Sun Stakes Claim on Financial Services

by Timothy Prickett Morgan

If governmental and educational laboratories are the test beds where crazy new ideas in IT either fly or die, then the financial services market is where the ideas that live get honed and turned into commercial products first. Banks, brokerages, and insurance companies will kill to save a fraction of a penny on a transaction or install a new application that lets them differentiate their businesses in some way.

Those labs and the engineering departments of manufacturers may have put Sun Microsystems on the map when that fledgling company created the Unix workstation market more than two decades ago, and but it was the adoption of Unix servers by financial institutions that made Sun one of the first and the biggest high-tech dot-coms and that allowed Sun to become to dot-com platform of choice during the boom years. In recent years, as IBM and Hewlett-Packard have been very aggressive with their Unix and Linux server technologies, Sun has lost some prestige amount the financial services community. Now that it has a coherent strategy to compete against Unix, Linux, and Windows alternatives with its own Solaris platform, Sun is trying get back to where it once belonged, and it comes as no big surprise that Sun chose New York, the financial services capital of the world, to launch its next set of quarterly product announcements.

Sun's top brass talked up the company's forthcoming Solaris 10 operating system, which will run in 64-bit mode across Sun's and Fujitsu's Sparc platforms as well as machines based on Intel's Pentium and Xeon chips and Advanced Micro Devices' Opteron processors. However, Solaris 10, which was expected in the October timeframe, looks to be due about the end of the year now. Sun's new software chief, John Loiacono, has said that he would get a 64-bit enabled operating system out ahead of Microsoft's extended Windows Server 2003 for the Xeon-64 and Opteron processors, and he got some breathing room when Microsoft pushed out its support for 64-bit mode processing on Xeons and Opterons until sometime in the first half of next year. If Solaris 10 is launched within 90 days or so, which seems to be the plan, Sun will still have a comfortable margin over Microsoft on 64-bit X86 iron.

It will be quite a trick for Sun to get back in the good graces of the financial services industries--at least to the degree that it was adored in the late 1990s. In the early 2000s, as IBM in particular was very aggressive with its Unix offerings, both in terms performance and price, Sun's UltraSparc product line fell woefully behind, and its pricing did not come down as aggressively as it would have needed to in order to keep Big Blue out of its accounts. (Sun's big telecom customer base has been more protected because these firms have 10 year planning horizons for their gear. Our recent IT recession is just a blip in that cycle.)

Sun has to encourage financial services companies that a cross-platform Solaris 10 that has the logical partitioning and virtualization features that Unix customers expect, married to a lower-cost X86 platform (while having commonality with Sparc/Solaris platforms) and having lower costs than commercial Linux (including support) is what they need to base their IT infrastructures on.

This will be a tall order with Solaris 10 still three months off, Sun's own Opteron-based server line still not spanning a broad enough performance range, and UltraSparc server iron that can scale up being three or four times as expensive per unit of work than Opteron iron running the same Solaris operating system. The promises of great features in Solaris and an even more distant promise of cost-effective Opteron and Sparc solutions based on the joint Sun-Fujitsu Advanced Product Line platform or Sun's own "Niagara" and "Rock" processors is not enough of a lever. And that is why Sun's president and chief operating officer, Jonathan Schwartz, is advocating innovative subscription-based pricing models in lieu of new technology. This is the one thing that Sun can do that its competitors probably can't, since many of them do not own their own hardware and software stacks in such a way that Sun does, or they will not embrace the Opteron platform with the fervor that Sun does.

The thing is, Sun has been testing these subscription models for many months, and has been talking about them for even longer. If Sun wants to make a sophisticated financial case for these new pricing approaches on servers and operating systems, then it should just make this pricing available across its entire product line--not dabble with an offering here or there--and let the financial services companies try it out. It is the banks, brokerages, and insurance companies who will be the innovators in IT pricing, just like academia has been for the IT technology for five decades. If anyone can do the math, they can.