Intel still stuck on the FireWire fence
It's in the 1394 patent pool so why doesn't it just back the technology?
Analysis What is Intel's problem with FireWire? Last week at the Intel Developer Forum (IDF), Chipzilla's bi-annual shindig, held in Palm Springs, it once more threw its mighty weight behind Universal Serial Bus (USB) 2.0, just as it did at the same gig six months ago. And this week, Jason Ziller, Intel's platform marketing manager, told the Infoworld Electric Web site that the prince of processors has no plans to support FireWire -- or IEEE 1394, as it's known in the trade -- in its core motherboard chips. Now, if Intel had gotten off the fence and categorically announced it has a real downer on FireWire, that would be fine. And you might well think Ziller of 'Zilla's comments are pretty definitive on the matter. The trouble is, they aren't. The Infoworld story cites an Intel Web page promoting USB 2.0. The page goes on about how "1394 deployment into the PC platform has proceeded more slowly than expected". Why? Because, in part, of "uncertainties about cost and licensing". Pretty damning, you might think, and a further defining statement of Intel's dislike of FireWire. Wrong. The same page also says (and Infoworld curiously neglects to mention) "Intel has been strongly committed for several years to the 1394 high-speed serial bus". As for the "cost and licensing" point, the chip giant is being damn disingenuous here. Back in January, Apple came under fire for its '1$ per port' fee for the use of certain FireWire intellectual properties it owns. Apple, of course, invented FireWire, and while it submitted the technology to the IEEE, the international organisation that controls such standards, it retained (as IEEE regulations permit it to do) certain portions of the spec. Hence the fee. On hearing this, FireWire's chief supporters, most notably Compaq, Sony, Toshiba and Phillips, called foul and eventually forced Apple into a compromise: it shared out its FireWire IP among its fellow supporters, who would together manage its licensing and set a (lower) fee for each licence. At this point, Intel could still grumble about the cost and licensing issue, and company VP Pat Gelsinger did just that at the last IDF six months ago when he introduced USB 2.0 to the world. So why, then, did Intel join the FireWire patent pool just three months after Gelsinger's IDF keynote? You see what I mean? On one hand we have Intel forging ahead with USB 2.0 on the back of the technology's promise to deliver FireWire-level or greater performance, even to the extent of leaving the technology off its motherboards -- and on the other we have statements of support backed up with a direct investment in the technology's governing body. At this point, Mac users might be wondering what the heck Intel's inability to get off the fence on FireWire has to do with the price of potatoes. After all, Apple's supporting it, many of Mac peripheral makers are supporting it and it's a real big deal in the consumer electronics world. The issue is that Intel pretty much has complete control over what features 90 per cent of PCs offer. That, in turn, determines what kind of connection schemes peripherals vendors offer. In the old days, when Apple stood aside from the hordes of IBM PC-compatible computers, that didn't matter too much. Now, when it needs to fit in, what Intel does here takes on a new importance. Intel's 'in one minute, out the next' attitude to FireWire is essentially because it's not sure it can win over the consumer electronics guys to USB 2.0. There are a heck of a lot of digital video systems out there that use FireWire (although few of them use that name for it) and Intel is unsure whether they'll be willing to drop it for an incompatible technology, ie. USB 2.0, particularly since USB 2.0 won't make it out of the labs before the middle of next year. (Yes, the PlayStation 2 is supposed to support FireWire, but it's a long way off, and its specs. can and will change before the console is finally released.) In short, it doesn't want to get left out in case Sony et al, decide to stay with 1394, thank you very much. In the meantime, though, it's doing its damnedest to persuade them to hop over to USB. USB 2.0 will be fast enough for digital video connectivity, and no one will have to pay any royalty to use it. Sure, they'll have to buy USB controller chips off Intel, but since they have to buy FireWire controller chips anyway, that's no barrier to entry. Intel's persuasive techniques involved spreading Fear, Uncertainty and Doubt over the future of FireWire, and if it succeeds here, that's going to put the Mac at a disadvantage. Apple is, to a degree, selling its machines on the back of FireWire because it's a clear point of product differentiation. If FireWire disappears or becomes perceived as an Apple-only technology, that differentiator vanishes, and Apple's sales pitch becomes harder. Of course, while USB 2.0 is in development and then while it's gaining ground in the PC marketplace, FireWire will be evolving too, and yes, it's going to get faster -- initially, doubling its throughput to 800Mbps then rising to 1.6Gbps, but so long as USB can get to FireWire's current baseline, as version 2.0 promises to, it's going to limit the need for the faster technology. That's not to say FireWire will become irrelevant overnight, but that it will increasingly find itself relegated to ever more specialist roles. All this is, at least, some way off. In practical terms, USB 2.0 is largely vapourware, as is FireWire 2.0. And in the meantime we have the prospect that other motherboard vendors will offer 1394 support even if (or more likely because) Intel isn't. However, it does highlight the very important for those companies that so support FireWire to make more of an effort to promote the technology. Harmonising on one name, whether it's FireWire, iLink or whatever, would held, by making it clear that you can connect a Sony DV camcorder to your PC. If, as is rumoured, the next-generation iMac has one or more FireWire ports, that will help, but Apple also needs to get its fellow patent poolers to do more. It's just a pity that, as a member of that pool, Intel can't just get off its fence and support the technology fully too. ® Full IDF Summer 99 Coverage Back to the start of the article
Sponsored: Hyper-scale data management