View from the Hill:-

Why Sun's Fortunes will Look a Lot Brighter in 2003

I'm starting the New Year with a more positive outlook on the prospects for Sun's SPARC server business. Methinks that Sun's revenue will grow and that Sun is going to look smarter than its main competitors (again).

No. It's not because I've found my white hairs turning back to grey. And I haven't been eating mouldy old Christmas cake (which can cause hallucinations) either. Here are my good solid reasons:-

• Sun's over dependence on ephemeral dotcom companies is now ancient history.
  
  The customers they've got now are the same solid organizations in government, banking, manufacturing and research which they had before the dotcom bubble. Their surviving telco and ISP customers are the ones which made it with solid business plans. You're only as good as your customers. Sun's customers are great companies. They have real problems with increasing server demands, security risks, and reliability etc which SPARC systems can solve quite well at a scalable price.

• Sun's rivals are going to falter in their transition to 64 bit computing.
  
  Sun's had its own fair share of trouble in recent years trying to get more performance out of its 64 bit servers. It seemed like every last drop had been squeezed out. But Sun had the advantage that they've been doing this 64 bit OS stuff since 1996. They don't have to solve every problem at once.
  
  Sun's rivals in the Wintel camp face the double whammy of trying to optimise everything at the same time with a newish 64 bit platform. It takes some time to optimise all those software tweaks and balances, and they're going to stumble and look pretty foolish while they do it. Some Intel server companies are not going to succeed, and will withdraw from the 64 bit race going back to safer proven 32 bit server solutions. That will just increase confidence in the company which has been doing it the longest. My guess is that Sun will introduce strategic 32 bit (external bus) SPARC appliance processors to help out with those parts where high megahertz does best. 2003 will be the long awaited year that SPARC servers leapfrog the competition again. Not all the solutions will come from Sun. Their partners will provide critical building blocks.

• HP's transient problems in absorbing Compaq (which were blamed for their slowdown in business performance in 2002) will be called something different as they take on a veneer of permanence in 2003.
  
  Let's just call it "HP Stodge". Unlike HP Sauce (the #1 ketchup in the UK) which has a rich brown color, and is full of flavor and goodness, HP Stodge is something which is gray, sticky and makes your wheels spin more slowly.You can't run a company that size and which consists of business units which are mostly #2 or #3 in their respective niches and expect to get first class results. Integrating a company is no different to integrating a computer system. You need best of breed components to build a best of breed system. A lot of HP's business units are followers rather than leaders.

• InfiniBand is a disruptive server technology and just what the company doctor prescribed.
  
  Sun's got a good track record of inventing, borrowing and running with new ideas in server I/O connection technology. In the past it did a good job with SCSI, then SBus (which gave Sun a lead of several years over PC makers until PCI came along to do the same job) and Fibre-channel. This year Sun will hype and race ahead with InfiniBand. Sun didn't invent it. But Sun stands to gain the most from promoting high performance servers at a low price, particularly if the low cost part is a 3rd party add-on and doesn't affect the profit margin in Sun's part of the System. Once the hype machine gets into play, you too will believe that an InfiniBand server from Sun will be a safer choice, and that you'll get it working faster than a similar server from everyone else.