There are two common points of view about Linux in the so-called enterprise. The first is that Linux is only capable of displacing Microsoft products on cheap low-end servers, and that proprietary UNIXes with their huge 64-bit address spaces and big SMP scalability are safe. The second is that Linux is mainly displacing UNIX, since it's easy to port software from UNIX to Linux, and that Microsoft with its difficult-to-port-from APIs is safe.
Both points of view are wrong. Nothing is safe. On page 44, Linux is running on a new 64-processor NUMA system. And, on page 52, Linux is displacing the nastiest Microsoft server to replace, Exchange—undocumented protocols and all.
In our December 2002 issue, Douglas B. Maxwell wrote about how he beat the graphics performance of a large SGI system with $15,000 worth of PCs. This month, we're celebrating SGI's release of a big Linux box by putting it on our cover. Do we have the world's shortest attention span? Aren't generic PCs taking over everything?
If they are, they're not done yet. If you have a big problem that you haven't figured out how to split into PC-sized chunks, or don't want to take the time to split into PC-sized chunks, the 512GB of memory on the new SGI Altix 3000 seems like just what you need.
One slogan at SGI is “do science, not computer science”. Do big problems the way you know how and get better results now. Of course, this is waving a red flag in front of the commodity cluster faction, and I'm sure we'll soon have plenty of articles pointing out how you can get previously unclusterable work done on a cluster.
The diversity of success stories in this issue makes it clear that any company that tries to compete with Linux in a fair fight will lose. So it's going to be an unfair fight for a while, with the non-Linux vendors pulling shenanigans such as bogus software patents, FUD-based marketing, copy-restricted content, carefully placed “donations” and “campaign contributions”, and who knows what else.
But most of the companies, and most important, the people, who are promoting non-Linux legacy products today are going to be part of the Linux business tomorrow. Since our community will survive and theirs won't, ours has to be able to welcome and do business with them in the future. So we can't engage in the same desperate nonsense they are. All we have is the best software and the truth, and that's plenty.
Finally, the one enterprise that's most important to pulling some people from just getting by up to knowledge and success is the public library. Your local library provides educational materials, entertainment, training and community programs. Now, that important institution's budget won't be wasted on expensive, inflexible proprietary software. The Koha Project is offering library cataloging and search software under the GPL, in the same spirit we have public libraries in the first place. Join your local Friends of the Library and read Pat Eyler's article on page 58.