alphaspirit - Fotolia
When compared with disk backup, tape backup is inexpensive, but restoring data is slow. Jeff Boles, a senior analyst with the Taneja Group, a storage industry research and consulting firm, said disk backup can simplify your backup infrastructure. "You can get a fairly performant infrastructure with disk," Boles said. "A disk solution may solve many of your backup infrastructure throughput and performance challenges and simplify your overall solution." And simple is a word most SMBs like to hear.
Many small- and medium-sized businesses (SMBs) are searching for reliable, easy and economical data backup storage systems. In this small business data backup tutorial, learn about the pros and cons of using a disk backup storage system, whether or not a disk backup is right for your company, and disk technologies like virtual tape libraries and data deduplication.
IMPLEMENTING DISK BACKUP STORAGE SYSTEMS: A SMALL BUSINESS DATA BACKUP TUTORIAL
When Network Supervisor Mitch Watson first joined Hein & Associates LLP, a Denver, Colo.-based accounting firm, it had 160 employees, a total of 500 GB of disk space backed up to tape, and "big, huge rolling vaults of paper files." As the company moved everything to the network and grew to its current size of 280 employees, the backup window grew to 17 hours, and Watson knew he had to do something. "I knew disk-to-disk backup was quicker because you don't have the read and seek that a tape drive has," Watson said, "it just dumps it to the drive." After the initial full backup, Watson's daily disk backups only capture data changes, so he's "only grabbing small bursts of data from the endpoint," he said. That also keeps the systems' bandwidth requirements lower during working hours.
Hein & Associates uses Microsoft's disk-based System Center Data Protection Manager (DPM) 2007 for fast daily backups. "We could start the jobs at 11:00 p.m. and they will be done by three or four in the morning," Watson said. Disk backup also allows for quick restores because he doesn't often have to pull tape. "We can restore a 300 MB or 400 MB file in probably a minute," Watson said. His disk backup system holds 30 days of backups. He also uses DPM to synchronize a critical auditing application every 15 minutes over the company's virtual private network (VPN) so his field accountants can see each other's work nearly in real time.
Disk backup has many positive traits; however, tape still has its place. Watson said he still leverages tape's low cost by using it to store long-term backups. To comply with Internal Revenue Service (IRS) regulations, Hein & Associates must keep files for seven years, and Watson uses Symantec Corp.'s Backup Exec software for long-term tape backups.
Disk data backup does have some drawbacks. In addition to increased costs as compared to tape backup, Boles said disk backup does add an additional layer of complexity if you plan to keep vaulting tape off site. You will likely have to add data replication technology to your storage arsenal to maintain your disaster recovery (DR) strategy. "If you are looking to replace your offsite tape," Boles said, "that should be on your shopping list before you invest in a disk-to-disk solution, because the capability to efficiently replicate data will vary widely by vendor." Boles said you may have to invest in a backup appliance for replication technology.
The implementation of multiple disk-backup devices into an environment is also a potential concern. "Another disk-to-disk challenge is repository sprawl," Boles said. "It can be difficult to load-balance across multiple repositories, migrate data across multiple repositories, or even handle data in an integrated fashion when you are working with multiple repositories." Disk-device sprawl can be exacerbated by the fact that data in a disk-backup system can be difficult to move into long-term data archiving. "If you want to keep a data backup set for a year, you may want to get it out of your disk repository," Boles said. "And getting it out and keeping your backup application informed that it's out there somewhere maybe difficult to do."
Disk storage also offers access to some of today's advanced technologies, including server virtualization by making it easier to back up virtual environments. "Disk backup may help you stream stuff out of your virtual infrastructure more efficiently than tape," Boles said. Some technologies like PHD Virtual Technologies esXpress were designed specifically to back up virtual environments to disk.
Traditional virtual server backup software operates within the virtual machine, so it can restore individual files well, but require multiple steps to fully restore all of the operating systems, applications, and data contained in virtual machine disk format (VMDK) files. PHD Virtual's esXpress technology creates virtual backup appliances as VMs on VMware Inc. ESX servers and takes a snapshot of each VM on the ESX Server so it can restore entire VMDK files. Traditional backup software vendors are catching on to these purpose-built systems. Quantum Corp. offers PHD Virtual's technology as Quantum esXpress Backup Software for VMware. Other VM backup solutions include Vizioncore Inc.'s vRanger Pro Data Protection Platform and Acronis Inc.'s Backup & Recovery 10 Advanced Server Virtual Edition.
Data deduplication technology gets the most out of your storage capacity by eliminating duplicate data from your disk storage. Target deduplication allows you to use existing backup software by presenting itself to the backup server as a network-attached storage (NAS) share or virtual tape library (VTL) target. The backup data is run through the deduplication process then written to disk. Source deduplication requires you to adopt new backup software. The duplicate data is identified on the originating server and only non-duplicate data is sent across the LAN or wide-area network (WAN). It is ideal for remote and branch offices.
You can chose between a deduplication appliance and a software-based solution. Target deduplication uses a deduplication appliance that sits between the originating server and backup equipment. Deduplication software is used in source deduplication, sits on the originating server, and dedupes data before sending it to the backup equipment.
If you have more than one device that produces data for deduplication, you'll want a solution that offers global deduplication, which dedupes data from multiple devices and multiple sites. It effectively treats all of the devices and sites as one large deduplication system, which can increase your deduplication ratios.
Deduplication is not only available for disk backup storage systems. You can deduplicate data to tape, as well. However, deduplication to tape may require the additional step of reinflating data using disk before restoring the data. There may also be an issue of deduplicating data from a file that constantly requires updating and backing up. You may save money by using tape, only to spend that money and possibly more getting the deduplicated data restored.
There are solutions out there that include advanced services as part of their standard backup products. ExaGrid Systems Inc. includes deduplication in all its disk backup appliances, as does Quantum's DXi disk backup series.
Virtual tape libraries (VTLs) allow you to integrate disk backup into an environment that is used to seeing tape backup technology.
Irwin Teodoro, the director of systems integration for Laurus Technologies Inc., said virtual tape libraries can be used as a temporary fix for environments migrating to disk backup. "It's a stopgap type of technology," Teodoro said. "It's for organizations that still need to interface with tape while they migrate to disk based backup."
If you are looking to purchase a disk backup storage system, Jeff Boles said to consider both your current needs and what you might need down the road. "Plan your disk-to-disk infrastructure with an eye to the future," Boles said. "Don't buy the bare minimum because a lot of products on the market don't scale well." If you don't plan well for data growth and are forced to migrate to a more scalable system, it can be painful.