Enterprise Unix Roundup: Eyeing the Horizon

Main

Tips of the Trade

In 2005, Linux and open source software continued their assault on Unix; Solaris went open source; SCO edged closer to the precipice; and the GPL revamp kicked off. What will 2006 bring?

Where, oh where, did the past year go? It seems like just yesterday, we were waxing poetic on 2004, and here we are again at the tail of 2005. It was a year that had us on the edge of our seats in enterprise Unix-land. Linux and open source software continued their assault on Unix; Solaris and nearly everything Sun sells went open source; SCO edged closer to the precipice, despite releasing a new version of OpenServer; and the GPL revamp kicked off.

It was a good year to be a pundit, and it was an even better year to be an open source or Linux vendor, as the latest wave of zeitgeist rolled in. Here's a brief overview of what developments of 2005 will have the most impact on enterprise Unix in 2006.

Sun on the Rise

Sun had an extraordinarily busy year, and it rolls into 2006 on an upswing.

But first, we'd like to take credit for being right about Sun for the second year in a row. When 2003 drew to a close we predicted the following:

"... if Sun pulls together and recovers from its slump, 2004 will go down as the year proprietary Unix held the line against Linux."

In 2004, it held the line. Fast. In 2005, it appears to have listened to customers, digested, and taken action. Solaris 10 has been open sourced and is 3.1-million-plus strong in downloads. Much of Sun's software crown jewels are now open source as well, as is the blueprint for its latest UltraSPARC chip. Community appears to be more than just a buzzword for Sun: It is a commitment.

Further sweetening its position are Sun's two newest boxes, the T1000 and T2000. Based on Sun's much hyped UltraSPARC T1 CoolThreads chip (nee, Niagara), the servers are priced more similar to Dell systems than the IBM boxes Sun has competed against in the past.

Sun may well see a new day in 2006. We've been saying all along that we wouldn't count the systems vendor out yet, and we're glad to see our hunch proving true. We wish it would lose some of its inherent smugness, but we won't be greedy.

We suspect that the new Sun understands the current market far better than the old Sun. It may have moved into commodity player territory (after all, McNealy did say he told the sales force to make up for the price cut in volume), but in doing so its pitting itself directly against both the Linux vendors and the hardware and systems vendors and may now be a bigger threat to the Linux vendors than Microsoft.

Not that we even need to say it, but we'll be following Sun with eagle eyes in '06.

GPL 3.0 Conquers FUD

This isn't much of a prediction, but we can safely assert that the GPL 3 license that will undergo final revisions in 2006 on its way to final approval in early 2007 will no doubt undergo the most rigorous amount of FUD-finding any license has ever seen.

That's because the GPL (or GNU General Public License, for those of you in the know) is the very license that the Linux kernel falls under. So any revisions to it is going to draw intense scrutiny from friends and foes of Linux. GPL 3 will first be seen as a first discussion draft when it is released at the International Public Conference Jan. 16 and 17, 2006 at the Massachusetts Institute of Technology. You can bet that detractors of the license (and let us assure you, there are plenty), will take whatever potshots they can at the new revision, which is supposed to have more patent law and international language when it's done.

More than the actual technology of Linux, it has been the GPL, originally written by free software guru Richard Stallman, which has most concerned proprietary companies. Because of the GPL, no one company can ever acquire all of Linux, which pretty much kills the old embrace-and-extend philosophy from Microsoft. Little wonder the Redmond company spends quite a lot of time calling Linux names such as "viral" and "cancer." It isn't the nifty Linux code that's earning the foul epithets — it's the GPL.

The revision process is going to be long and involved and, ideally, pretty open. We were tickled to hear Eben Moglen, general counsel for the Free Software Foundation, tell a session at last year's Open Source Business Conference that ultimately it will be Stallman who has the final say on any revision of the GPL because he wrote it in the first place.

So much of the FUD that will get tossed around should wither in the light of day. Still, at the end of the day, one thing should be noted. Whatever changes affected to the GPL will not necessarily impact the Linux kernel. This is for a simple reason many people have overlooked: The Linux kernel is licensed under GPL v. 2 — and is likely to remain under that license. Unless Linus Torvalds decides to re-license the kernel under the GPL v. 3, all the FUD aimed at new GPL via Linux will be invalid as all get-out.

Like FUD usually is.

The Year of the Linux Desktop?

Here's one you might not have thought of: What will the commercial distros end up doing to finally get Linux more widely deployed on the corporate desktop? Innovate a great new killer app? Merge the big two desktop environments into one unified environment? Keep dreaming. The answer will very likely be ...

... buy their way in.

For several years, the big commercial Linux distribution companies have followed in the footsteps of their older Unix cousins and aimed themselves at the enterprise server market. This, by many of the companies' own admission, is a simple matter of going after the low-hanging fruit. Sure, there are fewer big companies out there, the distros (such as Red Hat, Novell, and Mandriva) argue, but if you can score just a few of them, the sales numbers will be big.

While the Linux distros (and their hardware partners Sun and IBM) have been focusing on the server space, many IT managers have been wondering when the desktop functionality of Linux will be ready for them to use, too. After all, you can lock down all of your servers under tight Linux controls, but it does little good if half of your Windows clients are zombified because your employees decided to click on the "Please read this file quick!" link in their in-boxes.

But, for good or ill, there has been little actual progress made on getting the Linux desktop ready to go. We think it's ready now, but for whatever reason, the customer base doesn't. The big reason? A perceived lack of applications. Of the big vendors, we think Novell has made the best progress on the desktop, followed closely by Mandriva. The solution to close the perceived gap — acquire the technology.

