But the excitement in VTL isn't just an exercise in public relations -- industry analysts also see the inevitable march from tape to disk. A recent survey from the Enterprise Strategy Group in Milford, Mass., reports that over half of early VTL adopters expect more than 60% of their current tape capacity to ultimately reside on VTL and near-line disk storage systems. With so much activity in the VTL market, it's time to get a handle on the technology and consider the role of VTL as a tape backup supplement in your own organization.
Disk masquerading as tape
Tape backup still has a place for long-term archival storage, but tape is under increasing scrutiny in organizations demanding faster and more reliable backup/restoration processes. However, not every enterprise is willing to simply abandon its established backup software, backup policies and in-house tape expertise. VTLs are emerging as a popular answer to this dilemma -- allowing organizations to reap the performance benefits of SAN-based disk storage masquerading as a conventional tape backup system.
VTL is mainly touted for its convenience and performance improvements over traditional tape. Where ordinary near-line disk storage simply appears as a mass of storage capacity, a VTL organizes disks to emulate specific tape drives and tape libraries. Existing backup software seamlessly "recognizes" the VTL as corresponding tape equipment so current backup software, such as Symantec Corp.'s NetBackup or EMC Corp.'s Legato, can remain in place -- with only minor configuration changes.
The real benefit is convenience. "It makes for a much easier transition into the disk backup world," says W. Curtis Preston, vice president of data protection at GlassHouse Technologies Inc. "You don't have to learn anything new -- just tell your backup software product that it has another tape library."
Since VTL relies on disk storage, backups and restoration procedures are typically faster than tape systems, though analysts are quick to note that actual improvements will vary with each implementation. "Restores should be much faster," says Greg Schulz, founder and senior analyst at StorageIO. Backups, however, may not experience the same benefits. "You may not be addressing your backup performance bottlenecks." Issues that are external to the backup system, such as network bottlenecks, may limit backup performance improvements even after a VTL is installed.
There are secondary benefits to consider as well. The shift to VTL eliminates streaming problems that often impair efficiency. "Modern tape drives need to be streamed," Preston says. "If it's a 50 megabytes per second (MBps) tape drive, you need to make it go 50 MBps." If the tape drive isn't kept full of data constantly, it must stop and start, which reduces backup speed and causes excessive wear on the tape. Preston notes that disk technology does not rely on streaming and can write effectively regardless of data transfer speeds.
As files move and file sizes change, the elements (clusters) of a file can eventually become scattered across a disk -- a side effect of the file system known as file fragmentation. Tape drives naturally overcome file fragmentation by gathering the elements of a file and recording them to tape sequentially. When files are simply copied from disk to disk, fragmentation can follow the file, and fragmented files take longer to restore since the disk must work harder to locate each cluster. VTL systems further emulate tape systems by recording files to disk sequentially. This avoids fragmentation, which can improve restoration performance.
Getting VTL to work
Analysts agree that VTL is not a panacea for every situation. VTL systems can be costlier than disk storage arrays, so each storage administrator must weigh the pros and cons to determine how well VTL technology can address weaknesses in their current tape backup process.
The first issue to consider is need. Businesses that are currently meeting their backup window, RPO and RTO with a conventional tape system may simply not see the benefit in VTL. VTL also doesn't make sense for a data center that has already moved away from tape in favor of other disk-based storage technologies. "If you're using a replication approach, or if you're already doing straight disk-to-disk backups, such as snapshots or CDP, VTLs may not be for you," Schulz says.
So where does VTL fit? According to Schulz and other analysts, VTL technology appeals to organizations that need to leverage disk storage for shrinking backup and restoration objectives but aren't yet ready to abandon their current investment in tape backup policies, practices and software. Disk storage overcomes the common tape insertion problems and mechanical errors that can easily interrupt a large tape job, so VTL systems can fit when backup reliability is a high priority. Disk storage also eliminates the use of tapes, so organizations that are sensitive to tape handling and storage issues may find VTL technology beneficial.
