The advent of agent-less backup makes it easy to believe that the end of agent-based backup is nigh. Nothing is further from the truth. While agent-less backup addresses many challenges around the protection and recovery of virtual machines (VMs), agent-less backup is no panacea as compelling reasons persist for organizations to continue to use offer agent-based backup as an alternative to agent-less backup. Consider:
- Agents can better monitor server CPU and tune backup performance accordingly. Agent-less backups still incur overhead on the underlying physical machine as they consume server CPU and memory while the backup occurs. The consumption of these resources may, in turn, impact other VMs running on the host. Using an agent-based approach, organizations may better monitor the CPU and/or memory consumption on the host by the backup process. If the backup process does impact applications on other VMs on that host, an agent-based backup is better positioned to throttle the backup process to lessen its effect on other VMs running on that host.
- Agents better facilitate understanding of database or email applications as well as creating application-consistent backups. Almost every database and email application generates logs that must be captured and then applied to ensure their successful recovery. Capturing all of this metadata almost always necessitates the deployment of some type of agent on the VM to ensure this recovery. Even a number of backup solutions that promote themselves as “agent-less” still employ the use of an agent at the beginning of a backup of these types of applications to capture the data necessary to ensure a successful recovery. The main difference is that at the end of the backup they remove their agent while agent-based solutions have an agent permanently residing on the VM.
- Agent deployment and management is less of an issue. Agent-less backup certainly does eliminate the time, effort and management overhead associated with deploying agents initially and then managing them long term. However agent-based backup solutions have taken numerous steps over the years to automate the initial deployment of agent and then maintain them long term. Through their integration with management consoles such as Microsoft Management Console (MMC) and/or VMware vSphere vCenter, organizations may now deploy agents on VMs managed by this software as part of each VM’s setup and/or ongoing management.
- Agents gather more detailed, technical information about the volumes and applications within the VM. A key drawback for agent-less backup is their lack of visibility into the data contained within each VM. As such, when backups are completed, they often lack the necessary metadata to quickly restore specific components within the VM, such as specific files or folders. Rather the entire VM must first be restored and mounted before the administrator can navigate to and recover the data in question. Agent-based approaches capture and retain this type of metadata so administrators may more quickly do recoveries that require this type of detail to be at their fingertips.