Dominic Wellington (Twitter
@dwellington) had a great post at the BMC communities recently articulating the age-old discussion of cloud platforms solving a business problem, versus an IT geek spinnin’ VM Templates attempting to deliver a cloud solution.
I welcome you to read his great summation about how you should reverse-think your way about solving your cloud project woes, by focusing on the quarter-inch hole problem.
One point I would also recommend readers to realize is consistency and repeatability of build processes. That is usually one major flaw in the traditional silo IT delivery (e.g. compute team building a VM from a template), as traditionally in this model various automation technologies (e.g. VMWare template cloning, or NetApp RapidCloning) are shoestrung by manual process (attaining IP addresses, installing management tools, patching, installing software, and transitioning the system to the end user, etc).
Usually, that is where the proverbial “weakest link” comes into play. I have lost count of the times where a request for 10 exactly-alike windows servers end up having 10 different servers built, even with unintentional software installed which increases security risk and vulnerability/attack surface (e.g. why is IIS installed on my server?).
Don’t get me wrong – I’m a true believer of the “people, process, technology” triangle. I just don’t believe people are best suited doing repeatable machine work. But as good as a sys admin you may be, you are not perfect. Nor are interactions between human beings. We all have character/personality differences which add to the complexity when working in team and ensuring a smooth, consistent, IT build process. Heck, sometimes we just may be having a bad day! Now multiple that by 10 fold. Now 100. Now 1000 provisioning requests. What are the odds for process failure?
When I try to explain to users the value add of BMC Cloud Lifecycle Management, I try not to focus on the O/S build process, or how it is better than the latest cloning technology out there. I focus on the system stack build challenge, for which the OS is a key component but not the only thing that matters. An operating system is a commodity folks! Repeatability is key, and consistency is the only way to go. Unfortunately, the human process – not matter how good – are rarely consistent. Let’s not get started on End User Experience either……
The other factor to take into account is cost. There are huge pressures these days in the Federal sector to shift from Cost/Plus or T&M contracts to FFP (Fixed Firm Price) contracts. This significantly changes the model for software factories or traditional IT outsourcing companies. Before, where you were able to bill for each resource associated with a specific Task, now you have a ceiling on your revenue, while your operating costs are a constant challenge to keep under control while IT is semi-automated and labor intensive.
In this model, there is no better use case than to focus on shifting labor from repeatable process and activities to value-add, revenue generating activities. This is where a Platform, Infrastructure, or Software automation technology such as CLM comes into play.