re finding the shift to cloud really does require another level of analytics beyond what would traditionally be done with trending and charts,” he says. “The technology enables benefits, but you won’t get them unless you’re clever about how you use that environment. When we look at these environments, we consistently see huge optimization potential.”