Many businesses today are using virtual and physical servers at the same time, and the number of virtual servers is beginning to outpace physical server utilization. This hybrid reality drives the need to protect both virtual and physical environments. In so doing, IT managers can look to the following requirements as a basic road map for achieving a consolidated backup:

  • Requirement 1: Comprehensive protection for both virtual and physical servers.
  • Requirement 2: Reduction of the recovery time objective (RTO) from hours or days to minutes.
  • Requirement 3: Elimination of backup windows, which helps reduce the recovery point objective (RPO) to just minutes.
  • Requirement 4: 100% reliability and assurance of recovery; the right solution automates verifications, helping eliminate manual verifications while ensuring data is correctly captured.
  • Requirement 5: Universal recovery across virtual or physical environments — anywhere to anywhere — including restoring to dissimilar hardware; granular recovery ranging from files and folders to email messages, database rows and application objects; and ultimately full bare-metal server recovery.

A block-based approach

Because legacy backup products were not designed for virtual environments, they often cannot fully address these requirements. Virtualization adds complexity and additional perspectives from where backups can be sourced (see Figure 1). An agent might sit in the hypervisor’s management partition, in the external storage or even in the virtual machine itself. Each of these positions comes with its own strengths as well as weaknesses.
Figure 1. Various locations where a backup agent can sit in a storage environment

Figure 1. Various locations where a backup agent can sit in a storage environment

There is, however, one location across both physical and virtual infrastructures that presents a single, unified place for an intelligently designed agent to capture data: the disk volume layer. Instead of dealing with individual files and folders in a file system, the disk volume layer sees data as blocks on disk. When data changes, so does the disk block. Capturing those changes and replicating them elsewhere means backing up those changes as they occur. That backup happens in near-real time and in a fashion that can be replayed elsewhere when necessary.

Such a solution is elegant in its simplicity. Because data is treated as blocks at the disk volume layer, attaching a backup agent to this process means watching each block as it changes. Once that block changes, the change is replicated to the backup server (see Figure 2). This process avoids the entire notion of a backup window, because data is constantly being backed up — solving Requirement 3.

Figure 2. Replicating changes in block-by-block fashion

Figure 2. Replicating changes in block-by-block fashion

All kinds of useful things can be done with the data once it is at the backup server. Views of that data can be created almost instantaneously simply by replaying the data into a workable user interface. This reduction in RTO addresses Requirement 2.

Backups can be verified by running a series of automated tests. Those tests can validate file system integrity, as well as database functionality across a range of supported applications. This verification helps organizations achieve the 100% reliability and assurance of recovery of Requirement 4.

Data within virtual machines is also easy to restore in a block-based backup scenario. Such data can be gathered simply using similar views, without requiring an arduous double-restore process. In fact, with the right intelligence built into such a system, restoring supported data structures can be accomplished through the same process. It avoids the variety of competing steps that are necessary in other backup architectures, and addresses Requirement 5 for universal recovery.

Above all, a disk volume focus enables virtual and physical computers to be treated equally from both the backup and the recovery perspective. No longer must backups for virtual machines involve different steps and technologies than those for physical ones — resolving Requirement 1. From the perspective of the backup, they’re all just blocks on disk.

Greg Shields is a senior partner with Concentrated Technology and a well-known author on storage technology.

Download this white paper and others from Dell AppAssure: