Whether you encounter them in the C-level suite, in IT or in some other line of business, contrarians are fun. They excel at trying to help you see the way the world is and their perspective is the right one, facts notwithstanding. Hey, back in the day if contrarians had had their way, Columbus’ three ships – the Nina, Pinta and Santa Maria – would still be in dry dock. Contemporaneously, NASA would never have gotten its rockets out of orbit, much less be able to land a man on the moon.
Still, even contrarians are not beyond achieving a fresh outlook.
Which brings me to this excellent post by Ryan Huang on some common, albeit misguided reasons, for SMBs and enterprises alike not considering public cloud as articulated by Jason Cumberland, a vice president for SaaS solutions at Dimension Data. These anecdotes of public cloud avoidance, incidentally, are taken directly from actual discussions with clients.
Reason #1: “I don’t know how much it will cost me. My current costs are definitive.”
“It’s understandable that CFOs initially feel more comfortable approving a PO for a fixed amount for capital equipment, but that doesn’t tell the whole story. It’s one thing to be able to definitely forecast your costs for previously purchased equipment, it’s another thing entirely to be able to forecast your demand. If you’re able to do both, congratulations,” said Cumberland.
He pointed out the reality is that very few clients can predict demand, or can expect it to grow consistently without peaks and valleys. To optimize the economics of the public cloud, users must be consider both the hidden costs of maintaining unused capacity, and the opportunity costs of not being able to meet unexpected demand exactly when needed.
Reason #2: “Shared public cloud architectures don’t offer the performance my applications require.”
Virtual machines running on a shared infrastructure do offer less performance than their physical counterparts with similar specifications, admitted Cumberland. However, he pointed out using that as a reason not to migrate applications to cloud was accepting the strawman argument.
“Yes, virtual performance is inferior to physical, but virtual machines also cost considerably less, come with no long-term commitments, and allow for automation of operations in ways that are significantly more difficult when relying on physical machines,” said the Dimension Data executive.
The advance of hybrid cloud compute options and virtualized, high-performance storage solutions also allow a broad range of applications to effectively run in the public cloud regardless of the performance requirements, he added.
Reason #3: “We’re 80% virtualized, so using public cloud offers limited benefits.”
Virtualization and public cloud, while related, offer different benefits, pointed out Cumberland. “While virtualization is an important step on the path to public cloud adoption, hosting everything internally still forces you to deal with the inefficiencies of capacity management, equipment procurement, installation, and maintenance of the physical systems, including training of your staff to remain current on the latest technologies.”
“If you’re already largely virtualized, the good news is that you are one step closer to conducting a critical evaluation of your current costs and infrastructure and beginning an evaluation of the public cloud. During that evaluation, it is important to critically evaluate the extent to which the time and budget spent hosting various applications and systems internally deliver a competitive advantage to your business, or whether they’re hosted internally only because they always have been, and no one before you has done the hard work to bring increased agility and lower costs to your business,” said Cumberland.
Cumberland believes that if companies are still citing these reasons and others like them, they are in danger of being left behind by their competition. Which leaves the contrarians among them without the ability to ever say “I told you so.” And for that particular group, of course, that’s the worst possible outcome.