The business of making Linux easier to use

– By Robin ‘Roblimo’ Miller
The comparative complexity of administering Linux servers is a huge adoption barrier for small companies that can’t afford to hire their own sysadmins and don’t want to rely on outside consultants for all configuration changes or routine server maintenance. NetMAX is building a business by — as they put it — “Simplifying Linux Servers.” Their “underlying technologies” are Open Source, they say, but their GUI admin tools are proprietary. Is this combination the wave of the future both for companies selling Linux, as well as for companies that want to use Linux without hiring skilled sysadmins?Some months ago I was standing in the software section of a CompUSA store, looking at the — literally — hundreds of programs for sale there whose main purpose was to overcome one of Windows’ many deficiencies. I remember thinking, “If Linux took over as the world’s main desktop OS, an awful lot of these software companies — especially the anti-virus crowd — would go out of business.”

Or would they? Couldn’t they just start overcoming Linux’s many deficiencies instead?

Linux and other Open Source and Free Software packages do not always score high on the usability front, and many volunteer developers resent being asked to make their output easier for non-geeks to use, which is certainly their right. At the same time, usability issues plague Open Source adoption. So perhaps instead of looking at this situation as a problem, we should see it through capitalist eyes as a business opportunity.

This is what NetMAX does. For example, the blurb for their Professional Suite 4.02 says:

Our Professional Suite enables you to use the power and reliability of Linux as well as popular applications such as Apache, Sendmail, and Samba which are integrated into this product. The browser-based interface helps you to quickly and easily configure and manage the required network services, freeing network administrators from routine tasks.

As a former small business owner, I can tell you that this is a powerful sales pitch. The idea of being able to handle routine server admin tasks through an easy GUI instead of relying on someone expensive is a powerful draw. Indeed, this message is at the heart of Microsoft’s most successful business-directed sales pitches, and it trumps Linux’s vaunted reliability, security, and other fine features often enough that Windows is still the world’s most popular small business server OS.

Not free, but a lot less costly

The NetMAX Professional Suite costs $309, which is one heck of a lot more than downloading Linux, Samba, Apache, and Sendmail for free, but is also one heck of a lot less than buying Windows 2K or Windows XP Pro, IIS, and a Windows-based email package.

NetMAX product manager M.D. Squiers happily admits that his company’s products, while based on Linux and Open Source, are not Open Source, and spends no time apologizing for his company’s willingness to embrace a proprietary licensing scheme. But, he says, NetMax will soon be giving associated developers and VARs access to their source code.

(The formal announcement of this Shared Source-like NetMax change is scheduled for January 13, for those who follow such things.)

Squiers also talks about the company’s partnership with Toshiba. “They came to us,” he says, “and asked us to provide software for their Magnia SG20.”

He is not sure how many units this partnership is responsible for selling. “You’d need to ask one of the marketing guys,” he says. “I’m more on the tech side.”

Squiers says Toshiba is not the only company with which NetMAX is cooperating. He mentions EMUmail, a Webmail package NetMAX now sells along with its own products, and
notes that he and his coworkers have developed their own implementation of interchange, and are now providing commercial support for all interchange users because, he says, “Red Hat dropped [interchange support] a few months ago. It just happened that we were coming out with our own implementation, so we’re now providing commercial support for it.

Part of the reason NetMAX is doing well as a product is that parent company Cybernet Systems has its fingers in a number of high-tech pies ranging from online gaming to Internet-based medical monitoring systems. The company doesn’t rely on retail or business-to-business product sales for all of its income. It performs plenty of contract research for the U.S. Government, Ford Motor Company and the A.I. du Pont Institute, among others, and this keeps cash flowing in while Squiers and his coworkers work on NetMAX development, which they have been doing in one way or another — originally starting with *BSD — since 1995. It was originally an internal project started because, Squiers says, “We were having problems with [Windows] NT deleting files. In 1996 we started developing the GUI, so NetMAX has many years of development. It’s a stable product.”

What became the NetMAX value-add — the proprietary component of the product line — is the GUI-based configuration toolset, which is supposed to save so much admin time that no one minds paying for it.

