It wasn’t too long ago that many people looked at virtualizing servers in the production environment with great trepidation. There were legitimate fears over a lack of stability, performance, and availability, not to mention the overall risk of putting so many eggs into one basket.
Well, we have come a long way since those days. While it remains true that the ratio of virtualized to physical systems depends a great deal on the size, industry and focus of an organization, most companies have virtualized at least 10% of their servers, and many organizations are running at virtualization levels of 70% or more of their servers. The reason for this is that there are quite a few key benefits to be derived from a virtualized environment.
Here are my five favorite benefits:
- Reduced facilities costs
- Increased flexibility of deployments
- Better server utilization
- Simplified upgrade paths
- Better availability and easier disaster recovery options
Let’s take a look at how these items can translate to operational flexibility for your company.
Reduced facilities costs
Hardware consolidation allows organizations to significantly shrink their facilities footprint. This will translate into reduced electricity costs, reduced cooling costs, and ultimately the need for a smaller facility or server cage or server room. For those organizations that are using on-premises data centers vs collocation facilities, having a smaller data center footprint makes it much easier to procure and maintain office space.
And, given that the costs involved are not trivial, this can be helpful to the bottom line, or help fund other/additional projects.
Increased flexibility of deployments
In a properly sized virtual environment, bringing up new server instances can take just a day or so, taking patches and any custom software configurations into account. Add automation into the mix, and new server deployments completed in mere hours or less. No more waiting for new hardware (approvals, procurement, etc); no more racking and stacking; no lengthy hardware burn-in.
The time saved here is not only helpful for the technology teams that have to manage this equipment, but for the business units that are waiting on the new applications or new functionality that they are looking to take advantage of.
Better server utilization
One of the things that has plagued information technologists for quite some time is how to purchase servers that are sized well enough to handle both today’s and tomorrow’s computing needs, while not being so expensive that they are running at 10% of actual capacity for most of their lifespan. This conundrum has been a key factor in IT teams having to add more and more services and applications to servers that were originally purchased for one dedicated application or function.
Thankfully, server virtualization makes it much easier to size individual servers more appropriately. Yes, it does require that an organization size the virtual hosts correctly, but this exercise does not have to occur with nearly the same frequency as the creation of server instances, sizing concerns are greatly reduced. Plus, additions to the virtualization environment can be obtained and deployed incrementally.
Simplified upgrade paths
Most organizations have some sort of upgrade timeframe or methodology for servers, which can resemble one or more of the following:
- Set Schedule: Every n years (where n = 3 to 5 typically)
- Performance: When an influential stakeholder (VIP, customer, partner) complains about the performance
- Vendor Pressure: Whenever the application it supports is upgraded
- Spare Time: Whenever someone remembers and has enough time
Most normal organizations (read: those not awash in cash) are going to prioritize server upgrades, which means that critical servers are likely to see reasonable upgrades, but many of the behind-the-scenes servers used for infrastructure services, or those hosting otherwise unpopular applications, are not likely to see hardware refreshes on a regular basis. Additionally, in the physical realm, hardware refreshes will often be coupled with OS changes, so as to maximize the gain from the pain.
In a highly virtualized environment, however, adding capacity and performance to the virtual environment need not be a rip-and-replace affair. Not only is it possible to add performance through incremental hardware upgrades, but the performance gains can be felt by all – or, at the very least, by many – of the applications at once.
Let’s provide an example that shows the benefits for even a small shop. ACME LLC has a highly virtualized server environment, consisting of the following:
- A dual-controller iSCSI SAN with trunked 1Gbps ports
- Three (3) host servers, each with 6-8 guest server instances
If ACME LLC needs to add some faster disk for a key application – say, flash – or needs to upgrade one of its host servers to support better I/O for some key application, then the benefits of these changes can be felt by more than one of the guest VMs. Adding a new host server to the virtualization farm (or upgrading one host to newer hardware) will automatically provide performance improvements for all the guest VMs which end up on that host. Upgrades of individual guest VMs can now be limited to the operating system and applications, and not hardware.
Improved Availability / Disaster Recovery Options
Organizations employing server virtualization can more quickly prepare for disaster by backing up or replicating the guest images to another site which already has virtualized hardware (servers and storage). The amount of physical space and hardware needed for setting up a disaster recovery site is greatly reduced through the use of server virtualization.
It’s even better if you are using the various high availability features offered by VMWare, Microsoft and 3rd parties for dynamically moving guest VMs across your virtual farm.
Virtualizing servers makes it easier to provide high-availability in a cost-effective manner, whether in a local data center or in a remote office/facility. Organizations with multiple offices – assuming that they are already hosting their server equipment in those offices – can build out the virtualization hosts each office to be able to support another office’s servers. Or they can build their infrastructure to share the load across all the remaining virtual hosts.
These are just a few of the configurations that can be established to reduce the impact of a site or server disaster. And they cost less than buying and maintaining an extra set of physical servers for all services (or even for core services).
Virtualization In Conclusion
Employing server virtualization for your organization is a very good way to increase business agility, better manage technology costs over the long term, and reduce the impact of system and site interruptions. Yes, there is an upfront investment cost, but one that will quickly pay for itself as outlined above. Yes, there needs to be solid planning, but this is needed for all successful business projects.
If you have not yet taken the server virtualization plunge, or if your current implementation is rather limited in scope, now is the time to take full advantage of this technology to simplify both your business operations and your IT operations. You’ll wonder why you waited so long.