What is VM sprawl?

Virtual machines are simple to create and implement, which can easily lead to an overgrowth of VMs that lose their utility with time and disuse. VM sprawl often wastes both labor and resources.

Virtual machines operate as digitized versions of a physical computer. In most respects, a VM can perform the same computer functions as a physical computer. This includes running operating systems (OSs) and various programs, maintaining data storage and connecting to computer networks. A VM accomplishes these tasks while relying on virtual components, as opposed to “actual” resources like physical servers.

VM services are either considered to be “managed” or “self-service,” depending on how actively involved the user wants to become in administering VM creation and usage. The mimicry of VMs is so expert that VMs can even be synchronized to operate with the same chronological precision of real-time applications.

Virtual machines were first introduced during the highly fertile computing period that took place between the late 1960s and the early 1970s. VMs developed as an outgrowth of virtualization experimentation conducted primarily by International Business Machines (IBM).

Virtualization technology enables virtual software to successfully mimic the functionality of physical hardware—like servers, storage and networks. The virtual representations can be run on one physical machine, which significantly multiplies the utility an organization can derive from a single-computer system.

The primary goal of IBM® during this period was to find ways to improve performance on its history-making line of mainframe computers. Part of that effort involved developing time-sharing solutions. IBM first showed virtualization could work by unveiling its CP-40 research system in 1967. Its features included user-friendly commands, file-system commands, mapping of records to uniformly sized blocks and files that could be created simply by writing to those files.

Further refinements continued over the next five years, culminating in its watershed 1972 introduction of what is now seen as the world’s first virtual machine. The VM/370 marked the beginning of the IBM System/370 mainframes and was the first to support the use of virtual memory. The age of virtual machines that could offer full computer functionality within a virtual environment had begun.

Because they don’t require the same level of capital expenditure, working with VMs often yields considerable financial savings. The use of VMs often negates the need to acquire expensive physical servers and other equipment, which keeps VM pricing down. Further, VMs can capably manage workloads—especially if they follow traditional, legacy and monolithic workload types.

To coordinate their activity, VMs work in close conjunction with hypervisors. While VMs are files that duplicate aspects of physical computing, hypervisors are the software that executes and manages their functionality. Hypervisors enable the simultaneous running of different OSs on different virtual machines. VMs also work effectively with data centers, assisting them by helping them transmit cloud-based services through virtualized servers that utilize physical hardware.

Virtual machine sprawl, like many unfortunate situations, goes awry after starting off with the best intentions. A company’s IT department is tasked with developing a virtual machine for a very specific usage that is typically timely and carries some kind of urgency. That VM is built and put into operation. All is well.

But time passes. The business situation that warranted the creation of the VM may still be in effect, but the urgency that used to surround that VM has definitely cooled. Regardless of the cause for their disuse, the moment that unused VMs start to collect dust, they begin to be relics (and as needed as a hardware tool that’s purchased for one household project and never used again). Eventually, such VMs rapidly become little more than clutter.

The following are additional reasons that VM sprawl occurs:

• Simple creation process: Developers can quickly and easily make VMs, with a streamlined process for VM provisioning that can often be achieved through automation. This can begin a pattern where the first solution to any business problem is to create a new VM.

• Unnecessary proliferation: Because it’s so easy to create VMs, there can be a tendency to produce considerably more VMs for temporary use than are needed in the long run. And as business conditions evolve, organizations may feel led to produce an increased number of virtual machines in reaction to those changes.

• VMs saved for future use: A VM in disuse nonetheless represents a certain amount of work that’s been done purposefully and that VM still retains a specific amount of utility if put back into use in later projects. That’s why many companies are slow to winnow out their stock of VM resources.

• Expected performance stops: There are times when VMs can suffer increasing amounts of lag. Sometimes a VM’s performance simply needs to be tweaked. However, troubleshooting such VMs can prove difficult due to the many variables affecting VM performance.

• VMs off the radar: Another key reason for VM sprawl is that organizations lose track of all their VM resources. Such companies simply forget about past VMs they’ve used and still retain. If VM sprawl is severe enough, organizations could wind up needlessly remaking the same VM.

Some may wonder what the big deal is about VM sprawl. Sure, some idle VMs may be going unused, but is that reason to sound the alarm? Turns out that, yes, there is a sufficient number of reasons why VM sprawl warrants concern:

• Inefficiency: If an organization cares about running an efficient and well-controlled computing environment, it should regularly be policing its own storage and weeding out VMs that have outlasted their utility.

• Security vulnerabilities: In a perfect world, you could create VMs and then leave them laying around without significant worries about security risks. But if recent history has taught us anything, it’s that sensitive data can be attacked anywhere and no service providers are immune from cyberattacks.

• Waste: The twin pillars of resource usage and resource allocation can be wrecked by VM sprawl. In a system corroded with VM sprawl, ever more resources will have to be allocated and consumed, with only decreasing returns to show for that investment.

There are simple management techniques that can help limit the number of VMs an organization maintains to a manageable level.

Smart organizations monitor and regulate their VM creation and management. An effective VM governance policy addresses issues like capacity planning (which regulates the maximum number of new VMs allowed), VM performance optimization and VM lifecycle management.

Once a VM has been identified for elimination, the decommissioning process begins. The outgoing VM is removed from service. Then, the data contained within that VM is secured. Finally, the VM’s hardware and software are eradicated.

Organizations to need to observe data protection measures to secure those VMs that deserve permanent storage. This includes monitoring tools that identify disk space utilization and security safeguards to protect proper access control.

Even the most carefully constructed computer operation is subject to emergency situations in the neo-modern age. That necessitates an effective disaster recovery plan for VMs, which can be enacted during a worst-case scenario.

Once governance policies are in place, organizations need to make sure that all relevant staff members are trained in the agreed-upon IT infrastructure and backup plans. The governance policy should be clearly articulated to all IT teams and admins.

Another way to combat VM sprawl is to confront it directly with virtualization tools and technologies designed to help manage VM use and reduce the effects of VM sprawl. Such management tools include the following:

• VMware (now a subsidiary of computer-maker Dell Technologies) creates products that help manage VM use and reduce the effects of VM sprawl. One of its signature virtualization tools, VMware vSphere, addresses all key management points regarding VM operation.

• Another major player in this space is Microsoft. Its Hyper-V on Windows product mixes the virtual and the actual by letting users build and operate VMs on a physical host.

• VirtualBox is a free product that works with Windows, Mac and Linux and claims to have 100,000 registered users.

• The Linux® platform (developed by RedHat®) also supports VM use. VMs enable virtual environments, which lets users sample Linux distributions without requiring that those distributions be installed on physical computers. VMs also permit users to operate Linux on a CPU that runs on a different OS.