A series of recently issued market research studies and company earnings reports reveals the crosscurrents at work in the enterprise storage market. On one hand, the high-end storage arena is in the middle of a hardware product-transition cycle, a transition that's had an impact not only on the hardware vendors but also on the Fibre Channel SAN arena. On the other hand, EMC is using the transition to shift the rules of the game. Instead of providing storage management, EMC--which has completed three major software acquisitions--is positioning itself as a data-management company in which software and services are as important as hardware. How these crosscurrents work themselves out could have a significant influence on the nature of storage in the future.
Perhaps the most intriguing evidence of the enterprise storage market's new direction is a recent report from the Yankee Group. The report shows that the Fibre Channel component market declined in second quarter 2004 compared to first quarter 2004. The market researchers' second-quarter revenue estimate of $415.49 million is also below the average quarterly revenues of $475.8 million that Fibre Channel SAN racked up in 2003. According to the study, the Fibre Channel SAN component market showed weakness across the board in the second quarter. Each major segment--switches, directors, and host bus adapters (HBAs)--showed sequential declines from the first quarter 2004 to second quarter 2004. Only the director segment, which generated $132.8 million in second quarter 2004, showed growth compared to the average quarterly sales in 2003. That year, the director sector recorded sales of $117.7 million per quarter on average. As a result of second quarter 2004 performance, the Yankee Group now believes that only the director segment will hit earlier sales forecasts, and the sector as a whole could fall as much as 20 percent below earlier forecasts.
Yankee Group analysts offer several reasons for the dismal performance. Over the past year, Fibre Channel SAN vendors have targeted smaller enterprises, a move that has led to earnings declines. Moreover, the use of IP SAN has grown, so vendors have broadened their product portfolios. But perhaps most tellingly, large enterprises have held off investing in their storage infrastructure, thanks to the product-transition cycle that's underway at the high end.
The impact of the product transition is obvious in the third-quarter earnings reports from Hitachi, HP, and IBM. At IBM, despite a 27 percent increase in the midrange disk category, overall disk storage revenue dropped 1 percent year over year. According to Merrill Lynch analysts, that drop indicates that sales of IBM's high-end Shark fell dramatically in the third quarter. Indeed, according to Merrill Lynch, IBM's mainframe storage business dropped 26 percent in the third quarter. At HP, storage revenues dropped 15 percent year over year, with online storage falling a whopping 23 percent. Hitachi also turned in a weak performance in enterprise storage, according to a channel check conducted by Merrill Lynch researchers. In each case, buyers were waiting for new technology scheduled to appear on the market beginning in December.
However, the story is dramatically different at EMC, where revenues jumped 34 percent and net income jumped 37 percent compared to the same quarter in 2003. Core revenue--excluding the revenue generated by Documentum, Legato, and VMware--was up 19 percent.
Clearly, EMC gained market share at the expense of Hitachi and IBM, but those gains might be short-lived. The next generation of high-end, high-performance enterprise storage products from both Hitachi and IBM offer impressive technical specifications, and many industry observers anticipate that the new technology will help those companies regain lost ground. EMC, for its part, has also indicated that it's readying a new generation of technology that will catapult it back into the performance lead.
But, although EMC seems ready to fight the good fight of hardware specifications, something more significant might be underway. EMC is focused more on Information Lifecycle Management (ILM) and the software applications necessary to manage data storage over time rather than on the underlying hardware. Indeed, storage might now follow the same model as computing, in which the underlying hardware has increasingly become a commodity item, and most of the significant differentiation comes much further up the technology stack.
Over time, storage could emerge as a completely separate IT infrastructure, requiring a separate management strategy and separate tools. Enterprises would have production systems on which to generate data and an independent infrastructure in which to manage the data after it's produced. Should such a scenario come to pass, many IT organizations would be facing a major restructuring effort.
