We had an overwhelming response to our call for predictions regarding the future of enterprise data storage. Here are some of the prognostications we collected:
Three years from now, the terms SAN and NAS will be meaningless. SAN and NAS will evolve into a single technology that includes the benefits of both. This is the same thing that happened with routers and switches in the networking world. Today, almost nobody understands the difference between a router and a Layer 4 or a Layer 7 switch -- and almost nobody cares! People have already figured out that any serious storage infrastructure requires both SAN and NAS. In three to five years, customers will insist on multilingual systems that can speak block-based protocols, like Fibre Channel and iSCSI, as well as file-based protocols.
-- Dave Hitz, founder and executive vice president of engineering, Network Appliance Inc., Sunnyvale, Calif.
Automated networked storage is the way of the future. Over the next three years, direct-attached storage will play a diminishing role in the enterprise. The pace of decline will be faster than most analysts predict. Today, it's unusual to find a server that isn't networked. In a few years, it will be just as unusual to find storage that isn't networked.
-- Joe Tucci, president and CEO, EMC Corp., Hopkinton, Mass.
I envision a day when you walk into an airport and your handheld device hooks into the airport network and tells you your gate number, seat assignment and departure time. You'll even be able to relay to the airline whether you want the chicken or the fish for your in-flight special meal. All this will require a mountainous amount of data that will need to be distributed, cached, stored, managed and protected. Storage management software will need to be flexible and robust and have the ability to interoperate with a variety of different platforms.
-- Jose Iglesias, director of storage products, IBM Tivoli Software, Austin, Texas
E-mail servers and their related storage as we know them today will be dead in three to five years. The costs and management of e-mail servers continue to spiral upward, while today's PCs are more storage-rich than ever before. Their storage capacity is "free" and yet relatively untapped. Free PC storage, combined with aggressive server-to-PC e-mail migration and better PC data management, will solve storage problems at a much lower cost to corporations than traditional, centralized server approaches.
-- Dave Cane, chairman and senior vice president of research and development, Connected Corp., Framingham, Mass.
Government regulations in health care (the Health Insurance Portability and Accountability Act) and financial services (Gramm-Leach-Bliley Act) are creating a strong need for companies to archive much more information than ever before. Massive amounts of data need to be moved to storage rapidly and managed at a much lower cost. For at least the next decade, we're going to see a fast-growing demand for software that saves bandwidth and dramatically shrinks storage requirements at the same time. Just as cheap gasoline hasn't reduced the pressure on carmakers to get better mileage, cheap storage isn't going to reduce the pressure to get more out of the storage and bandwidth that's already in place.
-- Gordon Dorworth, president and CEO, Stampede Technologies Inc., Dayton, Ohio
Ultra Density Optical (UDO) drives, the first storage devices to use a blue laser, will debut in late 2003 and provide 30GB in a 5.25-in. magneto-optical form factor as well as transfer speeds up to 8M bit/sec. The relatively high cost and slow seek times will relegate UDO primarily to write-once applications that are increasingly important for regulatory requirements.
-- Michael Katz, managing director of PricewaterhouseCoopers LLP's Global Technology Centre, Menlo Park, Calif.
The widely held misconception is that end customers don't really care about storage. The new and preferred fashion for an information-drenched, -driven and -obsessed economy is customer-centric information storage. Record-keeping (and the infrastructure that supports it) is the key point of differentiation in all manner of services ranging from banking to cell phones. Storage is about ready to come out of the closet as a key corporate differentiator. Royal Bank in Canada is head and shoulders ahead of competitors in understanding this and providing services to customers seeking to control how information about them is managed. This is the high-margin sweet spot of the future.
-- Thornton May, futurist, Biddeford, Maine
Storage capacities in typical environments continue to grow over 60% per year. Clearly, IT staff is not growing at nearly this rate. On top of this, the performance, availability and recoverability requirements placed on these storage infrastructures are more stringent than ever. As a result, there is a looming storage management crisis, with a need for creative architectures and a new class of storage management tools that dramatically improve the efficiency and productivity of IT staffs.