You see, out there in Linux-land, away from the enterprise-oriented vendors, there's a whole cadre of desktop-oriented distributions working their coders' butts off to get "ready" for desktop users. We're referring to Linspire, Xandros, and Ubuntu, to name a few. Of the distros in this space, Linspire and Xandros have the strongest desktop offerings. Xandros has been partnered with Skype, the open source Voice over IP application, for more than a year now. Linspire has been engineering a lot of multimedia applications that normally Linux users have to hack their hearts out to run. Both have Windows interoperability with emulation software.

We think it is a distinct possibility that one of the big distros will acquire one of the smaller ones this year for the express purpose of enhancing its desktop efforts in one fell swoop. The likely candidates? We think Linspire is the most attractive of the bunch. Yes, technologically it is dead even with Xandros, but Linspire has already made some inroads into the small to midsize business (SMB) space and has more contacts with independent software vendors (ISV).

Who will buy? We're thinking Novell. It has the desktop know-how already, thanks to its acquisition of Ximian, and its SUSE engineers will be able to re-tool anything the Debian-based Linspire has under the hood. But don't rule out Mandriva as a buyer. It has desktop strengths, too, as well as a big international presence. The Debian hurdle may not be a deal-killer. After all, earlier this year, it was rumored that Mandriva was thinking about joining the Debian Core Consortium.

And, just to make things interesting, don't rule out Mandriva as an acquisition target. Novell would do well to pick it up, because it would increase its already strong desktop offerings and international channels.

So Goes SCO

At the beginning of 2005, it was clear which way the wind was blowing for SCO. If we were a betting sort and we participated in a Corporate Dead Pool ... oh well, never mind. It didn't take much of a crystal ball to see SCO's future. Even Groklaw is looking beyond the Lindon, Utah-based company, and we picked up a definite who-cares, no-legs vibe about the law suits among enterprises and ISVs at tradeshows this year.

SCO v.IBM and SCO v.Novell should be lumbering into court rooms some time in 2006. We'll be stifling a yawn and watching, unless SCO runs out of money or its assets are acquired before then. Neither of which we'd mind.

We do hope, however, that another ISV picks up OpenServer 6. It's good stuff, and SCO's foibles shouldn't influence the results of years of development efforts, although we can't see any enterprise in its right mind buying the product from its current vendor.

Enterprise Unix Efforts Dribble

Server sales in the first quarter of 2005 hit a water mark. According to IDC, for the first time Unix server revenue and Windows server revenue were tied at $4.2 billion. In the third quarter, Windows factory revenue hit $4.6 billion, making it the largest single segment of the server market for the first time.

With Unix and Windows each currently controlling one-third of the space, there's still a pretty decent berth for Linux expansion, and Linux continues to grow rapidly. Enterprise Unix sales, however, are on the decline. This has been attributed to both the rise in Linux servers, which don't seem to be moving in at the expense of Windows and the increase in x86 volume servers.

Obviously, Unix variants throughout the spectrum are feeling this. Possibly none more so than AIX and HP-UX, which have earned reputations as being the poster children for enterprise Unix.

IBM has thrown much weight behind AIX in recent months. The most recent development, announced last week, was a Unix Collaboration Center. Although it was described by some as being largely smoke and mirrors, and despite hearing word on the street that AIX's days are numbered and Big Blue's support for it will substantively turn into migration programs, we're not inclined to cry "abandon ship!" just yet. There are plenty of AIX shops still out there, and enough legacy AIX-based apps to keep the fires burning. However, there isn't a whole lot new about AIX and fewer new products running it. We doubt this will change in the future. Which, of course, means fewer new customers, and that brings us back to the rumors on the street.

HP, on the other hand, appears to be throwing its full weight behind HP-UX, having chosen it as the Unix OS of choice in its post-acquisition state. HP continues to release HP-UX-powered servers, and in early November released the BL60p, the vendor's first Unix blade. It's also the first Itanium 2 blade to hit the market. The BL60p runs only HP-UX but can share a chassis with Opteron and Xeon blades. We still think it smacks of customized-solution-gone-public, but it does make a statement as to HP's faith in HP-UX.

Still, we're not sure how many takers will be new customer wins compared to current HP-UX shops. With the surge in Linux deployments and contraction in Unix sales, our inclination is to think status quo, at best, for both operating systems 2006.

Prediction Wildcard

Finally, here's the one we think might happen, but we really hope won't. Google has been the bane of Microsoft and friend of all things open source for quite some time. But we have started to see little signs and inklings that Google may not be such a good friend toward all things open source much longer. Specifically Linux.

In recent weeks, Google has been getting friendly with Sun. At first it seemed a natural outgrowth of the animosity both companies have for and receive from Redmond. But, deep in the fearsome caverns of our hearts, we wonder, will Google ultimately shift toward a more Solaris-friendly stance? It makes some sense from a business standpoint. After all, in Sun you get one company, one operating system, and one platform — all, by the way, in various stages of being open sourced. Since Novell and Mandriva are in business flux, and no one knows what the heck Red Hat is doing, is it beyond the realm of possibility that Google will embrace Solaris? Certainly Sun wouldn't mind. Because beyond the immediate influx of revenue, Sun would also get something it really needs: access to a strong developer community (which, as we pointed out earlier, is a key component to making Solaris grow).

Too out there? Maybe. But in 2006, we see some big shake-ups coming for the enterprise status quo.

Enterprise Unix Roundup will not be published next week. We wish you a safe a joyous holiday season and look forward to bringing you the latest news and analysis in 2006.

>> To Tips of the Trade