The Four Hundred

Looking beyond the current economic uncertainty and global unrest, Gartner analysts are forecasting a general upturn for the technology sector. Gartner also believes IBM is the company best positioned to take advantage of the pending turnaround. Polishing his crystal ball at the Gartner symposium in San Diego last week, Gartner CEO Michael Fleisher explained why the IT sector is on the verge of rejuvenation, which will likely take shape late this year and blossom in 2004.

"It is hard to imagine how business and political affairs could be moving in a worse direction," Fleisher said as he set the stage for his message about how to succeed in bad times. The key point he pressed was that the long-term view always wins over the short term view. While in an evangelical mood, he reminded the audience listening to his keynote address that successful long-range planning requires perspective and distance from the current events, and that this planning may "fly in the face of conventional short-term wisdom."

"Today there is almost universal, deep pessimism. But I think we are getting lost in the moment, succumbing to short-term thinking instead of using experience and knowledge to find the best path," he said.

Reversing a Gartner trend of pessimistic predictions that began in 1999, Fleisher contends there is a high probability of significant improvement coming in 2004. The proof, he said, is that recovery occurs for a simple reason: Demand for IT products and services exceeds supply.

The demand is created by delays in upgrading and replacing hardware and software, which can only be deferred for so long. At a certain point it costs more to fix and maintain than it does to purchase replacements. Basic financial and operational pressures will compel companies to make significant system changes. In taking his long-term-thinking paradigm one step further, Fleisher pointed out that the current market situation allows companies to take advantage of favorable terms. And with the end of overcapacity in sight, pricing power will soon return to the vendors. Another factor that pushes the demand side is the financial benefits from price/performance improvements that have occurred since the last upgrade. "Companies will spend," he said, "if the benefits are clear and dramatic."

Server consolidation, innovative server technology, Web services, and reductions in the time and costs involved in building applications were cited as enhancements that will drive demand. Platform shifts involving Linux and wireless technologies will also play an important role. "Cisco Systems and Intel are literally giving away key technologies, not because they are altruistic, but because doing it will stimulate demand for core product lines and lock customers into their technologies," Fleisher said.

In a different presentation at the Gartner conference, Tom Bittman, a vice president and research director, presented the company's mostly glowing opinion on IBM, while lauding Big Blue for putting forth a "great strategy." Maybe the reason for all of this back-slapping is that IBM and Gartner share a common opinion about the future of IT departments. Gartner peppers its presentations and literature with the term real-time enterprises," while IBM prefers the term "on-demand computing," but both terms are concerned with the continuation and expansion of e-business strategies toward a utility computing model that has variable costs (as opposed to fixed capital costs) and a true distributed computing paradigm.

Managing this growing complexity is the goal of what Gartner calls the "real-time infrastructure," the lofty, theoretical state of an IT shop in which servers, applications, databases, and other IT assets can detect changing conditions and automatically adapt to them according to the shop's service level policy. "I know it sounds a little too good to be true, said Donna Scott, an analyst with Gartner, "but it is on the horizon."

The real-time infrastructure will combine several technologies, some that are nearing maturity, and some that will benefit the real-time infrastructure in the years ahead. Things like load balancing, failover, and blade servers are fairly mature, while the software that links IT resource management to specific policies and services level requirements is still a number of years off, Scott said.

The key to realizing the real-time infrastructure and the goal of policy-based resource management is the virtualization of resources, said Bittman. Some older platforms, like the IBM mainframe and the AS/400 and iSeries midrange platform, have some degree of virtualization built into them. These platforms can run mixed workloads within z/OS and MVS and within OS/400, and they can further dynamically allocate and reallocate resources on distinct platforms through IBM's logical partitioning technologies. However, the true real-time enterprise will be distinguished by an intricate combination of hardware, middleware, and application that all share some level of virtualization.

On Gartner's famed "hype cycle" graph, the real-time infrastructure and associated technologies are at the half-way point on the introductory hype curve. The hype cycle charts the highs and lows that promising new technologies will experience during their life. It starts with exponential growth in hype following a "technology trigger." This hype tops out at the "peak of inflated expectations" and is immediately followed by a rapid drop in hype, coming to rest in the "trough of disillusionment." As the technology matures, the hype level slowly rises again, through the "slope of enlightenment," and--if the technology is any good--it comes home to rest on the "plateau of productivity." The technology trigger for the real-time infrastructure struck about two years ago, Bittman said. "It's really very early in the hype cycle," he said. "We're just creeping up this curve. We think we have a long way to go."

While some systems vendors, like IBM and Sun Microsystems, have been touting grid technologies as the path to "on demand" resource availability and real-time computing, grids will play just a part in the real-time infrastructure, Bittman said. "Grids are being proposed by the vendors as the cure all for virtualization," he said. "No, it's not going to happen." At the same time, Bittman praised IBM for the clearness of its autonomic computing objective, giving it higher marks in "vision" than Hewlett-Packard, Sun, and Microsoft.

Bittman described on-demand as making IT infrastructure technologically more efficient, more flexible, and rapid. The keys IBM holds in this regard are its lead in autonomic computing and the extent of its business consulting, which will deliver the strategic message to the installed base. Autonomic computing is primarily about making software more self-managing. "IBM has the deepest and broadest vision in the industry," Bittman said. "You can't have flexible business processes unless you have flexible IT, and you have to be able to self-manage your infrastructure." This is particularly true as a company's IT department becomes more connected with customers and suppliers.

Software, Bittman said, is another reason why IBM has strengthened its position. "The future is with software," he said. "It is more strategic to maintaining customers over a long period. Once you have a WebSphere customer or a DB2 customer, it will take a lot of change to get that customer to move to something else."

On-demand makes integration a requirement. Primarily, this means software operating in a heterogeneous environment, and WebSphere and Web services come into play. WebSphere has become very critical to IBM. It has become a more focused software group and has matured so that it can be shared among the platforms. Shared components is a major focus at IBM, where divisions used to work within individual fiefdoms. Hardware used to be the key to locking in a customer, but it is becoming less important. IBM is not a hardware company anymore. According to Gartner statistics, hardware revenue accounted for 39 percent of IBM's total revenue in 2001. It will drop to 27 percent by 2006 if Gartner's bet comes in.

Servers will become more of a support player, Bittman said. They will support the open standards, the grid computing standards, and work in an autonomic environment. Unix and Intel servers will be the primary growth platforms at IBM. Those platforms will allow IBM to be aggressive in battles with the competition. Right now IBM and HP are neck-and-neck for the number-two position in the Unix market, and in the high end of the Intel market IBM is doing very well.

Bittman sees the iSeries as an important cog, but not one that will be in the forefront of IBM strategic moves. The iSeries (AS/400) started off as IBM's winner in the small and midsized business market, but Bittman points out that IBM now has its focus on software and particularly on software that runs on other servers, especially Intel-based servers.

The iSeries installed base won't go away anytime soon, he said. "In fact, IBM has made some aggressive moves to grow the installed base." Like most of the iSeries users, Bittman cites inadequate marketing and market perception as barriers to IBM's attempts at bringing more customers to the iSeries. "It's not technology," he said.

Gartner says Big Blue is in a great position to make substantial gains as the IT market picks up momentum. The real question about how successful IBM will be is in its execution.