VMware's ESX server virtualization has changed quite a bit about the enterprise data center in the last five years,...
but one area where it has room to grow is in its integration with advanced data backup and recovery tools.
That's the conclusion from users and analysts, even as VMware technology partners in backup and replication work to integrate with new vStorage APIs for data protection released with vSphere 4 this year, with the goal of more efficiently backing up virtual machines without requiring the deployment of agents in each guest machine.
"At this point, I wouldn't be surprised if 80% of users currently have agents in guests," said data backup expert W. Curtis Preston. There are two ways to use the current VMware Consolidated Backup (VCB) option, which transfers the quiescing of applications and execution of snapshots to a proxy server. One is a virtual mount, which means that the C: drive of a guest machine is transferred to the proxy server and backed up using regular backup software as if it was a local drive on the proxy machine. The other is to send an image of the virtual machine itself to the VCB proxy, where it can be backed up in its entirety.
According to Preston, "at the least it's a two-stage backup and restore" process to use VMware Consolidated Backup. If a virtual mount is used, the user must have a separate tool. Popular choices include VizionCore's vRanger Pro and Veeam Software's Backup & Replication to back up the guest machine itself. If an image-level backup is used, the full Virtual machine disk format (VMDK) file is copied to the proxy machine and then backed up. Should users need to restore a single file or object from an image level backup, the full virtual machine must often be restored. Even incremental backups often copy the entire VMDK file over to the proxy server, Preston said.
Symantec Corp. has an offering it calls Granular Recovery Technology that's part of both its Veritas NetBackup and Backup Exec backup software products, which allows users to restore individual files or objects without re-mounting the entire virtual machine or requiring multiple backups of the virtual machine image. "Symantec has about the best that can be done currently," said Preston, with the most widely used form of VCB, with VMware Infrastructure 3 (VI3) environments.
vStorage API integration
Enter VMware's latest release, vSphere 4, which began shipping this summer with new vStorage APIs for data protection partners that amount to an effort to make all VMware backups work more efficiently. The new vStorage API for Data Protection will offer direct API-based integration for backups, eliminating the need for VMware Consolidated Backup. VMware Consolidated Backup required a customer to add another software module to the backup server, and then the customer or backup vendor -- or both --had to join the backup software with VCB through scripting. VMware will now open up APIs to its backup partners, allowing backup software to query virtual machines directly.
According to Enterprise Strategy Group (ESG) senior analyst Lauren Whitehouse, a new Virtual Disk Developer Kit (VDDK) will allow for image-level and file-level recovery in one "hop," incremental and differential backups, backups directly from networked storage systems without having to pull data over to the proxy server, and doing away with the requirement that the proxy server be dedicated to virtual machine backup. "It can be a backup server doing physical environment backup, too," Whitehouse said.
So far, few data backup vendors have vStorage API integration shipping within their backup products, though almost all say they have plans. Currently shipping products with VMware vStorage API integration include CA Inc.'s ARCserve Backup 12.5 (day-one support); FalconStor Software Inc.'s Network Storage Server (NSS); and PHD Virtual's esXpress 3.6 (snapshots only; full support will follow in version 4.0; Quantum Corp. OEM's esXpress). Veeam's Backup & Recovery 4.0, slated for release this fall, will also integrate with the vStorage APIs.
Products with roadmap plans to integrate with the vStorage API include AppAssure's self-titled continuous data protection (CDP) product; Asigra Inc.'s Hybrid Cloud Backup v9 and v9.2 Service Pack; BakBone Software Inc.'s NetVault; Double-Take Software Inc.'s Double-Take Backup; EMC Corp.'s Avamar and NetWorker; Symantec Corp.'s Backup Exec 2010 and NetBackup 7.0; and Zmanda Inc.'s Zmanda Recovery Manager.
Some vendors said they were still evaluating the new APIs. As VMware continues to enhance the API framework to deliver enhanced capabilities (backup to tape, application consistency, etc), NetApp will evaluate and provide integration into the VCB framework appropriately," wrote a NetApp Inc. spokesperson in an email to SearchDataBackup. CommVault Technologies Inc. and NetApp argue their current integrations offer the same benefits to customers as the vStorage API.
According to Eric Burgener, VP of marketing for continuous data protection (CDP) and disaster recovery (DR) player InMage, which is OEMed by Hitachi Data Systems (HDS) in another email to SearchDataBackup, "Going forward, as more and more companies deploy production VMs using Citrix and Hyper-V, we feel that this API will be less important as people will be looking for a single data protection solution [whether that's for backup or DR] that can simultaneously manage all three environments the same way."
Another feature of vSphere 4 is the ability to add software management "plugins" to VMware's vCenter management console, allowing elements of the underlying infrastructure, including backup software and devices, to be managed by the VMware administrator. One example of such integration is a plugin from FalconStor to manage NSS snapshots; most vendors that plan integration with the vStorage API plan a vCenter plugin option.
For small- and medium-sized businesses (SMBs), vSphere 4 will include VMware's own version of an application written with the vStorage APIs called Data Recovery. The product takes a snapshot of a virtual machine and then does a granular backup from that snapshot to remove the processing load from the production machine. Data Recovery can also do file-level recovery from a VMDK backup. It's aimed at smaller customers without a good data backup tool and will be limited to disk-based backups of virtual machine images only. It will support up to 100 virtual machines or 2 TB of storage.
However, ESG's Whitehouse sees Data Recovery as another potential point of friction for VMware and its ecosystem partners. "Data Recovery is kind of done in a similar model as PHD Virtual," she said. "VMware walks a fine line between proving value for its customers and implementing features in their product that may take away opportunities for some ecosystem customers to add value."
Virtual Desktop Infrastructure
Burton Group analyst Chris Wolf said that while users are understandably cautious when it comes to changes in their backup environments, there are incentives to upgrade data protection plans for virtual machines as technology matures. "Backup is serious stuff -- if you can't get files back, sometimes people lose their jobs," Wolf said. "But individual agents inside guest VMs can be very taxing on the CPU and disk I/O, which can mean a lower consolidation density with server virtualization, which can drive up the total cost of the entire environment."
For now, though, Chris House, senior network analyst at The MetroHealth System, a US hospital network, said he's sticking to agents running inside his 160 virtual guests. "Our virtual machines aren't very large in terms of disks, and [IBM's Tivoli Storage Manager] TSM already does incremental backups by default," he said. "There's a comfort level with it -- all our physical servers have a TSM client, and it makes sense to do the same for virtual machines."