Server virtualization enables higher resource utilization, cost savings and improved efficiency. However, virtualization also creates new backup challenges.
Traditional backup processes aren’t well suited to the virtualized environment, with consolidated workloads and extreme data redundancy. Because legacy backup applications are not virtualization-aware, administrators must install backup agents on individual virtual machines (VMs). This quickly becomes a management nightmare as VMs proliferate.
Resource contention is another significant issue. Underutilized physical servers generally have plenty of processing power left over for backup processes. When multiple VMs share the same physical hardware, fewer resources are available for backup.
At the same time, there is more data to be backed up and backups need to be completed more frequently to meet recovery point objectives. The problem becomes exacerbated as the environment scales to hundreds or thousands of VMs sharing a common resource pool. Agent-based backup solutions place a significant burden on host servers and make it difficult if not impossible to complete backups within the available window.
Because server virtualization has become mainstream, all major backup vendors offer virtualization-aware products that can perform host-level backups without the need to install a separate backup agent on each VM. However, host-level backup presents a separate set of issues. These solutions typically use hypervisor snapshots to create image-level backups, then track block-level modifications. Challenges arise when VMs are moved from one server to another – a host may be running a different set of VMs than it was when the last backup was made. And while an image-level backup makes it easier to recover an entire VM, it requires at least a two-step process for file-level recovery.
In this scenario, backup is easy but recovery can be difficult. File-level recovery allows administrators to restore individual files in minutes without the time-consuming process of extracting the full VM image to a local drive. Organizations need this capability to support mission-critical applications running on VMs.
Best-of-breed backup solutions for virtualized environments offer both image- and file-level restore as well as replication rollback capabilities, protecting VMs from both hardware failure and software corruption. IT can recover individual items from any virtualized application, on any operating system, without additional backups, agents or software tools.
Hypervisor snapshots can be used to roll back a VM to a prior state, but do not provide adequate data protection for transaction-intensive applications. Storage snapshot capabilities eliminate the risks associated with hypervisor snapshots while increasing backup performance. The hypervisor snapshot is encapsulated within a storage snapshot that is then used as the source of the VM backup.
De-duplication further improves backup efficiency. When backing up multiple VMs, de-duplication software stores only one instance of similar blocks, saving time, network bandwidth and storage space. More efficient use of storage resources enables organizations to store more VMs for a longer period of time in order to meet recovery point objectives.
With VM replication, copies of mission-critical VMs are mirrored to a spare server and kept in the ready-to-start native format. This capability enables cost-effective disaster recovery for mission-critical applications and data. Should a failure occur, IT can run a VM directly from a compressed and de-duplicated backup file on regular backup storage, either in production or an isolated virtual lab.
Although it may seem overly complex to add a virtualization-specific product, as virtualization rates increase it is important to use backup and recovery tools that are built to manage the virtual environment. Virtualization-specific backup helps organizations overcome the limitations of traditional backup and gain fast, flexible and reliable recovery of mission-critical applications.