With backup vendors revising their offerings to take advantage of VMware vSphere 4's new features, an upgrade to...
vSphere in 2010 could protect your organization's data storage.
Because a virtual machine (VM) is an encapsulation of an operating system, applications and data, protecting a production virtual disk is vital. IT organizations face several challenges when it comes to protecting their virtual machine disk image (VMDK) files. Finite shared physical resources on the host system and in-guest I/O-intensive backup processes have the potential to bring VM performance to a crawl. The increase in the amount of data in VMware environments can impact the backup window.
VMware Inc.'s release of vSphere 4 in the first half of 2009 introduced several data protection enhancements over VMware Infrastructure 3-era tools. And now that backup vendors have had several months to revise their offerings to take advantage of its new features, an upgrade to vSphere in 2010 could greatly improve your organization's data protection.
Here are 10 backup-related reasons why you should consider a vSphere upgrade in 2010:
1) VMware vStorage APIs for Data Protection. VMware introduced VMware Consolidated Backup (VCB) command line utilities in ESX 3.0 to simplify data protection by offloading backup from ESX Servers to one or more centralized proxy servers. It allows a live system image to be captured without disrupting the VM-resident applications or overwhelming the host's CPU. With VCB, backups occur off-host so there's no "backup window" to contend with. This method also removes the need for an agent in every virtual machine and provides VM-level recovery. But it requires either a two-step backup or recovery to perform a file-level recovery. In vSphere, VMware completely rearchitected the facilitation of backup. Instead of acquiring VCB and implementing it with your backup application, your backup vendor now leverages vStorage APIs (virtual machine file system, or VMFS, drivers) to access data on VMFS volumes. This integration makes for a much better backup implementation.
2) VMware Data Recovery (VDR). For those who prefer to use a native backup tool, VDR is an agentless, disk-based (local or Fibre Channel, iSCSI or network-attached storage shared disk) backup solution that employs snapshots and data deduplication. Running in a virtual machine as a Linux virtual appliance, VDR backup and recovery tasks are launched through VMware vCenter Server. Because VDR takes advantage of the vStorage APIs for Data Protection, a VM-level backup is performed that enables fast backup and recovery. Granular (file-level) recovery is possible from the VM-level backup. Using block-level data dedupe, it keeps required disk space to a minimum. VDR is included with vSphere Enterprise Plus, Advanced and Essentials Plus editions.
3) Granular recovery. File-level recovery is now possible from a VM-level backup without a two-step process. This allows organizations to perform an image-level backup (with no impact on the CPU of the host system) and to later do either a bare-metal-like or individual file recovery.
4) Changed block tracking (CBT). In the past, copying VMDKs for backup purposes meant moving large files around your LAN, WAN and/or SAN. The effect? An increase in network traffic and backup time. VMware introduced CBT in vSphere 4 to markedly improve performance. CBT allows the VMkernel to track changed blocks of a virtual machine's virtual disk. Backup applications using vStorage APIs can immediately identify the blocks that have changed since the last backup and copy only those blocks -- cutting down the time it takes to capture and transfer data, as well as reducing traffic on the network.
5) No physical proxy server required. Previously, a physical machine was required for the VCB backup proxy server. But organizations can now reduce their infrastructure commitment because the proxy system that mounts the LUNs can be a virtual machine.
6) vCenter Server plug-ins. Monitoring and management improvements are gained via vCenter Server integration. Backup vendors offering vCenter Server plug-ins can provide views of a virtual machine's backup status, including success/failure, method of backup and schedule. Some even automate the process of assigning backup policies to new VMs as they come online, eliminating protection gaps.
7) Direct backup from shared storage. When the backup data path includes a proxy system, it can create network bottlenecks and inefficiency. Now, a direct-to-target architecture is employed, allowing backup data to go from the ESX host to the storage target. Leveraging the vStorage APIs, virtual disk data is read directly from the SAN storage device.
8) iSCSI enhancements. Organizations using software iSCSI initiators to implement shared storage in virtual server environments could see performance improvements with vSphere. More disk I/O throughput and reductions in CPU overhead benefit I/O-intensive backup processes.
9) Thin provisioning. VMware made some improvements in vSphere for thin provisioning: the ability to overcommit disk space. Not only is it easier to create, maintain and monitor "thin" disks, but using them could help improve backup performance, as well as the capacity of space used on host's data stores. Here's why: When a VM is created, storage is allocated. Virtual machines typically don't use all of the storage space assigned, so the VMDK file could be unnecessarily larger. Backup applications are able to recognize "empty" file system space to avoid backing them up, but the process takes up valuable time. Thin provisioning results in smaller VMDK files and faster backups.
10) Data deduplication. Data deduplication is not a feature of vSphere unless you use VDR for backup. Therefore, it's likely that you will look to your backup application provider for this feature. Since taking advantage of vSphere features, such as CBT, vStorage APIs for Data Protection and vCenter Plug-ins, requires you to also upgrade your backup application, you'll likely gain dedupe features in the upgraded backup application. Given the tendency for duplicate data in backup processes, eliminating duplicates -- at the "source" VM or the proxy server -- will reduce the amount of data transferred and stored, and improve performance.
BIO: Lauren Whitehouse is an analyst focusing on backup and recovery software and replication solutions at Enterprise Strategy Group, Milford, Mass.