Everyone knows Intel wants to sell processors but the question is does it really care what operating system is running on the chip? The answer is apparently yes. It's taken notice of the effect Linux is having in the IT market and it's reacting. For example, Linux has continued to grow as governments worldwide invest in Open Source software. These users want an end to proprietary lock-in (think Windows desktop upgrade cycle), to keep IT dollars local (OSS installed by local firms), and to reduce costs. According to a 2004 IDC study Linux is expected to be in use on approximately 17 million PCs by 2008 worth $10 billion in PC sales.

Intel has also taken note of ISVs like MySQL AB and JBoss, which have created new business models based on revenues from services and related tools. Intel is also looking to China were dollars spent by IT users could be allocated to hardware in lieu of proprietary operating systems like Windows.

An initiative announced by Intel in January heralded five new Intel platforms including Mobility, Digital Home, Digital Enterprise, Digital Health Care, and Channel Products. The announcement indicates that the new Intel will stop focusing on silicon and chip speeds and start focusing on platforms on which to build solutions including Linux. At Novell's BrainShare in March Intel marketing manager Matt Semenza offered a presentation called "Intel and the Linux Desktop" saying that Intel is working to ensure that Linux and Open Source software solutions perform optimally on Intel platforms.

Intel isn't exactly new to Open Source software. It's a co-founder of OSDL and Richard Wirt, general manager of Intel's Software and Solutions Group sits on OSDL's board. Intel's Matt Wichmann is chairman of the Free Standards Group's (www.freestandards.org) Linux Standards Base Workgroup (www.linuxbase.org/).

Intel has also let free software developers use Intel products that were formerly only available for a licensing fee, including Intel compilers, VTune analyzers, performance libraries, threading tools, and cluster tools. Intel was also an early investor in now industry-leading Red Hat so it can't be considered a Johnnie-come-lately to Open Source.

Intel has a realistic view of the forces at play in the desktop market. These forces include accelerators, which have both technological and economic components. It sees the evolving government policies favoring keeping IT dollars in country so American companies like Microsoft and Apple seem to be at a disadvantage where Linux can be substituted. Lack of choice in software has led to anti-Microsoft sentiment. Having a chance to choose between Windows and Linux makes more sense and prevents single vendor lock-in. The rampant propagation of worms, viruses, and other security risks are making desktop PC users take notice of less vulnerable or less targeted operating systems like Linux. On the other hand, the outlook for desktop or client PC Linux (an Intel moniker) isn't that rosy. There are significant inhibitors to Linux adoption that Intel hopes to mitigate like providing hardware driver support to making sure that its hardware is supported under Linux. It also acknowledges that there are feature gaps like a Plug N Play architecture that's as seamless as the ones Windows and Mac OS users enjoy on their desktops. Application availability is also a concern. Windows the long-time leader has many more applications available for its operating system. A shift in the market where Linux becomes more widely used will make it profitable for ISVs to start offering solutions on both Windows and Linux. And as the time comes for older applications to be updated, moving them to Web Services as Intuit did with its Turbo Tax for the Web opens the door to new operating systems. Finally, Intel participation in making sure that there's adequate validation and training in best practices will help to relieve potential Linux desktop users of doubts about the Linux desktop.

Intel's Linux Agenda
Intel is very aware of the spike in malicious software attacks, globalization, the increased need for technology mobility, and the driving demand for storage of traditional data like databases as well as digital voicemail and the like as a result of VOIP adoption. Intel is trying to drive the digital office through richer collaboration among PCs, data analysis, and mobility through pervasive connectivity. Intel has even put together a Linux Quick Start Kit for system builders offering drivers for video (Extreme Graphics Driver, Graphics Media Accelerator 900), LAN (Gigabit LAN, Marvell Yukon Gigabit Driver) and Audio (Intel High Definition Audio, AC '97 Audio Driver). These drivers are being supplied in conjunction with Novell's Linux Desktop.

Intel has committed to supporting key vertical market segments for Linux on client PCs including basic office automation, government, transactional enterprise, and education. Additionally, it's working with technology providers like Novell to develop robust Linux solutions.

Vanderpool and Multi-Core Processors
Until recently Intel has been using the codename Vanderpool to describe its new Virtualization Technology. This virtualization technology is expected to be available on the new dual-core Pentium D. It's expected that PC users will be able to run multiple instances of the same OS simultaneously. The next step, though not confirmed, could be to run Windows and Linux side-by-side using it.

AMD has launched its multi-core Opteron chip as well helping to usher in a new era of multi-core architectures. AMD has outlined the possibility of running multiple operating systems side-by-side using its new Pacifica technology, scheduled to be out in 2006. This technology will let different operating systems reside on the same hardware.

It means that Linux could start living on the same PC in harmony with Windows, setting the stage for an eventual migration to a Linux desktop. In theory it sounds plausible, that solution and a way to do a side-by-side comparison. The likelihood of the server market also pushing this trend is very good since speed is less a concern these days and more emphasis is being put on hardware utilization. Consolidating servers in a virtualized environment makes a lot of sense not only from a hardware prospective but the point of view of space savings and power consumption.

Summary
Bottom line Intel sees Linux as a valid platform not only for the server but the desktop. As a corporation that believe in investing in emerging technologies that compliment its products, it has chosen to promote desktop Linux and ensure its success. It obviously wants to be the platform of choice for Linux solutions. It will continue to push new innovative technologies like its Virtualization Technology while being involved in the Open Source software community. It's apparent from speaking to its staff that its approach is to help reduce risk and time-to-market for Linux solutions by lending its support.

SIDEBAR

Intel Offers Centrino Support
As much a personal peeve as a legitimate gripe about Intel was its lack of support for Linux on Centrino. Centrino is a collection of chips aimed at maximizing the utility of mobile computers with low power consumption by virtue of the Pentium M chip as well as wireless capabilities. However, things are changing. With the release of its Intel Pro/Wireless Network Adapter Drivers (http://support.intel.com/support/notebook/sb/CS-006408.htm), Intel has started to offer e-mail support on installation issues and it has started Open Source projects to help with its PRO/Wireless 2100 driver (http://ipw2100.sourceforge.net/) and Pro/Wireless 2200 BG driver (http://ipw2200.sourceforge.net/). Centrino launched in March 2003 and an Open Source initiative started in 2004. Going forward it appears that Linux will supported as a legitimate desktop platform in Intel's eyes and if not there have been successful projects to work around the wireless problem like Linuxant's (www.linuxant.com) DriverLoader for Wireless LAN (www.linuxant.com/driverloader/) devices that load native Windows drivers under Linux to power certain hardware including the Centrino wireless chips.

• Figure 1