Preparing the Data Center for a Move to Virtualization

Updated: April 14, 2010

Moving to a virtualized computing environment makes overseeing physical hardware easier because there are fewer devices, but the tradeoff is that it also makes running other components, such as software and storage, more complex. Operating systems and systems management tools were designed largely for monolithic servers, so their capabilities do not translate well to virtual environments. Vendors, such as VMware and Microsoft, have been trying to address these issues, but there are instances where virtual system features are not as robust as those found with traditional servers.

One challenge is figuring out how well each virtual server is functioning. Performance monitoring is difficult because a variety of applications run in different areas of a server but share elements, such as the device's internal and external storage. With so many moving elements, companies want to see which application(s) may be getting bogged down and where the bottlenecks are arising. Virtual system vendors do offer performance tools with their products, but in general, they provide users with limited information. In some cases, they may not work with real time data, and in other instances, they deliver broad rather than granular metrics.

Consequently, third-party suppliers, such as Uptime Software, Veam, and Vizioncore, have tried to fill the void. Since buying these tools adds to the overall cost of moving to a virtualized environment, companies should determine if their organization will need such products during the evaluation process rather than after a move to virtualization has been given the "Thumbs Up" by management.

Backup is another area where added complexity is evident. Consolidating 10, 15, maybe even 20 servers onto one platform appeals to companies who feel that their data center is now being overrun with hardware. While vendors have made progress with technologies, such as data deduplication, backup has remained a troublesome virtualized application. Running a backup application can take several hours on traditional servers. If a company tries to consolidate such applications onto a single device, there simply may not be enough time for them to get the job done. Companies need to determine their backup needs in their initial evaluations, so it does not become a "gottcha" as they begin their deployments.

In addition to technical issues, virtualization creates management challenges. This computing option blurs responsibilities among formerly distinct IT groups, such as server administrators, storage professionals, network engineers, and security teams. With all of these elements running on one system, there are no longer clear boundaries among these different groups. Server administrators need to understand how virtual LANs operate, and the network has been extended inside the host systems.

In order to deploy and manage their systems, companies will need to cross pollinate their staff's skill sets. So moving to virtualization requires investments (sometimes significant) in training. As server virtualization invades the data center, teams within the IT organization have to be prepared to work more closely together than they may have in the past. As a result, companies may also have to overcome the turf issues and infighting, as different managers and groups vie for control.