Gartner analyst Tom Bittman recently posted a blog showing that the majority of corporate private clouds are failing,“I was a little surprised that 95% of the 140 respondents (who had private clouds in place) said something was wrong with their private cloud. [1]” At that high a ratio, it doesn’t matter what technology you use for the private cloud – Microsoft, OpenStack, VMware or something else. His report gives a variety of reasons for these failures but the net-net boils down to this: The average corporate private cloud environment wasn’t built to the specifications of its internal customers but, instead, was a reflection of what IT was comfortable doing and wanted the private cloud to be.
Most of these clouds aren’t cloudy at all but are more, what Forrester Research Inc. Senior Analyst Lauren Nelson calls, “Enhanced Virtualization” environments. This means they are traditional static virtual machine environments with a little automation to drive up IT’s efficiency. They aren’t self-service to the developer, don’t provide fast access to fully configured environments, aren’t metered (let alone pay-per-use) and typically don’t support Chef or Ansible-based deployments, let alone containers. For private cloud administrators to capture the new enterprise applications, you need to rethink your approach and make the radical and culturally difficult shift from infrastructure management to service delivery. You need to learn from the clouds, but more importantly, reflect what your developers want from your private cloud solution.
Too many IT infrastructure managers see public cloud as the enemy and thus don’t bother to either understand why their developers are using them, or get their hands dirty understanding just how these environments work. This mentality serves only to reinforce the long-standing views held by these administrators, and if you are trying to appeal to your front-line developers, this approach is just wrong. A cloudwashed virtual server environment that takes two days to deploy new workloads, fulfilling requests through the help desk and having no cost transparency will lose every day to a public cloud.
And don’t even try the security or reliability card. Really? You honestly think your static VM environment hiding behind outdated firewall-based security hosted out of a 1990s era data center on servers bought in 2013 is going to trump a public cloud with clusters of next generation data centers in 19 geographic regions being protected by a team of top security professionals? And by pitting the public cloud as the enemy, you forgo any opportunity to partner with your developers around helping manage and monitor the new and often very important applications they are deploying to public cloud environments, which is where IT pros should really be concentrating your energies.
What you should be doing instead is teach your infrastructure admin community how to evolve from static in-house vs. cloud, to a true hybrid cloud portfolio. Help your IT infrastructure administrators understand why the front-line developers value self-service so much and how it doesn’t breed chaos in the data center — but just the opposite. Help them understand that a career path towards service definition and cost transparency is better than fighting to try and keep the company from using more public cloud – that ship sailed long ago.
Don’t agree? Agree but need more guidance to get this right? Let’s have the discussion. Come join me at the OpenStack Silicon Valley forum on Wednesday, August 26th where I’ll be leading a talk on this topic and debating the merits of this approach with colleagues from Google and CoreOS.
IT Pros are critical to enterprise cloud success but you have to evolve to own this opportunity. We can help and want to engage with you to ensure your success.
[1] Gartner Blog, Problems Encountered by 95% of Private Clouds, Tom Bittman, February 5, 2015