My IT department have been virtualizing even more, consolidating servers, sharing storage resources among multiple machines and converting NICs (Network Interface Cards) into virtual switches (I still haven't figured out how they did that).
The move into a virtualized environment is very useful for reducing energy consumption, decreasing physical server and storage foot print and driving up processor and storage utilization but it also has some side effects when it comes to data protection.
The problem begins at the same place that drove us into virtualization to begin with, resource sharing, You may now have 10 virtualized servers running on the same physical host, if your backup process consumed only 5% CPU and IO on your physical server, imagine what would happen if all 10 virtual machines kick off the backup process at the same time...
There are multiple valid approaches for providing data protection to those virtual machines and I’ll try to address each and every one of them in upcoming blogs…
- File based VS block based backups
- Keep your existing backup methodology (Agent-based backup)
- Perform the backup through the host (VMware console/hyper-v host OS)
- Hardware based snapshots
- Utilize vendor specific APIs that provide "agentless" or off-host backup (VMware's VCB and vStorage)
Other enhancements that might not necessarily be backup related but have to be seriously considered when virtualizing include
- Deduplication (client side or target side)