Other LWN stuff:
 Daily Updates
 Calendar
 Linux Stocks Page
 Book reviews
 Penguin Gallery

 Archives/search
 Use LWN headlines
 Advertise here
 Contact us

Recent features:
- RMS Interview
- 2001 Timeline
- O'Reilly Open Source Conference
- OLS 2001
- Gaël Duval
- Kernel Summit
- Singapore Linux Conference
- djbdns

Here is the permanent site for this page.

See also: last week's LWN.

Leading items and editorials

• Ben and Jerry's companies operate in a slow, "organic" growth mode, usually without large outside investments. The founders of a Ben and Jerry's business may have large and ambitious goals, but they realize that it will take many years to achieve them. These companies usually seek to make money from the beginning, because they have to.

• Amazon companies operate in the "land grab" mode - the goal is to get as big as possible as quickly as possible, with small details like profitability having little importance.

Obviously the two different types of companies are two very different places to be, though each has its advantages. Beyond that, however, the author makes two points about these company models that are worth some consideration. First is that every company has to choose one or the other and be serious about it. Attempts to mix the two approaches will fail.

The other is that the market the company is in can determine which model will be successful. Land grabs can be called for when there are large network effects, and there is little competition in the area already. The Ben and Jerry's model makes more sense when competition exists, and customer lock-in is relatively weak. Amazon had to grab the turf before somebody else did, while Ben and Jerry's could not have done an ice cream land grab whether it wanted to or not.

So how do these models apply to the business of Linux? Until sometime around 1998, Linux companies all followed the Ben and Jerry's model, for the simple reason that the funding for aggressive growth did not exist. Companies like Red Hat, SuSE, VA Linux, LinuxMall.com, and others grew slowly for years as Linux slowly caught on.

Much of that has changed over the last two years, as funding for Amazon-type companies became available. Now, it seems, being a proper Linux company requires regular acquisitions, constant growth, and wild expenditures of cash. Lineo, currently in the IPO process, is a classic example of the new type of Linux company. By way of no less than six acquisitions, it has sought to gain both size and market share as quickly as possible. Most other high-profile Linux companies are also operating in the Amazon mode.

But is that the way that Linux businesses should work? Land grabs work when there are strong network and lock-in effects. Linux does have the usual network effects associated with operating systems - every user who adopts Linux makes your Linux system worth more. But this effect is truly a tide that lifts all boats, because lock-in with Linux is nearly absent. That is, after all, one of the advantages of Linux. Changing from one distributor or support provider to another may not be entirely fun, but it is not that big a deal, either.

Thus territory occupied by land grabs is going to be very hard to defend. Massive marketing budgets and acquisitions may well succeed in getting customers, but those same customers will head for the door quickly if they are not happy with what they find - they have many other alternatives. When you use Linux, you are not locked in to any vendor.

Linux itself has grown in the Ben and Jerry's mode - slowly, patiently, and without the ability to "acquire" market share. It may well be that the most successful Linux businesses, in the end, operate in the same way. The model for these businesses may not be a corporation at all; it may, instead, be the Debian project. Debian continues to grow in users and developers both, despite its publicly-traded competition. Wouldn't it be interesting if the Linux business model that found the most success in the long term were based on some variant of Debian's constitutionally-driven cooperative, distributed setup?

Oracle's "Internet appliance" system is out. The "New Internet Computer" (NIC) is the latest result of Larry Ellison's long personal crusade to make non-Microsoft systems available to the world. It's aimed at people who only want access to the net; as such, it's essentially a $199 (without monitor) X terminal. It has a 266 MHz processor, 64MB of memory, Ethernet and USB ports, a CDROM, and a built-in modem. And, of course, a Linux operating system. (Some) more information may be found in the press release.

What it doesn't have is a hard drive. The NIC boots directly from CD, and the entire system is to be found there. A bit of flash memory is provided for the saving of bookmarks (and, presumably, cookies); otherwise the system is completely read-only.

The creation of a "no maintenance" system makes a lot of sense for the intended audience. One wonders, though, what will happen when the first nasty, remotely-exploitable security hole in the NIC is found. Will Oracle ship new CDs to every customer? Even if the company were willing to take on that expense, how quickly could that distribution be done? One can hope that the NIC has been configured with security in mind, but, even so, the possibility of widespread, unfixable security problems is a bit worrisome.

The CD-based nature of the system also makes customization difficult in general. Fortunately, the architecture of the NIC is more open than, say, the iOpener. If the device is successful, one can expect to see a wide market in value-added replacement OS disks. One could imagine creating NICs oriented toward a number of dedicated tasks, from recipe management to point-of-sale or factory floor applications.

It is far too early to say whether the NIC will be a successful product. The network appliance market is still unformed and untested, after all. But this system shows, perhaps, how Linux may take over the desktop - and countertop - after all.

Software Patents Still on European Commission Agenda. EuroLinux has issued this update warning that the European Commission agenda still includes the topic of software patents. Whatever your opinion on this matter, be aware that following the issue and making sure you have an opportunity to voice your opinion before a decision is made is important.

"The General Directorate for Internal Market at the European Commission has not changed its ideologic position in favour of software patents."

A wealth of conferences is on tap for next week. The O'Reilly Open Source Software Convention is happening July 17 to 20 in Monterey, California; LWN's Dennis Tenney will be attending, and we hope to have contributed reports from another attendee. Those interested in hard-core hacking, however, will be heading north to the Ottawa Linux Symposium, held July 19 to 22. LWN editor Jon Corbet plans to be there, if some last-minute complications don't get in the way. The Ottawa event, limited to 500 attendees, sold out on June 23.

Inside this week's Linux Weekly News:

• Security: ssh license restrictions a boon for openssh, watch out for patent issues, though.
• Kernel: Fixing the latency problems, memory management issues.
• Distributions: Debian's 0th conference, Red Hat HA server, SuSE on the IBM S/390.
• Development: geda, Mozilla developer meeting, Erlang conference, real time Linux.
• Commerce: Deals, new products, Napster and Sourceforge.
• Back page: Linux links, this week in Linux history, and letters to the editor

This Week's LWN was brought to you by:

• Jonathan Corbet, Executive Editor
• Elizabeth O. Coolbaugh, Managing Editor

Next: Security