Case against Private Cloud Projects
There are many reasons why I don’t like private cloud projects.
Almost every enterprise today already have work loads that run in their own data centers. I am not talking about those. Some of those workloads may stay within your DC for a long time to come. My point is that you should not continue to grow your data centers. Its OK have existing workloads as legacy workloads as long as you account for paying this legacy tax. I am talking about those workloads that are fit to move to public cloud and new workloads.
So, here are more reasons on why private clouds projects do not make sense.
1) Just-In-Time inventory
Your goal is to not hoard inventory of hardware. Neither should you have empty capacity lying around. Get rid of excess capability. This will help you save on power, electricity, head count costs, real-estate space and more. Trying to leverage excess capacity leads to more problems than its worth. This is nothing unique to your data center, this is true in many other production systems.
Public Cloud provides you the just in time inventory of resources. This frees up your developers, ops folks from having to follow more workflows. Even if you use a PAAS to provision newer resources, it requires excess capacity lying around. Public Cloud avoids this Capacity Tax for your organization.
2) Environmental impact
Public Cloud vendors are investing to make their data centers green. Please see: AWS and GCP .
An enterprise on its own will not have enough scale to pursue a full green data center. Private Cloud projects make the world less sustainable into the future. Please see here ( These data centers have generally made much less progress than their hyper-scale cloud counterparts due to persistent issues and market barriers, such as lack of metrics and transparency, and also misalignment of incentives. ) and here.
It is our responsibility to reduce energy footprint of our industry. Public cloud helps us make a dent into this problem.
3) Expensive
Private Cloud projects are expensive. You will overspend to meet non functional requirements ( security, availability, resilience and scalability). You will also have to hire more people to build and operate the infrastructure. This money is better spent on activities that directly help business outcomes, like profit. Please see: AWS Economics
Private Cloud projects are like wet cement. Flexible during vendor presence and then becomes rigid unsightly bricks. They become harder to change to meet evolving business needs. You will end up tearing apart your private cloud and start building another private cloud.
That’s all folks. Thoughts?
