Server Virtualization Piles on the Savings But Watch Out for Those Costs
Everyone is quick to tout the cost-savings that server virtualization provides - less server hardware, lower heating and cooling costs, smaller data center footprints, better utilization of existing hardware. In many respects, it's like a dream come true from a cost savings perspective for many organizations. But successfully implementing server virtualization is another story as it requires organizations do more than just remove the shrink wrap on the software, install it and then watch the savings pile up.
Server virtualization addresses many of the current operational costs that data centers face. But VMware's ESX server and Microsoft's Hyper-V server virtualization OSes are anything but plug-n-play, especially when it comes to protecting data in these environments.
Quantum's Product Marketing Manager, Mike Sparkes, cautions that organizations have to avoid the temptation to just look at the upfront cost savings in reduced hardware and software costs. The temptation can be to just calculate the cost saving by looking at the benefits that an organization will derive by virtualizing "x" number of existing servers now using a "y" number of servers.
What that "y" number can fail to account for are the performance bottlenecks and licensing costs associated with protecting the data created by the virtualized servers. "Once you create a virtualized environment, there are better and easier ways to manage backup performance and software licensing than what organizations experience with their current physical servers," says Sparkes.
Protecting the data in virtualized environments requires companies adopt some new approaches that both reduce the amount of data they backup as well as offload the overhead associated with backup. A couple of data protection options that organizations should look to adopt and implement server virtualization include:
The problem that organizations run into is that both of these options introduce new costs and complexity into the equation. To deduplicate data means the introduction of disk into the backup process. Further, organizations may not be using backup software that can make these specific calls to VMware's VC B or Hyper-V's VSS Writer that generate the snapshots and then do the off-host backups of the data. Because of this, organizations suddenly find themselves looking at new expenses associated with protecting this environment - the cost for disk and the cost for backup software and proxy servers that supports these new snapshot features plus the time needed to implement the solution.
No one argues that server virtualization can significantly reduce data center costs but as organizations factor in the costs and complexity associated with protecting this data, the appeal of server virtualization can lose some of its luster. It is because of this that we are starting to see more product offerings that bundle deduplicating disk storage systems with data protection software to minimize the expense and complexity associated with implementing data protection in these virtualized environments. In an upcoming blog entry, DCIG will take a look at a new entrant in this space.
Server virtualization addresses many of the current operational costs that data centers face. But VMware's ESX server and Microsoft's Hyper-V server virtualization OSes are anything but plug-n-play, especially when it comes to protecting data in these environments.
Quantum's Product Marketing Manager, Mike Sparkes, cautions that organizations have to avoid the temptation to just look at the upfront cost savings in reduced hardware and software costs. The temptation can be to just calculate the cost saving by looking at the benefits that an organization will derive by virtualizing "x" number of existing servers now using a "y" number of servers.
What that "y" number can fail to account for are the performance bottlenecks and licensing costs associated with protecting the data created by the virtualized servers. "Once you create a virtualized environment, there are better and easier ways to manage backup performance and software licensing than what organizations experience with their current physical servers," says Sparkes.
Protecting the data in virtualized environments requires companies adopt some new approaches that both reduce the amount of data they backup as well as offload the overhead associated with backup. A couple of data protection options that organizations should look to adopt and implement server virtualization include:
- Deduplicate redundant data. Virtualized environments can contain large amount of redundant data so when this data is backed up, the same data is stored multiple times. They also contain a lot of allocated but empty disk space. Deduplicating the backup data and storing it to disk, reduces the footprint of the backup data to a minimum which saves on storage capacity and energy costs while expediting backups and recoveries.
- Use native backup tools found in VMware's ESX Server and Microsoft's Hyper-V OSes. Both of these server virtualization operating systems provide base line data protection tools that minimize or eliminate the need for agents on each virtual machine (VM) that it hosts.
The VMware ESX includes the VMware Consolidated Backup (VCB) feature which creates a snapshot of a VM so that the backup of a VM can occur using a backup server without incurring a performance penalty on the hosting ESX server. This eliminates the overhead associated with the backup. Microsoft Hyper-V is similar in that it also creates snapshots of either individual VMs or the entire Hyper-V server.
Unlike VMware, however, Hyper-V uses the Volume Shadow Copy (VSS) Writer found in Microsoft OSes which gives it a slight edge over VCB in Windows environments since it can create application consistent snapshots of virtualized Windows OSes. The backup of the snapshot is then accomplished in a manner like in a VCB environment by using a proxy server. (For further information on Hyper-V backups, a blog recently appeared on Microsoft's Technet website that examines this subject.)
The problem that organizations run into is that both of these options introduce new costs and complexity into the equation. To deduplicate data means the introduction of disk into the backup process. Further, organizations may not be using backup software that can make these specific calls to VMware's VC B or Hyper-V's VSS Writer that generate the snapshots and then do the off-host backups of the data. Because of this, organizations suddenly find themselves looking at new expenses associated with protecting this environment - the cost for disk and the cost for backup software and proxy servers that supports these new snapshot features plus the time needed to implement the solution.
No one argues that server virtualization can significantly reduce data center costs but as organizations factor in the costs and complexity associated with protecting this data, the appeal of server virtualization can lose some of its luster. It is because of this that we are starting to see more product offerings that bundle deduplicating disk storage systems with data protection software to minimize the expense and complexity associated with implementing data protection in these virtualized environments. In an upcoming blog entry, DCIG will take a look at a new entrant in this space.
Leave a comment