A couple of innocuous words that have insidious consequences. A simple marketing tagline, yet it appears to be seriously undermining a vast segment of the computer sector. What is it? The phrase, “industry standard.”
It comes from the same warped (but ingenious) minds that coined the term “legacy” back in the 1990s. That harmless word served to forward the Windows NT – Intel processor gravy train at the expense of so-called inferior technologies like mainframe, OpenVMS and even Unix. Yet many would argue that despite all the genuine improvement to Windows in recent years to make it much more enterprise friendly, it still can’t hold a candle to many of these legacy systems.
The rumor goes that the legacy propaganda was originated between a Gartner analyst and someone in the Windows NT development/marketing camp. Whether true or not, it certainly helped sell an awful lot of Wintel gear. Now we have “industry-standard” weaving its way into the computer lexicon. The hidden intention appears to be to outlaw RISC and everything else except x86-Intel fare.
Go to Sun’s Web site and you will find numerous legacy references. In fact, Sun has an “industry-standard server” line. Even IBM’s site is full of it. And of course, HP bought into it chip, line and sinker. The reason I mentioned those vendors specifically is that they were or are the ones with the most to lose.
All three were RISC-chip champions not so long ago. HP had PA-RISC and Alpha, Sun had SPARC, and IBM had POWER. HP threw in the towel and switched over to Intel Itanium. Are Sun and IBM secretly contemplating a similar move over the long term? At first glance, you wouldn’t think so, as both continue to put plenty of R&D dollars into UltraSPARC and POWER. But you have to wonder why they have embraced the industry standard bandwagon so thoroughly.
After all, what’s so “standard” about it? Industry-standard processors are simply those that are sold on the semiconductor market to original equipment manufacturers (OEMs). They, in turn, incorporate these processors into their systems (as opposed to proprietary processors), which they sell as part of an overall computer solution. So what appears to be going on is standardization of all servers onto Intel processors (with a few consolation sales granted to AMD). That’s actually a monopoly, not a standard. Hopefully IBM and Sun will have the guts to stick it out and continue to fly the flag for RISC-based processors. That will keep things healthier on the innovation chip over the long term.
CSI – Chip Offenders Unit
I attempted to track down the origin of the term “industry standard,” but I didn’t have much success. Web searches are sabotaged by the fact that the term industry standard architecture (ISA) was coined in 1981 for something else entirely. So it isn’t even an original concept.
I first became aware of it about 18 months back when Sun insisted on using this for its x86-based servers. My thought at the time was that Sun was shooting itself in the foot. Why term x86 models as industry standard? It’s only a short hop from there to wonder if the rest of Sun’s servers are non-standard and inferior merchandise. If I had so much intellectual property invested in RISC, I’d never let anyone use the words industry standard anywhere near my company. Unless Sun has a hidden agenda? Who knows …
Further surfing revealed a surprising fact. While I became aware of the wording only in recent months, it’s been around for much of the decade. Nathan Brookwood, a processor analyst at Insight64 used the term in 2002. HP appears have first used it in 2003 to herald the dismantling of its RISC mainstays, as it moved toward Itanium. Even IBM has been throwing this propaganda around since 2005 at least.
So where did the “industry standard” wording originate? Which marketing genius, or propagandist-extraordinaire coined it? Perhaps it happened like the fabled Wall Street Crash meeting of financiers — a secret meeting on a train before they pulled the plug on the world economy. Or some secret cartel of server OEMs that assembles in an underground bunker to fix server prices and determine which companies survive and which will be dispensed with? If you know where it came from, let me know.
Drew Robb is a frequent contributor to ServerWatch, where this column originally appeared.