In my continual quest for knowledge and current news about virtualization and cloud computing, Deirdre Mahon’s article “Cost or agility: What is cloud’s true purpose?” caught my attention. I’d like to address the same cost or agility theme, but focus in on private clouds instead of the public cloud. My assessment is based on a more specific scenario, in which a company or corporation has presented its business units with the option of utilizing a shared infrastructure, in addition to the option of the public clouds that are readily available in the current technology marketplace.
What do I mean by “shared infrastructure”? Based on my experience, I consider the term to refer to company-wide managed services. I would expect to see significant investment in virtual infrastructure as well as some physical resources available to the business units. That is to say, these are typical, run-of-the-mill managed services; they are simply made available within the company, rather than sold as a service externally. In reality, there is little other distinction between internal and external managed services. For the sake of this conversation, I will assert that most managed services are based on virtual infrastructures that are in the process of becoming more cloud-like, where the end goal is to be a true private cloud platform.
The business units of the company in my hypothetical scenario correspond to the different “organizations [that] are switching to the cloud at a faster than ever pace” of Mahon’s article. When choosing the best vendor for an organization, options shouldn’t be limited to public cloud offerings only: they should also include internal shared services offerings. Although agility and speed are important in the decision process, at the end of the day, cost usually ends up at the top of the decision tree. While there are many other factors to consider, I will focus here on cost comparisons between the public and private offerings.
The public and private offerings are usually based on some sort of charge-back model, whereby a business unit or organization is billed based on the amount of resources provided. Many private offerings will have a difficult time competing with the public price offerings available. This is why managed services units are actively working on reducing costs and becoming more direct competitors with the public cloud price models.
How do you reduce costs? The cost of the physical infrastructure—computer, storage, and network—isn’t really going to change. The price of the physical equipment is what it is. Manufacturer partnerships can lower the price, especially as quantities of hardware go up, but those costs are fixed for the term of the contract until it’s renegotiated. If reducing the cost of physical hardware isn’t the answer, then reducing the cost of labor is the only way to reduce costs, in my opinion.
Modern-day managed services organizations are a collection of different technology towers, each one responsible for different technological areas. This model presents a series of boundaries that contribute to red tape and delays in overall completion of tasks and projects. This setup does very little to aid agility and speed when altering the infrastructure. It must change to provide agility and speed to customers.
As I mentioned earlier, managed service offerings are based on virtual infrastructures that are in the process of becoming more cloud-like, where the end goal is to be a true private cloud platform. A primary part of this transformation is the automation used to create and provide the capacity for self-service operations. The goal is to provide enough automation to cover the complete life-cycle of the server from cradle to grave: from the initial build process, through day-to-day operational tasks, to the decommissioning process to finish out the life-cycle.
The more automation and self-service options made available to the customers, the fewer the support people needed to support the overall environment. Making these options available will reduce the resource costs billed to the customer. More automation equals fewer people, and that is the transition that most virtual infrastructures are striving for in order to achieve both lower costs and greater agility.