In a practical example of how this can work, Squiers’ coworker Nathan Pitts says, “For instance, if you change your IP address it can impact all kinds of things — your Web server configuration, your ftp server configuration, your Samba configuration, your firewall, DNS (the obvious one) and your mailserver. What we’ve done is determine all the dependencies for those kinds of configurations. If you update one, all of them get updated that need to be.”

NetMAX didn’t do well through retail channels

NetMAX was originally sold through retail stores, not as one product but as seven separate ones. This was necessary, Squiers says, because “to be in the retail space, and to be in CompUSA, you have to have 40 different products [a bit of hyperbole – ED] just to be on the shelf.” But, he says, “That just wasn’t profitable.” Not only that, reviews of shrink-wrapped NetMAX products, while generally favorable, were not always praise-to-the-skies endorsements. This Feb, 2001 review is an example. Right under the headline, NetMAX FireWall worth the fickle installation, a subhead says, Good product, but marketing may be misdirected.

The reviewer’s main complaint was that this was a good commercial-level product, but had no business being sold as a consumer, off-the-shelf item.

Now the NetMAX line has been reduced to three products, all sold to businesses either directly or through ISVs or other commercial channel partners, either software-only or preinstalled on “plug and play” server appliances, notably by Toshiba on their Magnia SG20 servers (and sold through Toshiba system resellers).

However, Squiers believes the $1300+ typical list prices for Toshiba’s server software/hardware bundles may be too high for many small businesses. “I see a price point of around $500,” he says, and points to NetMax’s own offering of their Professional Suite and VPN products, both preinstalled on a (now obsolete, but powerful enough for this application) Magnia SG10 — for $495. And when these units are gone — which Squiers says “won’t be long now, as rapidly as they’ve been selling” — NetMAX plans to find another piece of low-cost hardware to serve as an appliance base. Squiers is confident that he will be able to find white box (or possibly brand name) low-end servers “in the $200 range, no problem.”

Can this business model be duplicated by others?

We’re starting to see more proprietary products “built on an Open Source base” coming to market, where the underlying software is licensed under the GPL or another OSI-approved license with some sort of proprietary GUI or configuration toolset added on top of it.

We’ve already written about Sourcefire and OpenMFG, which both follow this pattern. (Apple’s OS X also follows it, since OS X is essentially a proprietary GUI and usability layer built on top of an Open Source BSD core.)

No doubt, there are many other companies that are also quietly following it, because this seems to be a viable business model; almost all the companies we’ve talked to who use it seem to be making money or at least expect to turn a profit in the near future. This pattern doesn’t pass the “give the product away and make your money offering service and customization” Free Software ideological purity test, but it is easy for customers and investors to understand.

What OpenMFG, Sourcefire, and NetMax have in common (besides licensing style) is that they all make Free or Open Source software easier to use in some way, along with offering features that aren’t currently available in pure Open Source or Free software form.

The danger of basing a business on the idea of overcoming Open Source deficiencies is that volunteer programmers may eventually come up with programs that are just as good as yours, possibly better, that are available for zero dollars.

But then, aren’t all commercial software vendors now faced with this threat? And aren’t software vendors that are locked to a proprietary operating system in an even worse position than those who use an Open Source base?

Besides the threat of free/Free software competition, vendors who deal with proprietary operating systems run the risk of having the operating system vendor suddenly decide to make their businesses obsolete by incorporate something functionally equivalent to their products into the operating system itself.

This has happened many times to applications developers who decided to make their fortunes in the Windows or Mac “worlds,” and it is undoubtedly going to happen many times in the future to developers who decide to overlook the potential of Linux and *BSD development (or ports) in favor of tying themselves strictly to proprietary operating systems.

In the end, no matter what operating system or licensing style is involved, software development is a risky business, with more entrants in the field going broke than ever turn a profit. But companies that develop from an Open Source base tend to have lower initial investment requirements than those that rely on a proprietary base, and if they are producing commercial-level products that occupy an entire server they have a sales price advantage over proprietary operating system users because they don’t need to buy an OS license for each unit sold.

Will the “Open Source base plus proprietary add-ons” style eventually become the standard development pattern for all server-level software?

It is entirely possible. More and more vendors, from IBM to “kitchen table” software authors, seem to be heading in this direction.