The Linux challenge to NT in the enterprise
Just a small investment by Intel? Ah, but look at the other announcement...
Intel's spinmeisters are no doubt well-satisfied that reports of its excellent Linux adventure have so far focussed on the Red Hat aspect, while failing to notice the significance of the Linux grenade Intel lobbed in, a couple of hours before Netscape started blabbing to the reporters about Red Hat. The reason they'll be happy is because although Intel wants to broaden its range of friends beyond the Wintel straightjacket, it wants its little deviations to have a certain amount of deniability. This doesn't stop them annoying Microsoft immediately, and probably won't stop a more overt breach ultimately, but so long as Intel can still claim its relationship with Microsoft is still strong, it can carry on playing footsie with the opposition and claiming to be neutral. Hell guys, we just make the CPUs, we don't want no trouble… The Red Hat deal obviously helps Red Hat a lot, but from Intel's perspective can be portrayed as not mattering much. The (undisclosed) investment is relatively small compared to the total amount Intel has invested as venture in smaller companies, and much the same goes for Netscape's investment. Netscape and Intel can talk about the growing importance of Linux in the enterprise, and Red Hat can use the investment to set up an enterprise sales operation. But a challenge to Microsoft? Linux shipments are healthy, but they'd have to increase by a factor of ten before they seriously started to hurt Gates' company. That of course didn't stop Microsoft's ever-watchful and ever paranoid execs noting that Linux was their biggest threat last week. But this time the execs are right, because Intel's plans for Linux will hit Microsoft right where it hurts - NT. Intel didn't say much yesterday about the Red Hat investment, but it did roll out a series of initiatives (Intel puts its weight behind Linux - and Linus) designed to support Linux in general and to boost its uptake. These, which included Intel joining Linux International, constituted the really important announcement from yesterday. They were in effect designed to place Linux in the same kind of advantageous position as Windows is on the Intel platform. The essence of the Wintel alliance is really that it produces a commodity platform. Intel pitches its hardware as standard, cheap to build in volume for a ready mass market, while Microsoft does the same for the software. So NT's assault on the corporate network is based on it running on standard, cheap platforms, easy application integration, and easy deployment (NB, this is the brochure - we don't necessarily believe this stuff). At the same time as pursuing a strategy of close integration of hardware and software with Microsoft, Intel has been trying to do similar things with the Unix market. Intel-based servers can be presented as cost-effective, maybe not as scalable as the big Unix boxes from Sun and IBM, but you can always achieve the power you need just by deploying more of them. Which means Intel sells more boxes. But there are problems. There's no standard Unix that can go onto an Intel box as readily as NT does, and although the UDI initiative will help by moving towards standardisation of drivers, it's obviously not a complete solution. Intel also has problems with the vendors, because although there are versions of Unix for Intel from enterprise server suppliers, these companies have their own Risc boxes aimed at the enterprise, so for example you'd hardly see Sun cannibalising its revenues by pushing Intel Solaris. Paradoxically, this means Unix is missing a trick. On the Intel platform so far the only real challenger for large-scale enterprise networking applications has been NT, so the corporations tend to look at their current systems, quite probably Unix and/or mainframe-based, then look at NT and then decide not to do anything right now. NT sales have gone up, sure, but enterprise customers prepared to bet the business on it are still in short supply. You can see how this frustrates Intel, even without taking account of the fact that NT deployment is now stalled for about another year until NT 5.0 is out. So casting its eyes around, Intel identified Linux as a logical candidate (possibly the only logical one) for giving a helping hand into the enterprise. Corporate perceptions so far have limited the uptake of Linux in business, but endorsement from Intel and Netscape will help, and note also how Intel pitched its announcement. It was talking about Linux being a good platform for ISPs, i.e., it was aligning it against big, enterprise-class servers from Sun and IBM. The route Intel proposed for Linux (in the really important announcement, that is) could perhaps be seen as a sort of distributed Wintel. Intel likes to work closely with companies in order to make the hardware and software combination closely integrated, fast and commodity. Linux itself isn't a company, but Intel reckons it can do a similar job by plugging itself into, and offering a high level of R&D and technical support to, that thing out there which is the Linux community, and which appears to be able to establish and adhere to standards without anybody owning anything. If it all works according to plan, it'll be possible to buy systems that are to all intents and purposes shrink-wrapped combinations of Intel-based hardware and Linux software, and as Linux will finally have established a legitimacy in the business end of the market, actually going out and buying them will become more and more of a no-brainer. The next move, trust us, will be for a major PC vendor to start pushing Linux actively. Watch Dell and Gateway. ® Other current Linux stories include: Informix becomes latest Linux recruit Intel, Netscape buy stakes in Red Hat Linux Intel puts its weight behind Linux - and Linus Click for more stories
Sponsored: Becoming a Pragmatic Security Leader