By contrast, Sun's managers must be reaching for the Maalox. They need to figure out how to survive the on-going consolidation of the computer industry accelerated by the steep recession they are caught up in.

The opportunity must have been tempting for Big Blue to buy up its third largest competitor for as little as $7 billion—a little more than half the cash IBM has on hand. But it would have been hard even for the leviathan IBM to swallow Sun.

The two companies have significant overlap in their microprocessor, server and storage products. Customers of the two companies would want all the products to continue indefinitely. IBM's shareholders would demand the company rationalize its overlapping businesses for maximum cost savings—likely killing Sun's Sparc processors, eliminating many server blade and system designs and several disk and tape systems.

Anyway you walk that line some people are unhappy and the company spends time and money on transitions rather than innovations that move the market forward.

"It was certainly going to be messy," said Nathan Brookwood, principal of Insight64 (Saratoga, Calif.). "It takes the better part of a decade to wean an installed base off a hardware architecture," he added.

Martin Reynolds, a senior analyst with Gartner Dataquest, speculated if the deal went through IBM might have ported Sun's Solaris operating system to the IBM Power processor to move Sun users to IBM systems. Solaris is state-of-the-art in its multicore support, but the porting effort would not have been easy, Reynolds said.

So I expect IBM execs are breathing a sigh of relief today. In the end, it's better for IBM to let Sun twist in the wind than try to rationalize the two big server makers. What IBM really needs is an injection of communications technology—not more servers—if it wants to best compete with current and emerging rivals Hewlett-Packard and Cisco Systems.