Buying Frenzy

Oliver Diedrich

Open source businesses as objects of desire

2007 was the year when buying open source companies became a trend in the software market, and this trend seems to continue in 2008. The buyers were traditional software companies as well as larger open source businesses wanting to round out their portfolios: Citrix bought the virtualisation specialist XenSource for $500m. Groupware supplier Zimbra was swallowed up by Yahoo for $350m. Sun took over Cluster File Systems (CFS), the makers of open source cluster file system Lustre, and recently MySQL AB for $1 billion as well as German company Innotek, producer of the virtualisation software VirtualBox. Xandros has acquired groupware producer Scalix. Nokia bought Norwegian Trolltech, supplier of the Qt toolkit and the embedded Linux platform Qtopia.

But why would anyone want to buy an open source business? For the software? That's freely available. To influence the development process? A small open source business is hardly going to turn its nose up if a big company is prepared to offer it cash in return for ensuring that specific ideas get implemented. It's also not possible to remove a potential competitor from the market by buying it up: Once software has been released as open source, it can't be made to disappear. And it's not about customer bases or turnover. XenSource, Zimbra, CFS, Scalix, Trolltech, Innotek, even MySQL AB – none of them earn so much money or have such great customer bases as to make them attractive targets for this reason alone.

There must be other reasons – and a common denominator can indeed be discerned in all cases. These companies are buying the knowhow of the developers and the rights to the code. As far as knowhow is concerned, all companies have plenty to offer. The XenSource development team is the driving force behind Xen, a virtualisation solution which is now being seen at as competitor to VMware. Virtual desktops may be the next great trend in virtualisation – and a big threat to Citrix's traditional terminal server business. Innotek is a virtualisation specialist, too – Innotek was one of the original developers of VirtualPC, now owned by Microsoft.

Zimbra demonstrated what could be done using AJAX and a browser for its groupware, the most spotlighted component of its Zimbra Collaboration Suite. Yahoo is clearly looking for a counterweight to Google Apps. CFS brings expertise in distributed storage, a boom segment in which Sun hopes to make OpenSolaris the system of choice. Scalix produces one of the most advanced competitors to MS Exchange – and Xandros is specifically positioning its Linux server as a solution for Windows networks and as an alternative to Windows Server. MySQL is the most popular database on the internet, and Trolltech has a lot of experience in embedded Linux.

Similarly valuable are the rights to the code. Xen, Lustre, Qtopia, MySQL and VirtualBox are licensed under the GPL, Scalix under the MPL - both licences are copyleft licences; they require that modifications to the code are also released as open source. Thus if Citrix wants to integrate a modified version of Xen into its products, or Sun wants to integrate Lustre's cluster file system, the VirtualBox technology or the MySQL database into Solaris, or Nokia wants to produce mobile devices based on Qtopia, the modified code must be published as open source – unless they own the copyright, in which case they can produce their own versions under other licensing schemes as and when they like.

XenSource did precisely that when it developed its virtualisation solutions based on the Xen hypervisor. Zimbra, Scalix, Innotek and Trolltech offer commercially licensed "improved" versions of their products with additional features. Adapted versions of Lustre are integrated into storage and cluster solutions. For a company which wants to do more with an application than just use or distribute it, it is the copyright to the code rather than access to it under the terms of an open source licence which is critical. This allows them to do what they want with the code without being subject to external licensing restrictions.

Developing or publishing software as open source in no way devalues the intellectual property, despite what many people think. On the contrary, it helps popularise and stimulate interest in the software, gives other companies the opportunity to look at what can be done with the code – and in fact, as long as the vendor retains the copyright, leaves all options open.

If you class selling a company for lots of money as success (Citrix paid $500mfor XenSource, Xandros $350m for Scalix, Sun $1bn for MySQL) these examples show what makes an open source company successful – expertise in a hot field, proven by an application to which you have the rights. This is no different to other startups, just that being open source makes it easier for other companies to evaluate the expertise of the developers and the usability of the software. VirtualBox, for example, was originally developed as a closed source product, but Sun did not notice the software until it was released as open source one year ago.

And what becomes of open source projects which acquire new owners in the course of a takeover? Citrix is a classic proprietary software producer, which clearly has no great interest on continuing to maintain an open source project such as Xen – they would like to pass this role onto a consortium with other businesses with an interest in Xen. Sun, itself an open source supplier, will continue to maintain Lustre, VirtualBox and MySQL but by obtaining the rights to the code has circumvented any potential licensing problems relating to OpenSolaris integration. Xandros also has no reason to allow the open source version of Scalix to wither on the vine – as a Linux supplier, they are anyway reliant on open source. And an open source application allows Yahoo to do a little to brush up its image compared to its competitor, the all-conquering Google. (odi)