Oracle's Open Source Database Endorsement
Linux Outlook Research Brief

Hybrid Open Source Business Models


I was at the Web 2.0 conference a few days ago and spoke to Dan Farber of ZDNet, whom I run into routinely at open source conferences.  Dan mentioned that he thought a lot of the companies that are promoting open source really have more of a hybrid model.  For example companies like SugarCRM, JasperSoft, Al Fresco, Pentaho all have a very capable open source offering as well as additional commercial offerings which are closed source.  The open source products are hugely popular and fully functional.  There's no crippleware or limited demo; it's what-you-see-is-what-you-get fully open source software under an OSI compliant license.  And all of these companies have significant communities and a growing business that enables them to continue to invest in open source. So the products, both open source and commercial, continue to get better and better with each release.

In my view there's nothing wrong with the hybrid model as long as people are clear about what is open source and what is not.  At MySQL we have a model that all our software is open source (under the GPL).  In fact we've even open sourced software that was previously closed, such as the MySQL Cluster technology that was originally a closed source high performance telcom database.  We have a business model where we offer MySQL Network, a subscription service that adds value on top of our Community Edition with certified software, support services, a knowledge base, indemnification and software advisors.  When we launched MySQL Network we were very clear to everyone that our goal was to add value and not take anything away from the community.  That's a good approach and it works for us.  But there's more than one way to skin a cat, and other companies may choose a very different model.  In my view it doesn't make them any less open source than MySQL, Apache, Red Hat or JBoss et al.  If you like their software, use it.  If you don't, add to it or write your own.  There's no lock-in and you can always stay open source. If you want the convenience of additional features or services, then you can buy those. 

SugarCRM has a bunch of very smart guys who are totally committed to open source.  They were the first application company to dive into the open source pool and prove that open source can continue to move up the stack.  They happen to be down the street from us in Cupertino and users of MySQL so we compare notes on occasion about business, technology, SQL tuning tips and what not.  They have a free open source version of SugarCRM published under a Mozilla license and they have a Pro Edition that has some additional group capabilities that are not open source. (You can actually get the source code, but its not published under an OSI license.) But there's no hidden agenda or surprises and you're free to use the open source offering forever, to add to it, and you never need to buy anything from SugarCRM if you don't want to.  Dan Farber in his blog talks to John Roberts, SugarCRM CEO, about their model and John makes a pretty good case for why they chose it. 

SugarCRM is able to get the best of both worlds by building on open source and they commit over half of their resources (well over half for many releases) to working on the open source product.  They also have a thriving community that helps create add-ons, new features, translations and more.  All of that code stays open source.

What do you think?  Is there room for diversity of business models in the open source world?



Only one correction. Alfresco actually has more of a MySQL model, with the network/subscription component (everything open source but some services (e.g., enterprise-grade security) only available through the network). It's a case of closed network, but open source, much like MySQL, Red Hat, etc. It's not a SugarCRM model, as nearly as I can tell.

In my opinion, the open source business model will grow and take a bigger share of the market. Sun mentioned, they realised that Open Source products having the fastest quality improvements. Getting more comparable or even better Open Source products like Linux and OpenOffice this business model will win more companies which making money with services and value adding services.

The twin track play open source/commerical license options, I think has become the favored model for fundamentally open source based organizations. There remains companies who provide highly specialized but none-the-less indespensible components such as relational connectivity drivers based on standard APIs(JDBC/ODBC) that will play out more in a mixed source model. The advantage is to the developer - they can pick and choose the app server, or database they want - but given the understood criticality of database drivers, they have the optimal solution.

Kudos to SugarCRM for such a good product - both in terms of what it does and how easily it can be extended.

My one concern is that customers do not (fully) understand the benefits of the dual approach and often the 'Open Source' side of things need to be played down in case it scares them away.

This will probably change (slowly) but even then there will be two distinct groups of customers.

Zack - excellent points. What I find interesting is that as MySQL's Network evolves (and RedHat's and JBoss' too) into areas of provisioning, updates, monitoring etc, they start to look a lot like a hybrid model. Whether commercial value added functionality is delivered as software or as a service, it is still commercial value add to an open source solution. In the end, there may not be much difference between these models.

Some good points Zack. I agree that it is essential to the hybrid model to have a clear definition of what is open source and what is commercial, but I don't think that is enough. The open source product has to continue to be developed and improved. Its not enough that the open source products isn't crippleware today, it has to stay useful.

At JasperSoft, we have a clear line between the open source and commercial products. JasperReports (open source) is a reporting library, and JasperDecisions (commercial product) is a server with full intergration with enterprise directories, identity management systems, a repository, etc. Any new feature that is a core reporting feature goes into the open source JasperReports.

Another test of the hybrid model is whether or not there is any migration of features from commercial to open source. I think that as features are added to the commercial part of a hybrid model that one should expect other features to be migrated into the open source project, as possible.

Speaking of hybrid model, you also have the collaborative source approach ( which tries to limit (a bit) the Open Source Definition in order to enforce a strong qui pro quo among the users: those who collaborate get some free licenses, the others are "taxed" by some license royalties which will allow the payment of a dedicated full time team of developers. This looks to me like an enhancement to the other hybrid approaches as, if you are interested by contributing and by the commercial "Enterprise" edition, you do not have to contribute code while still being taxed for the enterprise features...

I do not believe this

The comments to this entry are closed.