With the advent of multi-core CPUs, the deployment and use of virtual servers has become commonplace. One of the easiest places introduce this technology is within your DR/Business Continuity plan.
Most DR plans include physical backup servers that are tediously kept in sync with the primaries. These backup servers are housed in off-site locations and represent a duplicate cost component that impacts every aspect of your technology plan and the overall budget. Due to the inherent problems in the standard DR plan the end result is usually underutilization of resources, because you instinctively over-provision, and server management issues resulting from the continual server sprawl.
The deployment of virtual backup systems provides immediate cost and management benefits and it’s most likely something you can tackle with your existing staff. While the concept is complex, the implementation is fairly straight-forward. Here are some general steps to follow.
1. Assessment of current DR plan – specifically the application servers that will need to be converted. Consult with your application vendors. Discuss your plans to virtualize and inquire about any licensing issues that might need to be addressed. Vitualization is recognized as a change in hardware by most OS and applications vendors and should not require any additional licensing.
2. Select a virtualization platform – In no particular order, Microsoft, XenSource(Citrix), VMWare are among the most popular choices.
- XenSource(Citrix)
- Xen (Open Source)
Common among all platforms is the use of a hypervisor and an (optional) management tool. The hypervisor is basically the software that allows you to install multiple servers on a single system. It handles the process threading to individual CPU cores and provides a virtual network architecture on the hypervisor system. For small DR deployments a hypervisor might be the only component necessary. Some vendors actually provide their hypervisor software free of charge. For purposes of hardware provisioning, every active virtual server must be paired to at least one cpu core therefore a single socket Xeon quad core server could support 4 active virtual servers.
3. The final step is actually converting physical servers into virtual servers.
This step simply involves installing utility software on the physical servers that will automatically create a duplicate virtual server on the new hypervisor platform. The final end result should be a single powerful server that is now taking the place of multiple physical servers.
While taking on a virtualization project sounds overwhelming at first, keep in mind that you already have the skill sets necessary to accomplish the goal. It’s the same servers you are currently supporting, just converted into a more manageable and scalable environment. I hope you find this overview of virtualization helpful and consider putting virtual server deployments on your list of goals for 2011.
MegaNet Communications maintains a secure, fault-tolerant, carrier-neutral, datacenter in Providence RI perfect for your DR collocation needs. If you have any questions regarding DR or collocation please feel free to contact one of our highly experienced sales engineers sales@meganet.net.
-J.S.