-- Scott Robinson, chief technology officer, Datalink Corp., Minneapolis
If we continue with the current evolutionary migration path of linear recording and helical scan magnetic tape technologies, we see the capability of delivering a 1TB tape cartridge by 2008. However, the successful development of new read and write technologies could enable the delivery of a 1TB tape cartridge even sooner.
-- Saurin Shah, director of advanced technology and applications, Imation Corp., Oakdale, Minn.
While storage media are making huge strides, significant advances are also being made on the software side to minimize file size. New compression file formats are already able to reduce file sizes by 500% to 1,000%. For example, a document stored on the NASA Web site can currently be downloaded as a PDF file at 147MB. Given the latest compression file formats, this same document could be downloaded at a mere 3MB. The next five years will see a continued drive for even greater file compression and hardware storage capabilities.
-- Joe Wharram, president and CEO, Computhink Inc., Lombard, Ill.
On top of the list of ideas that have run their course (for example, the earth is flat and accountants are working for shareholders) is the misplaced belief that data storage will take care of itself and that data accuracy and management are the sole responsibility of some wonk deep in the bowels of the IT organization. Storage management will be a key IT competency in the very near future.
-- Thornton May, futurist
Today companies manage networks with a relatively mature set of tools, while these same companies manage SANs with a relatively immature set of tools. Today companies manage networks and SANs with two distinctly different organizations. Tomorrow companies will manage the IP and storage I/O transport with a mature set of tools and within the same organization.
-- Marc Duvoisin, director of enterprise services, Dimension Data Holdings PLC, Atlanta
Research and development of materials and systems for storing large numbers of images in the volume rather than on the surface of a media has been under way for several decades. Recent advances in polymer materials are beginning to demonstrate promise, but there are significant challenges to commercializing these technologies for general use in data storage applications. Continued industry investments in this field could result in commercial holographic recording systems for data storage applications by 2005 or 2006.
-- Rusty Rosenberger, director of business development and product line management, Imation Corp., Oakdale, Minn.
By 2005, SCSI vendors will begin selling high-performance 2.5-in. drives for transaction processing applications. These will provide a combination of desirable attributes, including low-power consumption, fast access times and capacities in excess of 200GB per platter.
-- Michael Katz, PricewaterhouseCoopers
Within the next three years, enterprise storage will be fully automated. When assets are put in place, they will be automatically discovered, policies applied, and provisioned into the IT environment -- all without human intervention.
-- Gary McGuire, senior vice president, Computer Associates International Inc., Islandia, N.Y.
In the next few years, autonomic storage software will enable simplified storage hardware, accommodate heterogeneous files and protocols, automate more administrative functions (such as data placement and backup) and will proactively react to workloads by making changes to maintain customers' policy-based performance and availability goals. A bit further in the future, predictive capabilities will anticipate swings in workload, enabling the proactive adjustments to be even more effective and invisible.
-- Jai Menon, IBM Fellow, IBM's Almaden Research Center, San Jose
Serial ATA drives [which feature plug-and-play connection, low cost and high transfer rates] will account for the majority of new desktop drives sold beginning in 2005. The Serial ATA II server drive specification, to be finalized this year, will give rise to EIDE drives that can be connected via storage bus by 2004.
-- Michael Katz, PricewaterhouseCoopers
Refrigerators and tungsten lighting are nowhere to be seen in the data hub -- the year is 2005. Tiny network nodes hover to and fro silently controlling and monitoring inexpensive, subservient disk trays. Much sheet metal has been recycled. The mesh of switches connecting applications to spinning media is slave to software. Downtime and accompanying expletives have been retired from IT's unabridged dictionary.
-- Ziya Aral, chief technology officer, DataCore Software Corp., Fort Lauderdale, Fla.