SNIA: Toward The Information Utility - Storage Networking Industry Association on storage area networks - Storage Networking Industry Association - Technology Information

Computer Technology Review, Sept, 1999 by Fred G. Moore

Vision is not seeing things as they are, but rather seeing things as they will be. Fifty years ago, people would often store their money in their home under a mattress or in a box. Even though banks were common and gaining popularity, many thought that the risk was too high to let someone else keep their most valuable asset: their money. But the banks had vision, they realized that if they could guarantee that people could more safely store and retrieve their money than they could at home and that it would always be available, they would create a financial storage utility that people everywhere could access. To further encourage new business and growth, the banks decided to add value to the money deposited and increase its value by offering interest and dividends. It wasn't too long before the banks became a financial utility for virtually everyone.

The banking model also serves as an example for the exploding data storage industry. As money was the most valuable asset for the individual a half century ago, the next century promises that data will be the most valuable asset of the information-based society. Can we guarantee that someone can safely store and retrieve their data today? Not just yet, but we are getting closer. The IT industry has undergone an evolution in data storage since its beginning. Throughout the 1970s and 1980s, storage deployment could be described as a Many-to-One with all storage devices connected to the server, typically a mainframe. During the late 1980s and 1990s, a one-to-one relationship developed between the server and storage devices.

Devices were dedicated to a particular server, as departmental and distributed servers became widespread. Servers and storage were co-located. The LAN had now materialized. This decentralized approach created islands of computing and storage that were costly to manage and replicated hardware, software, and personnel across these isolated nodes of computing and data. Most often in this computing model, data was not accessible to another server in the LAN in case of failure and applications that weren't in a clustered or an SMP (Symmetrical Multi-Processor) designed system were unavailable until they could be restarted. The most critical and valuable element of the enterprise, the data, was too often unavailable and the costs for an unscheduled outage remained severe for most every business. Like the bank, the evolving storage industry also has a vision and it has arrived in the form of the Storage Area Network (SAN).

The long journey for the SAN is just beginning, but it promises to bring to the Information Age a roadmap needed to transform islands of data to the level of a true utility, universally accessible, sharable, and ubiquitous. The SAN may be ahead of its time, but is not ahead of its vision. The longer view for SANs significantly challenges and eventually replaces the traditional practice of using a general-purpose server as a storage repository. The high-speed network fabric constructed from storage interfaces connects a pool of data storage devices to a group of servers and allows storage devices and servers to be added and removed independently of each other. In the SAN, data is shared rather than connected to and owned by a server. An unscheduled outage on a server need not affect access to data in the SAN model.

Where are we on this journey? Is the roadmap developed and supported by all of the providers of SAN components? Do we presently have agreement on standards, the network fabric, security rules, interfaces, common storage format, and the necessary software to manage the SAN resources and standards as the banking industry has? Answers to these questions actually form the basis for the SAN roadmap and most of these issues remain unanswered. SNIA (Storage Networking Industry Association) must be successful. We must make it work. Our choices are yesterday or tomorrow. SNIA gets us to tomorrow. Without quick and effective work from this critical group for the storage industry, we will arrive at yesterday's model and will spend years trying to tie numerous proprietary-computing islands together. Today, we continue to tactically debate over Fibre Channel, SSA, and SCSI interfaces as to which one is better for the SAN fabric. Longer-term, it doesn't really matter which interface we choose since interfaces will continue to evolve, improve, and be replaced. There will be several choices. Will wireless ever become a SAN interface? Step out of the box for a moment. Imagine in our vision the day arrives that the wireless SAN (another fabric) evolves and each one of us becomes a node on a SAN. Just as an example, biometrics is the discipline of life measurement and early tests are underway to help patients, athletes, and airline pilots, among others, monitor bodily conditions. Using wireless and wearable computing appliances, continual biometric feedback about our physical condition will be constantly sent, analyzed, and returned to us to warn of critical bodily conditions.