HP + Compaq: Server powerhouse

Color me puzzled. Wall Street doesn't seem to like the proposed merger of HP and Compaq. What's not to like? In a declining PC market, two major forces are joining to survive a withering market. But what's making IT managers stand at attention is that two major server vendors are coming together.

Compaq was poised to be a major mid-range server player about four years ago. The company's then chairman and CEO Eckhard Pfeiffer had bought out Tandem, with its Himalayan line of massively parallel systems, and then moved on to acquire DEC and its OpenVMS, Digital Unix, and Alpha servers in early 1998. By April 1999, however, Compaq's board was running rampant with internal dissension, and it also panicked at a bad quarter, deciding that the cure to its ills was to fire Pfeiffer. Little did the company dream what bad times would be in store for it.

Though you could argue--and I do--that Pfeiffer's firing was the beginning of the end for Compaq, it certainly was the end of Compaq's mid-range computing initiatives. After a series of reorganizations, any advantage gained from the DEC and Tandem acquisitions was gone.

And now, HP and Carly Fiorina enter the picture. Though Fiorina hasn't been the superstar everyone hoped for, she's managed to keep HP floating higher in the water than Compaq and many other technology companies. One reason is that she's supported the PA-RISC architecture and the HP-9000 mid-range computers it powers--and the HP-UX that makes it all run. So what's in store for the server market? First, Intel will be a winner. Alpha's been on its death bed for what seems like forever, but it will be no more after 2004. PA-RISC has only been hanging on because HP correctly thought that the Itanium would take its own sweet time being released. By this time next year, expect to see HP-9000s and the renamed AlphaServers rolling off the production line with Itaniums under the hood. We won't see AlphaServers (ItaniumServers?) disappear, because there are still enough OpenVMS and Digital Unix customers out there to make it a small, but viable, market.

The NonStop Himalayas will enjoy more attention. For the longest time, Compaq hasn't realized that it had a goose that laid golden eggs with Himalayas. Not everyone needs a computer that won't stop running, but when you need that kind of reliability, you really need it. Sun and IBM have made great strides in the non-stop marketplace because of poor Compaq management, but I don't see that continuing with HP in charge.

On the high-availability front, Compaq products will get a boost from the acquisition. Technically, as the August 1 "Single-System High Availability" report from D.H. Brown Associates shows, Tru64 Unix and AlphaServers do better in clustering situations than IBM and Sun solutions. HP will turn technical superiority into market gold.

As for the lower-end HP Netserver and Compaq Proliant networking servers, the Proliant (always a troublesome line, in my experience), will gradually be phased away in favor of Netserver. I'll be surprised if there's a new 2002 line of Proliants.

On the operating system front, Digital Unix and Open VMS will continue. They're reliable, solid, and their customers love them. I don't see much money being spent on improving them, though. Tru64 is another matter. I expect an Itanium port sometime in 2002. HP-UX has been doing better on both workstations and servers; if any proprietary Unix comes out strong from this union, it will be HP-UX.

Linux, though, will be the real winner. Both companies have strong Linux commitments that will grow even stronger. Red Hat and Caldera, both of which have strong relationships with Compaq, should do the best here. That is, if Compaq, as rumor has it, doesn't buy a Linux company between now and the merger. If I were a betting man, I'd put my money on seeing Linux as a standard offering across the hardware line (except for the Himalayas) within the next year.

Bottom line? HP, with the best of what Compaq has to offer, will finally be in a place to duke it out with IBM's RS/6000, AS/400, and xSeries NUMA lines, and Sun's complete product line. It also means that HP will be ready to fight it out with companies offering Microsoft .NET-based servers. With this acquisition, HP has taken a giant step forward in high-end servers and mid-range computing.

Steven has written about technology for more than 15 years. He was previously a programmer and network administrator for NASA and the Department of Defense. Steven is also currently chairman of the Internet Press Guild.