Its performance and packaging continue to improve. But, with the exception of grid computing-which is still largely a prisoner of its own hype--server architecture is stuck serving the status quo. Likewise, networks may be getting bigger, faster, cheaper and smarter, but there is nothing revolutionary on the horizon.
On the surface, the current state of innovation in storage technology also appears to be advancing in first gear. Over the next three years, however, new technology will redraw the established order in the storage industry and force the reinvention of IT storage strategies.
What's behind this coming transition?
For starters, hardware components like microprocessors, application-specific integrated circuits (ASICs), disk drives and interface cards are dirt-cheap. These hardware goodies enable companies to build champagne storage devices at off-the-shelf beer budgets. This new breed of storage equipment will offer sophisticated capabilities such as failover, load balancing and data replication that once forced CIOs to pay big bucks.
The continuing merger of storage and networking technologies is also paying unexpected dividends. Over the next few years, Fibre Channel will fade away, and standard Ethernet and Internet Protocol (IP) will become the dominant storage transport and protocol. That should further decrease the cost of storage as commoditylike Ethernet components replace more esoteric--and more expensive--Fibre Channel installations.
|Technology change will redraw the established order in the storage industry and force the reinvention of IT storage strategies.|
There's also a side benefit here: A veritable army of Ethernet-savvy IT workers will be able to contribute their expertise and help improve the reliability of storage operations.
IP will change the concept of storage over distance as well. Storage bits traveling over common IP pipes will open up new possibilities for storage applications such as backup, remote mirroring and content distribution. Goodbye, goofy niches like channel extension; hello, common virtual private networks (VPNs).
This storage transformation will be marked by the emergence of network intelligence. Historically, storage software lived at the end points on either servers or storage subsystems. This model does not scale and is rife with overpriced proprietary solutions that lock users into vendor-controlled gulags for years.
But cheap processors and appliances will move storage intelligence to the network and offer new options for storing, protecting and distributing bits. What's more, these intelligent networks will expose storage applications to standard development environments. Proprietary application programming interfaces (APIs) from vendors will give way to Java 2 Enterprise Edition (J2EE) and .Net-based Web services, opening up high-end storage architectures to the masses.
Clash of titans?
The transformation won't come without a fight. Industry leaders such as EMC, which have the most to lose, are trying their best to freeze the market. These storage giants are dragging their feet on interoperability standards, spreading the gospel of fear, uncertainty and doubt to customers and prospects while introducing proprietary appliances in a last ditch effort to retain control.
In the end, none of these tactics will work. The storage industry would be best served by following a lesson learned from similar past technical transformations. For example, IBM was the mainframe king competing effectively with companies including Unisys, Amdahl, HDS and even Digital for years. But even Big Blue couldn't outdo client/server computing, an industrywide trend. In a similar vein, storage leaders won't be able to slow down the cheap, network-based storage steamroller headed down the highway, so they better figure out how to jump on board before they are flattened.
|CIOs ought to pay close attention to this trend and view all short-term storage purchases this year and in 2004 as tactical.|
CIOs ought to pay close attention to this trend and view all short-term storage purchases this year and in 2004 as tactical. Briefly put, buy the minimal amount of products possible and keep beating your vendors over the head to cough up the best prices. At the same time, prioritize longer-term storage needs and educate yourselves on new products, vendors and options. Plan on beginning your own storage transition by the beginning of 2005.
It's difficult to know which vendors will wind up the winners and losers, but it is safe to say that the storage industry landscape will change drastically by 2006. Over the next three years, this much is certain: New, still little-known storage companies will become industry leaders, while the dinosaurs just fade away.
Jon Oltsik is a senior analyst at the Enterprise Strategy Group.