The Desktop Dilemma

By Timothy R. Butler | Posted at 4:55 PM

As the economic downturn has taken its toll on GNU/Linux companies, many surviving companies have decided to move away from the desktop market, and focus on the server instead. This escalated to an alarming pace after the death of Eazel, a startup that created the Nautilus file manager, and while the focus-on-server mania has calmed down lately, it is still very much alive. Yet, it seems to me, that focusing on the server will in the end cause these companies to lose not only the desktop, but the server as well.

Mandrake's Move up the Charts

The first exhibit that I present to you to show that focusing on the server will lead to loss of marketshare, is Mandrake's rapid ascent in Netcraft's server survey. In recent months, it has been found that Mandrake's Apache/AdvancedExtranet server (a specially tuned version of the Apache web server, along with servers for other functions) has surpassed the usage of RedHat's Stronghold Apache server. In fact, AdvancedExtranet's overall growth was six times that of the web's average.

“Yet, one has to wonder why RedHat is losing ground to a company that is traditionally aimed at the desktop.”
The irony here, as some of you may note, is Mandrake is primarily a GNU/Linux distribution focused on the desktop/workstation market. That is not to say Mandrake Linux's server offering is not an excellent product, all accounts seem to say it is. Yet, one has to wonder why RedHat - a company focused on the server - is losing ground to a company that is traditionally aimed at the desktop.

The answer really is not that hard to figure out. As one Open for Business reader noted, when he discovered that Mandrake Linux was better for a desktop than RedHat, he switched to using it on his desktop. As the one in charge of managing the servers for a company with over a billion dollars in revenue, his increasing knowledge of Mandrake on the desktop lead to the eventual replacement of all RedHat-based servers with Mandrake counterparts. This demonstrates the power of supporting both the desktop and server - one deals with the desktop all the time, so if you can use the same product for the server as well, it makes sense to do so.

The Palladium Paradox

A handful of months ago a new technology surfaced in Newsweek called Palladium. The purveyor of the technology - none other then Microsoft - claims the purpose is to herald in the era of secure computing. Naysayers claim it is just another attempt by the Redmond software developer to gain complete control of the PC. I say it really could not matter less.

Let me say that again - it really does not matter whether Palladium is meant to secure or control the PC industry. That is not to say what Palladium does is a non-issue, rather what it does on the desktop is a non issue. What we really should be concerned about is what Palladium does to the server. By moving to a proprietary “secure” internet protocol on the desktop, the server is handed over to those who support the protocol by default.

Obviously the server would become a shoe-in for Microsoft just by leveraging its current 95% marketshare in the desktop, no special desktop antics required. In possibly the most brilliant move in the company's history, it would reverse the server tables, becoming the only company equipped with the software needed to talk to most desktops. This is something the world's largest software developer has been unable to do by competing directly in the server market, as UNIX and Linux-based solutions have the upper-hand in reliability and cost.

In brief, Palladium could force the same type of migration as my reader took willingly. While switching to Mandrake servers was something that made sense, with Palladium it would not matter if it made sense or not.

“By conceding the desktop market to Microsoft, or anyone else for that matter, in essence they cede the server market as well.”
I believe this is the critical flaw with most of today's Linux companies' narrow focus on the server. What they fail to understand is that their strength in the server market will never be secure so long as they ignore the client market. By conceding the desktop market to Microsoft, or anyone else for that matter, in essence they cede the server market as well.

The situation gets worse if we extend our scenario a bit. Even if Microsoft decided to release the API's for “Palladium 1.0,” by the time the Free Software community could support it, Palladium 1.1 could be pushed out via a service pack. In fact, using a recent addition to Microsoft End User License Agreements (or EULAs), it may even be that to use a Palladium compliant system, you would already have given Microsoft the right to update your system at will.

It may sound paranoid to suggest Microsoft could get away with changing protocols whenever it suits their fancy, but the whole Palladium idea was about security, right? It could easily be claimed - and to a point be correct - that the protocol must change to avoid being cracked by script kiddies. After years of people railing against Microsoft for its woefully insecure systems, those people would appear to be hypocrites for attacking Microsoft's making a “secure” platform. Obviously this would not be the case, but the company's PR team could easily make it look that way.

What Blue and Red Should Do

Both Mandrake's increasing marketshare, which is a competitive threat to RedHat, and Microsoft's Palladium, which is an anti-competitive threat to the entire industry, could be solved if companies with an interest in the server market would take the desktop seriously. Companies like IBM have made great contributions to Apache and other server projects, and it would be relatively easy for them to push the GNU/Linux desktop the rest of the way it needs to go to succeed.

Big Blue has a lot at stake here, considering that its entire product lineup, from the mainframe to the ThinkPad laptop, is threatened if Palladium takes hold. The company's recent move to drop Linux support on the ThinkPad shows the exact kind of narrow thinking that could cause IBM to again snatch defeat out of the jaws of victory.

While I do not see the entire Palladium scenario I laid out above actually happening, I think it does prove with great force the point of this article. That is, without securing marketshare on the desktop, one's share of the server market will never be secure. Whether it is a French Linux distributor nibbling away at its bigger counterpart in North Carolina, or the behemoth in Redmond pushing “secure computing,” the message is clear - ignore the desktop market at your own peril.

Timothy R. Butler is Editor-in-Chief of Open for Business. You can reach him at
tbutler@uninetsolutions. com.