Over the past decade, organizations have revolutionized their data center architectures and replaced legacy one app/one server environments with highly virtualized systems capable of making efficient use of expensive data center resources. At the same time, hypervisor vendors kept adding more features, which has resulted in a data center that looks nothing like it did just ten years ago.
In the figure below, you see the old way of doing IT: One app, one server. In this scenario, every application ran on its own hardware.
Virtualization has been a boon
The virtualization boon is no surprise. After all, it’s one technology that has mostly lived up to the hype and saved companies both time and money all while enabling new ways to do business. In short, it’s been a huge success story.
In the figure below, you’ll see that an additional software layer has been added to the application stack. Here, the hypervisor and its associated management tools enable workload migration between any of the physical servers. This abstraction has made it possible for organization to plan downtime more easily, improve overall availability, and make sure that they’re buying the right amount of hardware to meet ongoing business needs.
This is the point at which most organizations are today.
The cloud offers new opportunities
As you look to the future, new opportunities abound. First, VMware, while it retains a major leading position, is under fire from Microsoft’s Hyper-V, which has seriously closed the feature/performance gap with vSphere in Hyper-V 2012. That alone opens new opportunities for frugality in an organization. Now, it’s theoretically possible to create tiers of service based on hypervisor capabilities. Use the expensive vSphere tier for mission-critical enterprise workloads and the relatively free Hyper-V tier for lower-priority workloads.
Further, given the rise of both public and private clouds, organizations have new options for workload placement. Now, based on any number of factors, workloads can run on vSphere, Hyper-V, Xen, in a private cloud environment or in a public cloud environment, such as Amazon or Azure.
In order to operate in such an environment, organizations need to adopt tools that enable seamless management of the entire environment, as per the figure below.
In such a scenario, workload abstraction is taken to a new level and the data center becomes just another target object when it comes to deciding where a particular workload should run. Workloads can be seamlessly migrate from Hyper-V right to Amazon where they simply continue to operate.
We’re already seeing what this new world would look like in some software tools that are on the market today. System Center 2012, for example, includes Virtual Machine Manager 2012 SP1, which enables migration of workloads to and from Windows Azure. VMware’s vCloud Director product includes APIs necessary to integrate it with various cloud-based providers. In addition, other big companies, such as BMC, provide comprehensive management platforms that can be extended to support cloud environments.
These products are true behemoths in this multi-hypervisor/on-premises/cloud management space. As such, they can be difficult to get up and running. Other companies, such as HotLink, aim to bring simplicity to this space through less expensive, but still capable, tools. I fully expect this space to become much hotter in the coming year.
Action Item: CIOs should consider all of their options when it comes to where workloads operate. While a single hypervisor solution was all but required in years past, newer management tools are providing additional workload abstraction opportunities that may enable additional cost savings, even higher levels of availability, and seamless workload shifting on demand in order to enable a set of technology services that flexes and shrinks with the ebb and flow of the business. By looking at the data center as but one object in a series of options, CIOs can expand their viewpoint for the good of the organization.