Although object-based storage systems have been on the market for only a few years, they have been relatively successful, in terms of total number of sales and overall capacity sold. Customers as well as the press have shown significant interest in these products.
Vendors currently selling object-based storage systems continue to improve their functionality, while additional competitors continue to enter the market. But while the opportunity to prosper in this market remains substantial, IT organizations' deployment of object-based storage systems is still in the early stages.
Object-based storage systems take a new approach to storing data. The most familiar storage system implementation is block storage, wherein blocks of data are sent to the storage system from the host over an interface, and the identity of the data is based upon the volume and the logical block address. File-based storage systems are really remote file systems, where a storage system (typically NAS) stores data as files. In reality, the NAS device or file server turns it into block storage.
With object-based storage, the data is stored as an object with the application that stores and retrieves data defining objects. This creates new capabilities in dealing with objects that can be exploited by applications and management software. An example is the use of content addressable storage (CAS) to eliminate multiple copies of data being stored.
By dealing with objects and not the specific physical placement requirements of block storage systems, the object-based storage system should have some self-management capabilities regarding data placement and access, relieving storage administrators from that task.
The metadata kept about objects is really the key to enabling new capabilities for object-based storage systems. The content of that metadata is both information (attributes) that the storage system adds, such as size, date, access, etc., as well as information that the application includes for use by applications. The metadata can be used for content searches, use/access identification, regulatory compliance requirements and other high-value functions. These functions usually require exploitation by software (from ISVs) through an API for the object-based storage system. This can be viewed as a specialized form of I/O.
The success of object-based storage systems is growing. Several heavyweight storage vendors have products with impressive sales, led by EMC's Centera. A big contributor to this success is the number of ISVs that have changed their software to use object-based storage product APIs.
For those vendors that haven't been able to get ISVs to use their APIs, the mapping of object-based access to NFS or CIFS file protocols has enabled unchanged applications to use their object-based storage systems. The mapping for NFS and CIFS access may be done as an integral part of the system or through another server of some type in front of the storage system.
One problem is that the APIs differ from product to product. Some standards work has been done, but advanced capabilities are being exploited on a non-standardized basis. Without a common method of exploiting the capabilities, there will continue to be different versions of application software for different vendor products, which will skew the value of ensuing products.
The capabilities that customers are most heavily exploiting are CAS, regulatory compliance based on metadata information and content search and retrieve. The benefits of these advanced system functions have become competitive product features.
Looking at the growth of object-based storage to date, it is easy to predict ongoing and expanding success for the future. The wealth of emerging opportunities will encourage more vendors to add and differentiate their products. The most significant development will likely be the added exploitation of object-based storage capabilities by application software and, even more importantly, by operating systems. Beyond that, object-based storage will move into other market segments, such as the small-to-medium-size business arena, as prices drop.
The major vendors will eventually all have product offerings. The key to success will be the specific applications that can exploit the capabilities. Over time, standards that include APIs will eventually be a necessity for vendors, as well as a boon for customers. New capabilities will include necessary functions such as standardized backup and remote replication.
Although the integration of object-based storage with storage grids has plenty of positive potential, this marriage will be confusing to customers. Grids may initially be delivered as file-based storage, but object-based storage technology will quickly become a major part of this technology. Unfortunately, the increasingly competitive market for object-based storage will spawn a plethora of new and proprietary "definitions" of what object-based storage is and isn't.
By all measures, object-based storage systems have had a successful entry into the storage market. But the products are still in the early stages. As they mature, more capabilities will be added that will bring value to customers, and usefulness will expand beyond the initial application areas.
For customers, object-based storage systems represent great opportunities for advanced capabilities to help automate storage management and meet new requirements. For vendors, they provide new opportunities for sales in not only current storage markets but also for information that may not have traditionally been stored online in the IT environment.
In the end, object-based storage systems will represent another solid choice when customers develop their storage strategies.
Randy Kerns is a partner at The Evaluator Group. He can be reached at firstname.lastname@example.org.