Over the past several weeks, three seemingly unrelated announcements suggest that a major shift might be underway in the storage arena. First, at its European roadshow in Milan, Italy, market-research company IDC announced that it expected the worldwide storage market to reach the $71 billion mark by 2004. IDC analysts anticipate that software will be the fastest growing segment of the market, generating around $9 billion in sales, and services will account for $24 billion in revenues. If you do the math, you find that software and services together will make up almost 46 percent of the total worldwide storage market.
Second, IBM announced a deal to combine its Hard Disk Drive (HDD) manufacturing unit with Hitachi's parallel unit in a new company. Hitachi, which will own 70 percent of the new company, paid IBM $2.05 billion for its HDD assets. The deal essentially takes IBM out of the disk drive manufacturing business.
Third, Network Appliance (NetApp) unveiled an agreement with SAS Institute to develop storage products specifically geared to business-intelligence applications. According to NetApp officials, business-intelligence applications require a storage infrastructure that integrates high-performance online storage for data that must also be readily accessible throughout the enterprise with lower-cost, near online storage for archived data that's used less frequently. Under the agreement, NetApp will supply the hardware and management software, and SAS Institute will create a processing layer. According to company officials, NetApp hopes to sign similar agreements with Hyperion and other vendors of business-intelligence software packages.
So what do these three separate announcements demonstrate? The storage market is undergoing the same kind of metamorphosis that the general computing market underwent in the late 1980s on the PC level and is undergoing now, at least potentially, on the server level. That is, distinct hardware and software markets are evolving. Before the advent of personal computing, computer hardware manufacturers also supplied the OS. Microsoft changed all that. At the server level, UNIX was supposed to be the unified, hardware independent. (Although that scenario didn't work out, Linux might still play that role.)
Although IBM is pulling out of the hard drive business, the company is very much staying in the storage arena, particularly as a software and services supplier. If as IDC's analysis indicates, storage hardware, software, and services will become distinct arenas, two quite different trends should emerge. On the one hand, the pressure for open standards will escalate. Administrators won't want hardware choices to dictate software choices and vice versa. On the other hand, the market should open up to more innovative, entrepreneurial activity as companies perceive new storage-related opportunities, and both hardware and software storage products will probably proliferate.
Where does NetApp's announcement fit into this picture? In the general computing arena, vendors and consultants commonly advised companies to base hardware selections on the applications that they wanted to run. However, that hasn't been the case with storage, which companies generally view as a piece of their overall infrastructure. But in time, applications might become the most crucial choice in the storage industry as well.
The NetApp-SAS Institute deal suggests that in the future, storage choices will be increasingly tied to applications. NetApp has already inked a series of deals with the major database manufacturers, and new storage protocols such as Direct Access File System (DAFS) could have a major impact on database performance.
The storage market will probably follow the same route as the general computer market, with consolidation among hardware vendors, an emerging common software infrastructure, and competition at the application level. Because the changing approach to storage is in its early stages, I'll get to monitor the accuracy of my prediction as the storage market evolves.