Relinquishing Traditional Data Center Controls

I recently had a meaningful conversation with a customer as we spoke of a three-year plan to move “everything” into the cloud. I hear this statement over and over again, and it always brings visions to my mind of jumping into the deep end of the pool without knowing how to swim. For most, the perception is that the cloud provides a robust data center for multiple use cases—anywhere from dev, test, disaster recovery, and, of course, production workloads. Organizations look to the cloud as a way to solve their challenges in operating on-premises primary, secondary, or multi-tier data centers. I certainly agree with that mindset. Cloud would not be a multibillion-dollar business today if it wasn’t ready for primetime. 

However, businesses today often think of the cloud as an extension of their organization. It comes without the cost or effort of owning data center infrastructure—or all the technologies that need to be managed to support it. These technologies have no direct influence on how a business’s products and services are delivered to the end customer. Customers do not care if the business has a diesel fuel contract to power an enterprise standby generator 24x7 in the event of a catastrophe; they just want to transact their business and have their services fulfilled. IT organizations want that same level of service from their systems, as well, and they don’t want to spend their time and energy focusing on these systems when they do not serve the immediate needs of the end customers. All of these things make cloud an attractive choice.

Going back to my conversation with the customer, we went through the operational processes (on-premises) and compared them to the cloud. The thing that stood out during our conversation is that cloud providers do not give IT organizations access or control to any physical resources. IT organizations have to operate at a level where they no longer have access to the physical switches, host servers, hypervisors, storage arrays, backup device, and so on. All of these things are abstracted in a shared multi-tenant, elastic environment. 

While my customer understood this, he started to dig deeper into the repercussions of this change. For example, he would no longer have the ability to do virtual machine image level backup on his cloud virtual servers, and his on-premises backup system that’s fully integrated with the on-premises hypervisors and storage could not function in the cloud. He would have to deploy cloud-born backup systems or services that do not necessarily have the same feature sets—and all at an additional cost. This may affect any business-driven or compliance-driven policies for data protection. 

In short, many of his on-premises operations and controls may no longer exist in the cloud.

This is not really a bad thing. It is a paradigm shift that makes cloud a very viable solution for numerous businesses that do not want to be in the IT infrastructure business. It requires a cultural change from a hardware-defined to a software-defined environment. Even running a hybrid on-premises/cloud data center has several considerations once you dynamically move or scale workloads out to the cloud. Think of it as going from your house to a hotel. The facilities are different, and you can’t personally go to the maintenance closet to fix an issue that arises.

Going to the cloud has advantages and intricacies that need to be discussed. Let’s have that conversation.