Without a doubt, disk backup systems are now a key staple in backup infrastructure, while tape has become a storage...
option for specialized needs.
Taneja Group lumps disk backup systems into two main groups. A virtual tape library (VTL) is a system that emulates a tape library and typically operates over Fibre Channel (FC). The second approach is to backup to a storage device such as a NAS appliance. This approach has sometimes been called “disk-as-disk” backup, as the system does not appear as tape to the backup software. The small- and mid-range market, including branch offices, have shifted almost entirely to the latter approach. And now, there is an emerging class of disk backup appliances, such as Symantec’s 5220 and Quantum’s vmPRO 4000, that ship preloaded with backup software and are optimized for backup and restore.
Backup purchasing criteria have also moved significantly beyond just performance. IT architects now realize that backup storage is a critical part of the infrastructure. And business managers are realizing that backup storage is a key center of both capital and operational expenses within the IT infrastructure.
Disk backup challenges
Acquiring backup infrastructure has always been plagued with the challenge of accurately forecasting needs of the business three years or more from date of purchase, and then trying to right-size that infrastructure for that point in time. Traditionally, tape systems were bought with overprovisioned performance, but could be expanded in capacity in the future through either core library expansion or the export and archive of tape.
Today, disk backup systems still require overprovisioning of performance, but often come with fixed-capacity configurations that add another limitation to the backup infrastructure. This increases the likelihood that customers will exceed the usefulness of the system before a targeted end-of-life date. Moreover, disk-based protection can increase some operational costs by a significant degree.
Even with good forecasting, the limited capacity of disk backup systems, increases in disk capacity, and performance requirements mean disk systems are replaced more often than tape systems.
Along with this comes the operational pain of migrating data. Disk drives are also more subject to wear and tear than tape media. Plus, using disk consumes more power and requires more cooling than an idle tape cartridge in the data center.
Those challenges can offset the gains from high throughput, simple deployment, and ongoing ease of use. They’re also driving factors in how customers decide on new backup systems. Let’s take a look at emerging capabilities that can offset some of these challenges.
Editor's note: This chapter excerpt offers important technology considerations for modern data backup and recovery systems.
Disk backup systems continue to evolve
With adaptability in mind, a number of vendors are pursuing paths to easier and simpler scalability. Many first-generation grid approaches suffered from sometimes being little better than capacity-aggregation technologies. But vendors like Symantec -- in both software and hardware solutions -- are innovating around deduplication algorithms and balancing scale, in terms of both performance and capacity, against efficiency, and in turn delivering truly scale-out solutions.
In reality, small trade-offs in efficiency can make an expensive storage asset much more scalable, and it will in turn last much longer, with minimal additional operational cost. Simultaneously, other vendors, such as Quantum, have recently introduced capacity-on-demand licensing that delivers unusually high performance with initial small capacity configurations. Customers can add capacity without worrying about racking, cabling, or reconfiguration complexity.
Meanwhile, nearly every vendor has been pressing forward with technologies that can reach beyond the boundaries of a single box. These technologies let businesses scale their deduplicating performance by harnessing server compute cycles or just more efficiently harvesting data from physical or virtual servers.
Symantec’s OST, EMC’s Data Domain’s Boot, and Quantum’s DXi Accent are all software components that are coupled with disk-based appliances and work to extend deduplication scale and efficiency beyond the disk-based appliance. But with this idea of more efficiently connecting the deduplicating appliance to the infrastructure, scale is not a single-facet challenge.
With scaling data protection for the virtual infrastructure being one of the stickiest challenges, vendors like Acronis, EMC, Quantum, Symantec, and Veeam offer their own take on reaching into virtual systems and capturing data more efficiently with either virtual server-specific protection technologies or source-side deduplication. In highly virtualized environments, these technologies may compete with other scalability technologies and go further in allowing the backup storage appliance to be used beneath more workloads and applications. For this reason, we’re seeing some vendors directly integrate these technologies. Case in point, Quantum’s repackaging of its Pancetera acquisition as DXI-integrated vmPro.
Editor's note: For our latest disk backup system news, check out the disk backup section of the site.
Disaster recovery functionality in disk backup systems
Meanwhile, leading vendors in this space continue to focus on protecting these precious disk backup systems. With the possibility of more frequent device failure, data loss, or compromised performance, data protection and redundancy are more important than ever. Underlying disk-failure protection that doesn’t compromise performance during device failures should be the customer demand, but few vendors can deliver on this promise.
Instead, the best alternatives today revolve around protecting disk backup systems with complementary technologies. As one example, systems that can vault to tape or the cloud have considerable value in the enterprise. Taking aged data offline may not substantially impair the majority of recovery activities, and may significantly reduce backup software licensing costs while adding capacity that requires lower-cost maintenance. It also provides the ability to offsite media easily and affordably.
As another example, disk backup systems can leverage replication between devices to protect data on more than one system or across sites. But such replication may vary in sophistication. Solutions should transmit with high efficiency over precious WAN bandwidth, and should allow sophisticated mixes of right-sized devices to suit business needs in the future. WAN efficiency is not always a strength of D2D and VTL vendors, and WAN optimization can reduce the amount of data D2D and VTL appliances send across the WAN.
Editor's note: Learn more about how organizations employ replication for disaster recovery.
Next-generation solutions offer plenty of opportunity to move disk backup systems toward higher levels of reliability, flexibility, and create more sophisticated backup infrastructures.
While some capabilities have begun to seem common with fewer notable differences between vendors, those abilities offer a clear opportunity to overcome first-generation device shortcomings. They can also make full use of the bandwidth and ease of use of these systems with even more competitive capabilities and operational cost containment.
Jeff Boles is a senior analyst at Taneja Group. He can be reached at email@example.com.