Storage systems that have file-level awareness (NAS and CAS) can perform operations on data independent of applications. These operations can include performing searches on documents that may span different applications but are all related. Search capability for file system-based storage should become a major initiative over the next few years. Customers create files using various applications that may all be related to a project, a customer, a patient or a business. Being able to globally search for them is extremely efficient especially as digital archives continue to grow.
Another major benefit that can be provided on NAS or CAS systems but not as easily on SANs (never say never) is de-duplication of data. This is a new concept that we see being implemented primarily with disk-to-disk (D2D) backup and CAS active archive products. The concept is simple -- there is a large amount of duplicate data being stored on the same storage system. Instead of storing all of that data, delete all but one primary version of the data (and keep a copy of it just in case) and keep an index of all the others. It is not uncommon for customers to have a 20-to-one reduction of data. In other words, they are storing 20 terabytes (TB) of data on 1 TB of disk capacity. Think about the economic and operational implications of this. It changes the game entirely.
Today, there is a bright spotlight on D2D backup products. However, we should consider why, if the data is already on disk, it is being backed up to disk? In a SAN environment the answer is very clear: In order to recover data you need backup software to retrieve any type of data from any type of application. But with file system-based storage the answer is not as obvious. You can make snapshot copies of the data and restore them easily without having to use any backup application. Why add an extra step and more storage systems? The missing piece is that there is no single GUI or engine that allows customers to recover data across all of the file system storage.
File level awareness also allows for policies to control and manage data. For example, write once, read many (WORM) is being used in conjunction with retention policies for corporate governance and to meet compliance regulations. Additionally, other policies can be set to transparently move data between different tiers of storage or replicated to remote locations for greater data protection. More sophisticated policies and controls can be put into place as software vendors and storage systems companies work more closely together.
Having said all of that, SAN is not going away any time soon. Fibre Channel (FC) SANs are tried and true, and deployed throughout the world supporting mission-critical applications. Additionally, FC SANs are considered the high performance king of the hill. SANs support any application including databases, e-mail and file storage. Customers, vendors, system integrators, professional services organizations and application providers have all invested time, money and sweat equity into SANs. And the fact that SAN technologies and products are so widely implemented, qualified and proven is extremely valuable. Most customers want to make the safe choice and not the cool choice. iSCSI drives down cost and complexity, which should also further adoption of SANs. Additionally, there has been a great deal of innovation with SAN products including differential snapshots, thin provisioning, clustered network storage and automated tiered storage.
Who will win -- the Athlete (SAN) or the Geek (NAS and CAS)? The answer is that all approaches will co-exist for the foreseeable future. Over time, the efficiencies of file system performance, continued improvements in processing power and the use of clustered network storage may overcome the performance delta between SAN and NAS/CAS. NAS and CAS will continue to add functionality and leverage its superior intelligence. SAN will continue to be considered the high performance storage network that is field proven with enormous adoption.
About the author: Tony Asaro is the senior analyst for Enterprise Strategy Group.