Storage System Interoperability.
Interoperability. For those creating an enterprise storage solution, just the word stirs great emotion. For most, the painful experience of integrating storage system components from a variety of manufacturers that have not been thoroughly tested together, interoperability conjures up memories equivalent to root canals, medieval torture, or worse, but for those who have successfully integrated a system because of either their own interoperability testing or a manufacturer's thorough interoperability program, come feelings of empowerment, dominion, confidence, and, most importantly, peace of mind.
How can one word create such extreme and polarized responses? The reason is that little is more critical to the success of an enterprise storage solution than the interoperability between components.
The need for thorough interoperability testing is not a new revelation. Over the past few years, articles addressing the concerns of interoperability have been written ad infinitum. Rarely, however, does an article provide the practical solution to the interoperability problem. Who should take responsibility for interoperability testing? Should the integrator be forced to incur the expense and time of testing every component and configuration? Or should the manufacturer provide a list of qualified components that they have tested as being compatible with their product?
Today, interoperability testing is even harder as storage technology continues to develop at an exponential rate and every component of an enterprise storage solution is breaking new ground, pushing technological limits. Even though individual components may be built within standard industry specifications, that does not guarantee that those products will work together. The required technological advances of the computer industry are increasingly complex and as the need to process more and more data grows faster and faster, the demands on each component grow, as well.
So who should take responsibility for interoperability? An enterprise storage system includes a wide variety of components from disk drives and disk drive enclosures, a RAID controller and the memory inside the controller, to the host bus adapters, UPSs, hubs, and switches. Out of all of these components, only one is centrally located in between the servers and the disks. That component is the RAID controller. Located at the heart of the storage solution, the RAID controller interacts with every device. No other component has such universal interaction as the RAID controller. No other component can so clearly define the interoperability of the entire storage system.
Even though today's market is ever increasing the demands of the enterprise storage solution, no matter how fast the data is written, or how much data is stored, the data is useless, unless it is always available to be read and the data being read is accurate.
Due to these demands for high data integrity and availability, the RAID controller, an acronym for Redundant Array of Independent Disks, has quickly secured its position as a necessary component in all levels of storage systems. Depending on the applications, RAID controllers are now found in everything from the simple desktop to the grandest of computer systems. By providing redundancy throughout the storage system, a quality RAID controller can maintain data availability and integrity even through the loss of key components such as a disk drive, power supply, or a RAID controller itself.
In a storage system, the RAID controller sits in the center of activity, receiving data, securing the integrity of the data, and then writing the data to disk. The RAID controller also monitors other key components within the storage system such as the disk drives. If a disk drive were to become inoperable, the RAID controller not only alerts the user, but also maintains the data availability by using either the mirrored or parity data to provide uninterrupted access and rewrite the "lost" data from the downed disk to a spare disk designated for that purpose.
In today's marketplace, however, few RAID controller manufacturers have aggressively addressed the interoperability issue. More common are manufacturers that develop a product and leave the interoperability testing to the integrator. The RAID controller manufacturers that do interoperability testing often require that the entire storage solution be purchased from them, eliminating the value and flexibility of being able to choose and integrate individual components. Although the time and effort required to perform a thorough interoperability program is great, the rewards for the end-user, integrator, and component manufacturers alike are even greater.
Thorough interoperability testing by the RAID controller manufacturer optimizes the development of the storage system for both the component manufacturers and the integrators. For the integrator, interoperability testing allows attention to be focused on the end-user's specific needs and goals. By assuring the integrator of performance tested interoperability between storage system components, the integrator has the confidence to build the very best storage solution without spending the time, effort, and money to test each component before deciding on a final configuration. In short, manufacturer interoperability testing means lower expense, therefore, increased profits for the integrator.
For the component manufacturers, a good interoperability program by the RAID controller manufacturer strengthens the working relationship between component manufacturers especially during the development process. This working relationship is essential because no product in the computer industry remains static, but is forever changing. Often, it is during the initial development of a product that the interoperability testing begins. Therefore, the interoperability testing can be used as a pseudo beta-site for both the RAID controller manufacturer and the other component manufacturers. During the development stages of a product, issues that arise during the interoperability testing can be identified and corrected.
Although qualifying a component as interoperable with a RAID controller requires extensive research and time-consuming testing, this is but the first step of a thorough interoperability program. Each product qualified will undoubtedly experience upgrades, and whether it is new firmware or hardware, or an entirely new product based on the existing architecture, complete testing must be performed. Once again, a strong relationship between manufacturers insures that any incompatibility issues are resolved before the upgrade is released to the field.
No one questions the need for quality interoperability testing. One would be naive to believe that just because a product claims to meet industry standards, it is compatible with the rest of the components in a storage system. Specific testing must be done. The question becomes who should perform the tests. The integrator has often been saddled with this responsibility, but it is the RAID controller manufacturer who needs to grab the reins, for it is the manufacturer who can resolve compatibility issues long before a product reaches the integrator.
When it's a high-availability, high-integrity storage system that is sought, the cornerstone to begin the building process is the RAID controller. Using a RAID controller that has already developed a thorough interoperability program will prove to be the most effective and profitable solution because the functionality of the RAID controller includes all of the other components in the storage system.
Robert Wood is the supervisor of technical writing at CMD Technology (Irvine, CA).
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Technology Information|
|Author:||Wood, Robert L.|
|Publication:||Computer Technology Review|
|Date:||Feb 1, 2000|
|Previous Article:||ask THE SCSI EXPERT.|
|Next Article:||SAN Gets The Ink, But NAS Does The Work.|