BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
These days, many IT professionals are asked to do more with less money, and increasing the efficiency of your virtual host resources may be the only answer in lean budgetary times. Luckily, there are numerous ways to squeeze more VMs into an existing infrastructure.
With that, here are 10 ways extend and expand virtual host resources on the cheap.
1. Get lean
Rein in virtual machine (VM) sprawl and take inventory. Does Team A really need all of those test VMs? Are there orphaned VMs in the environment? Are there legacy VMs with data that you can import into another product?
Taking back VMs is often hard, especially if you haven’t implemented a chargeback model. Every VM deployed has an associated cost due to use of various virtual host resources. At least if you have a chargeback or showback model in place, you can report to upper management who is on board with the organization’s goal of saving money. Otherwise, it may take a strong public directive from upper management to succeed.
2. Shut down VMs
Many times, VMs in test and development environments are left idle for days, weeks or months in between testing cycles. Shutting down VMs between these times can save significant virtual host resources. It also allows for a natural ebb and flow for various testing cycles for different groups.
You may also be able to provision more VMs. Because of the sporadic need for test VMs, you wouldn’t need additional hosts or resource increases. Going a step further, creating self-service policies for application teams and restricting the number of VMs that a person or group can run at once may also help.
3. Forego hardware replacement
If your normal server refresh cycle is 3 to 5 years, there still may be plenty of life left in your virtual hosts for another year or two. The number of VMs that you can house on these hosts may diminish throughout the years because of the natural increase in VM resources. But you may be able to hold off on purchasing new equipment for a period of time.
Even if there is enough budget for new virtual hosts for production workloads, during tight times, you could use these older hosts in test/dev environments for years. On aging hardware, these workloads may not function at optimal performance, but the equipment will still provide the functionality to keep these second-tier environments viable.
4. Extend support contracts and recycle hardware
Extending hardware past the normal refresh and manufacturer-support contracts is a lot like hoping your paid off car does not become problematic and costly. Sometimes capital budgets are limited, but operating budgets are not.
Money may be more readily available for operating expenditures, allowing you to extend your normal hardware-refresh cycle. This type of limited expenditure could help you weather any failures and lessen the risk by protecting your aging virtual infrastructure.
5. Enforce usage policies
Creating and enforcing virtual machine-usage policies can save host resources, which may help you avoid that next purchase. Many times, when trying to extend your current virtual host resources, this is the most unpopular strategy, because it usually requires changes to current practices and behavior throughout multiple teams. And, without high-level managerial support, this option will fail.
No one likes change, and there is usually resistance at first. If these policy changes are perceived to come from a lower level or peer team, without higher-level managerial support, expect overt resistance or out-and-out sabotage.
6. Adopt a new hypervisor vendor
Sometimes, forgoing purchases during tight times is not the only answer. Making smart purchases to improve the cost effectiveness of your virtual environment may be something to consider.
Hypervisors have cost and functionality sweet spots for different VM operating systems. Look at your existing hardware and hypervisor-licensing options, which may reveal the need to change virtualization platforms -- especially if you foresee long-term budget restrictions in+- your market segment.
But switching hypervisors may require changes to your technical staff, so getting teams and management onboard is very important for this option to succeed.
7. Add memory
Memory is often one of the first resources exhausted on a virtual host. For little expense, additional memory may allow you to provision more VMs, eliminating the need to purchase new equipment.
8. Use existing features
Features like Hyper-V Dynamic Memory and VMware Memory Overcommit may allow you to provision more VMs on a virtual host. These technologies distribute only the amount of memory that VMs need, instead of allowing VMs to consume a predetermined amount of virtual host memory.
If you are not using these features to some degree, you are leaving money on the table. Not all VM workloads are suitable for these features, but investigating each VM’s memory habits may increase the number of VMs per host through the better management of virtual host resources.
9. Switch Virtual Hard Disk type
In most environments, IT administrators deploy thick or fixed Virtual Hard Disks (VHDs) to ensure top performance. For production VM workloads, this type of VHD is still a best practice. But, for test/dev environments, consider using dynamic or thin-provisioned VHDs.
All modern hypervisor vendors have a method to transition fixed disks to a dynamic/thin-provisioned VHDs. As you transition these VHDs, you will free valuable physical hard drive space, which will allow you to provision more VMs on the same amount of physical storage. By doing so, you can avoid or delay the need to purchase additional storage resources.
10. Upgrade to the latest hypervisor version
New hypervisor features often increase the efficiency of a virtual environment. Hyper-V Dynamic Memory, which debuted in Windows Server 2008 R2 SP1, is a prime example of a version upgrade that increased VM-to-host ratios and didn’t require new hardware. Even if an upgrade requires some investment in new hardware, you should weigh if there is still a potential budgetary benefit to making the purchase.
Ultimately, data center growth does not shrink but budgets do. Expectations of growth are still the same, so it will take some creative reconfigurations in your virtual environment to stay afloat. There are still some gotchas in the aforementioned plans, as you are trying to squeeze the most out of your existing virtual host resources. So, be cognizant of pushing the limits of your shared resources too far.
In tough times, finding the balance between accommodating more VMs and overextending your virtual hosts, which may cause increased instability, will be the main goal. Walk the tightrope carefully.