It's important to note that while VTL is certainly capable of eliminating tapes, it is almost universally used to supplement tape systems. "The vast majority of users buying VTL are still keeping tape," says Fara Yale, research vice president at Gartner Inc. "You're complementing your entire backup process by putting virtual tape between your primary disk and physical tape." In most cases, a VTL is installed as a "front end" to a tape system. The backup software transfers data to VTL disks, which speeds the process and retains backup data for a prescribed period. Backup contents are then periodically dumped to tape from the VTL without disrupting the production network. That resulting tape can then be taken off site for final archival storage.
Most vendors incorporate both hardware and software into their VTL product, though some software-only products can utilize existing hardware already in the storage infrastructure. The choice of hardware vs. software VTL products will depend on the level of integration that makes most sense for the enterprise. "The vendors that provide an integrated hardware/software solution say they have an advantage because they are able to "control" the environment better," says Heidi Biggar, analyst at the Enterprise Strategy Group. "Software-only folks (e.g., FalconStor and Diligent) say they have an advantage because they can integrate with existing or preferred hardware."
Although VTL adoption is still in its infancy, there are numerous vendors staking their claim to the marketplace. The largest VTL offerings include EMC's Clariion Disk Library, the Pathlight-VX from Advanced Digital Information Corp., the DX Series by Quantum Corp., HP's StorageWorks 6000 Virtual Library System and Sun Microsystems Inc.'s Virtual Storage Manager (VSM) and VSM Open, in addition to offerings from IBM and Overland Storage Inc. But there are also many smaller corporations and startups attracting attention, such as FalconStor, Diligent, Neartek Inc., Sepaton Inc. and MaXXan Systems Inc. Even backup software vendors, such as Arkeia Corp. and Atempo, are including VTL features in their backup products.
Analysts note that the last three years have been a period of development for VTL, and adoption seems poised to accelerate in 2006. "Now that users are more familiar with it [VTL], and the vendors are out there talking about it more, it's just really starting to take off now," Yale says. The products are out there, but VTL adopters must identify products that integrate well with their environment.
Picking the right VTL product can be problematic. Vendors like FalconStor and Sepaton are known for their software-based offerings, while most other VTL vendors provide hardware-based systems with software built in-house or OEMed externally. For example, EMC and IBM both use FalconStor software with their VTL products, while HP employs Sepaton software for their VTL servers. Although an argument can be made in favor of both approaches, analysts recommend a more turnkey selection.
Unless you've got the time to be an integrator yourself, you're going to work with a solution provider," Schulz says. "If you're out shopping for software and shopping for hardware, that says you're in the integration business. Do you have time to do that?"
Picking a VTL should involve a careful evaluation of attributes and decide whether or not these attributes meet your business objectives. While there are many possible features to consider, the most important will relate to performance and compatibility with your existing backup software. For example, know the maximum usable capacity. Since there are no tapes to remove and store, all data is held on disk so there must be enough VTL storage capacity to hold a complete backup and some period of incremental backups. This is often 15-30 days worth of data backups. Understand the backup and restoration speeds (e.g., Gigabytes per hour) since these will impact backup and recovery time.
It's also important to recognize the VTL's compatibility with your current backup platform. The VTL must successfully emulate the tape library or tape drives (e.g., LTO, AIT, DLT) currently in use. Otherwise, new backup software may be required and that would defeat the premise of maintaining your current backup investments.
Three emerging features to consider are compression, de-duplication and encryption. Compression stores more data on less disk space; deduplication identifies repeating data and saves just one iteration of that data; and encryption encodes data so that it can only be accessed by users with the appropriate credentials.
The future of VTL
Analysts agree that VTL technology is not poised for any dramatic or revolutionary leaps, though users should see a larger selection of features become available. Today, VTL is mainly seen as a convenient avenue to disk storage adoption without a major re-architecting of the backup infrastructure. "For some it's a long-term solution, but for others it's definitely a transitional technology," Schulz says. "It's a way of transitioning away from traditional tape."
In the near term, expect greater scalability for better capacity growth, and look for powerful features like data deduplication to decrease disk space utilization -- this makes VTL more cost-effective and attractive for long-term storage. Over the long term, however, VTL will likely morph into a more ubiquitous feature in general disk storage systems. For example, future disk arrays may be able to handle VTL just as they now tackle functions like snapshots or replication.