Companies have spent billions if not tens of billions of dollars putting in place the necessary hardware and software to virtualize their data centers. These numbers fail to take into account the countless man hours that they have also spent planning, configuring and implementing these environments. Yet just as many are ready to take a deep breath and enjoy the fruits of their labors, they are coming to the realization that virtualizing their environment and creating private or public clouds only laid the foundation for the real goal of data center automation.
It is rare in my conversations with end users, resellers or vendors that the topic of virtualization or the cloud does not come up in some context. While virtualization may not be overtly discussed, it is minimally presumed that the IT organization is looking to initiate or more fully take advantage of virtualization, the reseller wants to position itsr stable of products so it can help “enable” virtualization or the cloud and the vendor always talks about its deep levels of support that its products have for one or more of the various hypervisors (Citrix, Linux, Microsoft Hyper-V and VMware.)
The motivating factors for all of these parties implementing, selling and/or supporting virtualization and the cloud are two-fold. Initially it was cost savings as it enabled organizations to buy less hardware and software and then better utilize what they acquired. While that rationale has never gone away and remains valid, the flexibility and agility that virtualization offers have emerged as equally if not more compelling reasons to virtualize data centers than its cost savings.
Yet what people on all sides are beginning to realize and more fully grasp is that despite virtualization’s benefits, it can be a real pain to manage virtualization well. This is why unless data center virtualization turns into data center automation, the initial cost savings and flexibility that virtualization provides may evaporate.
In looking at the value of data center automation, it is important to differentiate and distinguish it from data center visualization. Visualization is equally important and a good first step toward automation as it provides organizations important insight into such items as:
- What physical devices are in the environment
- What VMs reside where
- What physical resources the VMs are using
- How the physical and virtual environments are laid out
Data center automation differs from this. It takes into account all of the resources in the virtualized environment and then looks to place the VMs on the hardware where they will be most optimally used.
For instance, the vMotion feature in vSphere gives organizations the flexibility to move a VM from one server to another. That’s great. But what vMotion does not necessarily do is evaluate all of the available physical machines in the environment and move it to the least utilized physical machine in the environment.
Even then, the next logical question is, “Should the VM even be moved to the least utilized physical machine?” There might be a number of reasons the physical machine is the least utilized. It may be an older physical machine and was in the process of being phased out of production. The application on the VM may be processor intensive and the new physical machine is not able to handle the VM’s workload. The VM may require the physical machine to be configured as highly available.
In short, just because virtualization gives organizations the flexibility to perform a specific task that they could not do before does not mean they should now do it. Rather what is needed is data center automation software that capitalizes on the possibilities that virtualization makes available and automates these tasks.
As it performs these tasks, it needs to do execute upon them in an intelligent manner using real-time information from the environment (resource availability, application workloads) and business-driven policies (this physical machine is being phased out so do not put VM workloads on it.) It is only when virtualization and the cloud begins to offer these levels of automation that organizations will be able to more fully treat and manage their data center as the business asset and enabler that it is and was always intended to be.