Mad Dog 21/21: Big Blue’s Sun Strategy Gamble–IBM Without i
September 21, 2009 Hesh Wiener
Users and resellers of Sun Microsystems iron have been shaken up by Oracle‘s acquisition plans. IBM is trying to exploit the situation, picking off as many Sun accounts as possible. Big Blue says its Power-AIX systems have won over many customers and adds that its mainframes and System x products are making big inroads, too. What is missing from this picture is the i-on-Power line. This could prove problematic. IBM’s dropping the i may mean dropping business machines, too, opening market gaps that Oracle and Microsoft can exploit, even in accounts where IBM is nominally the prime vendor.
IBM seems to believe that its EBCDIC technologies–the Power servers running i 6.1 line and mainframes running z/OS–don’t have a lot of appeal in the Sun base. IBM’s posture suggests that some (and maybe a considerable slice) of the Sun sites it wants to capture may have formerly been traditional IBM customers; it refers to its marketing targets as win-back users. So even as IBM boasts of its success in moving Sun customers with many servers to few large z platforms (and sometimes just one), the z systems IBM is talking about have Linux engines, not the general-purpose engines that are needed to run z/OS. IBM sells the Integrated Facility for Linux (IFL) engines for a lot less than it gets for general purpose engines. Like the System z with z/OS, Power Systems with i is perceived as also carrying a premium price. IBM maintains that both its EBCDIC computing lines provide superior value, but, in the ‘nix world at least, IBM has a hard time getting traction with that story.
The IBM emphasis on cost savings is a big part of IBM’s sales pitch when it offers a Unix or mainframe Linux system, once again sending a signal that IBM’s Power and z IFL servers are costly products except when IBM offers special deals. Sun customers moving to a Power or z Linux server can trade in their old Sun boxes for $8,000 per core, although they don’t get cash for their Sun servers. IBM gives these customers credit that can be applied to software, services and other IBM offerings under a set of rules a prospective buyer has to carefully review with and IBM sales rep or authorized reseller. The credits are real, but they are not available to every prospect and they may only pay off in a big way if a customer’s acquisition plan closely matches IBM’s incentive offering.
IBM doesn’t offer the same incentive to customers who move to an X64 platform. The IBM System x line is less expensive than its Power or z lines and, for savvy negotiators at least, comparable in price to similar servers offered by Hewlett-Packard or Dell . . . or, for that matter, Sun, which has its own line of Xeon and Opteron servers.
IBM’s single-minded emphasis on hardware costs may turn out to be a risk strategy right now, compared to a broader marketing effort that might emphasize hardware value for customers whose plans are clear and whose software stacks are based on Unix or Linux environments. But in today’s world of economic turmoil, IBM’s position may be weak compared to that of hardware and software competitors when it comes to winning the startups that will become tomorrow’s major customers.
Not all customers, not even all Sun customers, are in fields where the Internet is the be-all and end-all. Quite a few are in segments where computers do the bookkeeping, support various business activities, automate some aspects of sales or support, and provide a collection of clerical and communications services. IBM’s emphasis on what it believes is the best way to chip off chunks of the Sun base could be distracting it from a strategic necessity that HP, Dell, Oracle, and Microsoft seem to grasp more effectively. What IBM is missing, at least in its pursuit of the Sun base, is the very heart of a corporate philosophy that enable Big Blue to create and sustain the mainframe, to invent smaller data processing cousins going back to the days of its General Systems Division, and to develop the System/3X and AS/400 minis and what is current the i-on-Power line, an information processing system that addresses the huge population of companies for whom Web 2.0 is largely irrelevant or just another headache for in-house or outside Web geeks.
