What you will learn in this tip: Server virtualization technology can provide some unique solutions when compared to traditional disaster
recovery (DR) methods. It can reduce the cost of implementing DR while providing increased flexibility, and ensure recovery time objectives (RTOs) are met. However, every technology has its pros and cons, and there are situations where using virtualization as part of a DR strategy makes sense and other situations where it may not be a good fit. In this tip, we’ll examine some of the advantages of server virtualization for DR, and when you shouldn't use server virtualization as part of your DR strategy.
When implementing an effective DR strategy, budgets and RTOs/recovery point objectives (RPOs) usually dictate the type of solution that you implement. Server virtualization can help you stay within your budget and meet your RTOs/RPOs. Even if you don’t use server virtualization in your data center today, you can still leverage it at a DR location as a recovery platform for your physical servers. You can easily convert physical servers into virtual machines (VMs) using conversion tools that transform the operating system on a physical server into a virtual machine image. For companies that have already implemented server virtualization at their primary site, there are more options available for copying VMs between sites. In addition, server virtualization can provide many simple and advanced options for transporting VMs from site to site.
The advantages of server virtualization: Cost tops list
Perhaps the biggest advantage of using server virtualization at a DR site is cost savings. With server virtualization, you no longer have to purchase one server at your DR site for every server at your primary site that you want to protect. Consequently, you need far fewer physical servers at your DR site, which can save you money when purchasing server hardware. Virtualization also allows you to consolidate and reduce your network and storage requirements. Less network ports are needed as VMs share network adapters in a host. Less storage space is needed as VMs can be thin provisioned, and since VMs share storage volumes, there is less wasted storage space.
Cost savings with virtualization extend beyond your initial hardware purchase. You also gain ongoing operational cost savings as well. Having fewer physical servers means less rack space is required at your DR site, which can result in reduced operational costs in shared data center environments. Additionally, if you are paying for power and cooling at your DR site, fewer physical servers also reduces power and cooling requirements. Having fewer physical servers also means you have to spend less time and money maintaining physical servers at a DR location as well.
One of the biggest challenges with DR is performing failover testing. You have to validate the servers you put at your DR site are going to work properly when needed. Server virtualization can provide easier failover testing as VM operations can easily be scripted and automatically orchestrated. Failover can also be achieved using automation products like VMware Site Recovery Manager (SRM), which can build complex recovery plans that can be implemented at the push of a button.
Another challenge is with failback. Server virtualization can make this easier as well when using a VM replication product that can easily change replication directions without involving storage products. When testing a DR site, you often need to isolate the environment so it doesn't impact your production environment. Virtual networking provides greater flexibility for testing at a DR site without impacting production servers as virtual networks can be isolated from production networks.
Replication products are often used to keep a DR site in sync with the main site. Storage replication can be very costly and complicated to implement. Server virtualization offers a simpler and lower cost solution by replicating VMs at the virtualization layer. This provides you with the freedom of using dissimilar storage devices at the primary site and DR site. Even local storage can be used, which can reduce costs because you don't have to purchase an expensive shared storage device.
Hardware compatibility is another advantage of using server virtualization as part of your DR strategy.
Finally, hardware compatibility is another advantage of using server virtualization as part of your DR strategy. Often times data centers have many different models of servers that span hardware generations. Trying to restore a server onto dissimilar hardware can cause many problems with drivers, disk partitions and applications. In addition, if you plan on using any type of replication, you need similar hardware on both sides to prevent issues. Server virtualization can eliminate hardware compatibility problems—VMs run on virtual hardware that is emulated by the hypervisor and there are no hardware compatibility issues caused by differences in the underlying physical hardware. A VM will see the same virtual hardware regardless of the physical hardware that the hypervisor is running on. If virtualization is used at both the primary site and DR site, you can use different hardware without having to deal with compatibility issues. This provides greater flexibility to use any type of hardware and also makes upgrading physical hardware at the DR site easier.
Why you shouldn't use virtualization as part of your DR strategy
As compelling as server virtualization technology is there are some reasons why you may not want to use it as part of your DR strategy. One of the biggest concerns with virtualization is having all your eggs in one basket. When host failures occur in a virtual environment, they can have really big impacts. Even in the smallest servers today can scale pretty big and, as a result, very high consolidation ratios can be achieved. To achieve maximum ROI with virtualization, you need to get the most VM density out of your hosts as possible, which increases the impact that a host failure will have on your environment. To lessen the risk, you can scale out instead of up by using a larger number of less powerful servers, which spreads the VMs across more hosts and a single host failure will have less of an impact. You can also leverage high availability features such as VMware’s High Availability (HA) or Fault Tolerance (FT) feature to help quickly recover VMs on failed hosts or prevent host failures from impacting VMs.
If you haven’t used server virtualization yet, dealing with the complexities of virtualization can be challenging. Server virtualization impacts just about everything in the data center and the changes brought about by virtualization can be difficult. Proper training can help provide the skills needed to effectively manage virtual environments and to help avoid the most common pitfalls. There are also many consulting partners that can assist with all phases of a virtualization project from the early planning phase to implementation, and can help provide you with the experience that you may be lacking.
Server virtualization is touted as a cost-saving technology, but these savings aren't immediate and occur over time. You’ll need to spend a lot of money up front to implement virtualization. This includes hardware costs, host licensing and any other products you'll need to manage your virtual environment. Hardware specifications for virtual hosts are typically much greater than traditional single-use physical servers because they must support a large number of VMs so you can achieve the maximum benefits from virtualization. To help lower costs, you can get creative in your virtual environment design and leverage lower cost shared storage, and utilize some of the features that can help you reduce the amount memory and disk required for your VMs. Virtualization vendors may tout their free hypervisors, but if you want to scale your environment and use advanced features, it’s going to cost you. To help reduce the costs, look for licensing bundles or packages that offer discounts compared to buying individual licenses. While the virtualization vendors all sell their own management products, there are many more affordable options available from third-party vendors that often cheaper and better.
While there may be a few disadvantages, the advantages of server virtualization outweigh the drawbacks, which can be avoided with careful planning. If you have already implemented virtualization at your primary site, then you are most likely past the learning curve and already have the skills and experience to successfully implement it at a DR site, too. If you spend the time and effort to properly plan your virtualization strategy at your DR site, you can avoid the drawbacks and enjoy all the advantages that server virtualization provides.
About this author: Eric Siebert is an IT industry veteran focusing on server administration and virtualization. He is a very active member in the VMware Vmtn support forums and has obtained the elite Guru status by helping others with their own problems and challenges. He is also a Vmtn user moderator and maintains his own VMware VI3 information website, vSphere-land. In addition, he is a regular blogger and feature article contributor on TechTarget's SearchServerVirtualization and SearchVMware websites.
This was first published in July 2011