In association with heise online

01 October 2010, 14:54

Guest commentary: Open Core is over

A response to the article "The hype is over" published on the 1st of September

by Peter H. Ganten

Commercial open source software fills a market need – and it works. What has outlived itself is the open core idea.

Peter H. Ganten
Peter H. Ganten is the managing director of Bremen-based Univention GmbH

One of the core statements Oliver Diedrich makes in his article The hype is over is that commercial open source software has outlived itself. According to the author, an important reason for this is that most of this software belongs to the "open core" category – products where only the software core is available under an open source licence, while other components which are equally important for professional environments are provided via classic proprietary licences

I believe that the opposite is true: Commercial open source software is becoming an increasingly important and strategic IT aspect for more and more businesses and authorities, and the sustainability offered by the open source concept is only just being discovered in many places. However, this isn't so with "open core software" which, strictly speaking, isn't even open source software because it doesn't offer users the advantages of the open source concept. Which is also why I know hardly any vendors who have had lasting success as open source companies while following the "open core" philosophy. At Univention, we also started out with a kind of "open core" strategy, but we already switched to 100% open source back in 2006. Interestingly, we started being really successful right at that time and have consistently recorded double-digit growth rates ever since.

"Open core" always used to be the problem with SUSE (proprietary Yast) and, since the take-over by Novell, the company's proprietary components have been called eDirectory or Groupwise. Red Hat with its "perceived" 99% of open source components has been considerably more successful. As correctly pointed out in the article, people quickly realise that "open core" doesn't actually represent an advantage, and our company also encountered the described "dialectic problems" before we made the decision to release our software as 100% open source code.

However, this doesn't mean that commercial software, and especially commercial open source software, has become redundant, because professional software users often look for a supplier who will accept responsibility. This means product liability, support, maintenance, SLAs and so on. And who could provide such services better than the original vendor? In the case of commercial open source software, this often involves a classic proprietary software licence, although the same software can also be downloaded free-of-charge under an open source licence – but without liability, support or SLAs. Red Hat is a good example for this approach: A Red Hat subscription is, in effect, largely a classic proprietary software licence; however, users can also choose the CentOS alternative, which is built from the same sources. Ingres is another company which follows this approach, and so are we.

What's interesting about such commercial products from a business or public authority perspective is the calculable costs and risks. That companies with a €900-per-year support contract offering full ('flat rate') support and a response time of 4 hours for a complete Linux enterprise distribution can't be allowed to fiddle with their code has so far been plain to all our customers, who haven't perceived it as a disadvantage. People who want to fiddle either know what they're doing or need to factor in higher costs when a service provider first has to identify and iron out their mistakes.

What remains is a massive advantage over proprietary software, which in my view also includes "open core software": more sustainability. Every user of commercial open source software such as our UCS can, at any time, decide to cancel our services but can continue to use the software, develop it (have it developed) further, or obtain third-party support. However, this is normally unattractive because the original vendor can usually offer the best and most economic service in terms of ongoing development and support.

If the original vendor fails to do this, however, for instance because the company has ceased to exist or been taken over by another vendor, or because the original vendor has simply become too presumptuous or arrogant, then such alternatives as CentOS, MariaDB or all the *buntus quickly appear on the scene. I'm sure that, should Univention ever cease to exist (we can't do more than our best to prevent it), someone would very soon take our code, hire a few of our people and offer an attractive maintenance package to our hundreds of existing support customers. And that's exactly what isn't possible with proprietary software.

This advantage is increasingly being acknowledged: Not long ago, the head of the IT department at one of the larger medium-sized companies (and a UCS customer) told me that his company recently had a Basel-II audit where software escrow and contractual regulations in case of vendor unavailability were among the topics discussed. He said that he could instantly tick off wherever open source software is in use, because the typical risks don't exist with such software. These are immediate, tangible advantages of the open source concept for companies.

Therefore, many professional software users require commercial open source products (rather than deceptively packaged "open core" software), commercial open source software requires vendors as well as ordinary, comprehensible licensing models, and both increase the IT flexibility, sustainability and profitability in companies and public authorities.

It is true, of course, that we commercial open source vendors only provide a fraction of the open source software, and that, when it comes down to it, we are dispensable. But this very aspect is one of the particular advantages :-)

Print Version | Permalink:
  • Twitter
  • Facebook
  • submit to slashdot
  • StumbleUpon
  • submit to reddit

  • July's Community Calendar

The H Open

The H Security

The H Developer

The H Internet Toolkit