It’s very possible that Sun’s customers and even the IT user base in general have grown so technologically savvy that integrated packages that include servers, database software, applications, networking software, and terminals support have little appeal. But somehow we think Microsoft’s Windows, Exchange, and SQL Server business (with its ties to Outlook and productivity applications) is hugely popular because it is in fact what customers want. Oracle sees the same kind of future, but from a different angle. Oracle’s database software is a high-end offering, one that most of its customers, including those running on IBM hardware, think is the richest middleware available. Oracle has worked very hard to treat Unix and Linux environments on Sun, HP, and Dell boxes as members of the family. Oracle has tried, with slightly less satisfactory results (but only slightly), to get high-end Windows customers to buy its software. Even without the acquisition of Sun, which will bring MySQL to Oracle (unless antitrust authorities force a divestiture), Oracle has put quite a bit of thought into ways it can reach down into small businesses and start-ups without wrecking its established high-end revenue stream.
Basically, an X64 server running Windows (or Linux) and SQL Server (or MySQL) is a close but not identical rival of a small i-on-Power system. But IBM and its resellers can’t seem to make the winning case for the i that was relatively easy in the past. IBM failed to make its friendliest technologies–the software behind green screen applications (and green screen applications that have grown GUIs)–compellingly affordable. IBM’s failure in the EBCDIC midrange mirrors its failure at the high end. For even though z mainframes running z/OS and DB2 and CICS remain the source of a significant revenue and more impressive profit for IBM, Big Blue is really having a hard time persuading customers that it can deliver superior value.
As The Four Hundred reported last week, IDC says that the market for what it calls enterprise servers (priced at $500,000 or more), which more or less means IBM mainframes and big Unix boxes, is getting crushed right now. The mainframe is not going to bounce back. Instead, based on trends that IDC believes will continue, the high-end the server market is going to be cut in half during the next five years.
By contrast, IDC believes the midrange server segment (where machines cost $25,000 to $499,999), where i-on-Power is a contender, will suffer in the near term but bounce back by 2013. At the low end (where i-on-Power enters) conditions are similar. Worldwide trade in servers smaller than mainframes is going to survive even though computing budgets are going to remain under pressure for a while.
So, while IBM edged out HP to lead the industry in the dollar value of servers sold during the first and second quarters, according to Gartner and IDC, HP and Dell handily beat IBM in unit shipments. Undoubtedly, many of the servers in Gartner’s count are the sole boxes in small offices or one or only a few servers in companies with small payrolls, and IBM could very well feel it has little interest in such customers. But even in these troubled times the computer industry is likely to ship 6 to 8 million servers worldwide in 2009, and that means there’s plenty of room for the relatively small volumes of IBM EBCDIC systems it would take to give Big Blue a memorable year in market segments it has failed to win over for quite a while.
IBM has failed to see the value of a back-to-basics marketing plan aimed at companies forced to retrench by miserable economic conditions, a scarcity of new opportunities, and the daunting cost and complexity of the kinds of Web technologies that the leading edge developers claim is the IT of the future. (The Web visionaries may be right in a world that’s smart enough to live out their dreams, but it’s not so easy to show that we have such a smart world, or even a world that’s very smart at anything anymore.)
But simplification, backing up a few paces and heading in a different direction, has been a major theme in the IT world for the past few years. Intel has spent the past few years making a transition from hot chips to cool ones, from fast chips to slower ones, from uniprocessors to quads and beyond. Microsoft has had to ditch or at least postpone some of the impressive but mainly decorative additions to Windows it tried to promote in Vista. Windows 7 may not reverse all the bloat that made Vista an unsatisfactory choice for many customers, but it will certainly be as much of an admission of the need to rethink its strategy that Microsoft has made in software since it figured out the Internet was important. And Apple, goaded by Google, may yet stop fighting prospective customers who would love its software but not on its costly hardware.
IBM sometimes acts as if there’s a huge conflict between the idea of offering Unix boxes and also providing a more friendly line of systems for businesses that prefer not supporting ornate and arcane but technically advanced solutions. Those users, who might buy a small i-on-Power if IBM made that choice a bit more attractive, may soon reach the limits of their tolerance of Unix and Linux systems that are, for them, on the exotic side, and Windows systems that have their own drawbacks, limitations that could open as many opportunities for an intrepid IBM as Sun’s dissolution.