What are the best OSS business models for CCII and CEVA?
Matt conclusion is that “such thinking, among other considerations, led Appcelerator to drop the GPL for Apache […] at least for commercial open-source projects.”
Bill Burke from RH/JBoss added its own comments:
http://bill.burkecentral.com/2009/08/13/your-oss-license-is-mostly-irrelevant/
http://bill.burkecentral.com/2009/08/16/success-for-apache-or-for-asl/
And finally Eric Barroca, CEO from Nuxeo, wrote another interesting blog post on the same open source subject last week.
As Eric is telling it, there no such think than an Open Source business model. And the source of articles on this subject is not about to dry up. So here is my own piece of comments as I wanted to highlight some differences of OSS business models in regards of the software stack being exploited: either a CCII (Common Content & Information Infrastructure) or a CEVA (Content Enabled Vertical Applications) - cf: my previous blog post on this subject.
OSS business models applied to the CM industry: CCII vs CEVA differences
Limits between what are Content Infrastructure solutions (certain will say ECM ;-) ) or a Content Applications (e.g: WCM, DAM, RM,…) are currently not clearly formalized. This of course impacts how Open Source licenses are applied to Content Management Software and to their underlying commercial business models.
For marketing reasons organizations tends to model their OSS strategy behind one single OSS business model and this whatever the number and type of products available in their portfolio.
I personally think this is not the best strategy: CCII based solutions will best leverage some business-friendly type of OSS licenses which will ease widespread distributions while CEVA oriented solutions are better coupled to OSS licenses which allow improved economics.
a) Needs for a standardized, de facto, commoditized Content Infrastructure components
There are still a lot to specify and to standardize at the CCII stack. In order to be widely used and to avoid recreating another new generation of content silos, the Content Management industry needs some robust but widespread OSS frameworks. They will act as the next de-facto standards and help promote the fast and wide adoption of new paradigms. Recent examples could be found in the Apache Jackrabbit (Reference Implementation of the JSR170/283 standard – courtesy of Day Software for the larger part) and in the Apache Chemistry project (sort of RI for the Oasis CMIS project - courtesy of Nuxeo for the larger part). There are of course other examples (Apache Pluto as the RI for the Portlet API initially provided by IBM or Apache Shindig by Google). You will note that most of these projects are hosted under the Apache umbrella as one also needs to leverage a certain notoriety in order to rapidly become a “de-facto” standard: releasing a project under an ASL or BSD license on SourceForge is not enough.
It then looks like software projects that need to be rapidly and widely distributed will greatly benefit from a business-friendly OSS license rather than a more restricted one (or which include some derivative effect such as the GPL copyleft). This will then also probably be the case for the CCII platform or any sub-components or framework for the next years.
For example we may question ourselves in order to know if Terracotta, a vendor of open source caching solution, would not better benefit from using an Apache License contributed to the Apache foundation (e.g in replacement of Apache JCS?) rather than using a semi-viral Mozilla license with some badgeware (TPL License).
b) Needs for more controls and ROI for CEVA solutions
On the other hand Apache never really succeeded to be a nice place to develop finished ready-to-use solutions (aka CEVA in the CM industry). Who would be the interested to develop an Apache based DAM, DM, Collaboration, WCM,… under an Apache business model? Granted, there are couple of examples of finished products (Apache Jetspeed in the Portal industry or Apache Lenya in the WCM field), but did they really created any real traction on the market? Most successful CEVA solutions do not need to leverage someone else notoriety (e.g: SugarCRM; Zimbra, Alfresco, …) and rather need to control their reputation (either through improved controls on their codebase and contributions or on their brand).
Usually customers are also more entitled to buy a finished ready to use product rather than some raw semi-finished components. The ROI of R&D is then clearly emphasized on CEVA while being considered as a kind of academical research work on the CCII stack. CEVA are then often considered as a profit center by Software Vendors while CCII initiatives are more considered as a cost center. Then editors tends to find a business model which allow them to convert any users of a CEVA applications into some sort of paying customers while they try to leverage in the same time co-optition and shared work with 3rd parties on their needed CCII projects in order to lower their costs structure.
Synthesis
As mentioned by Eric Berocca companies like Day or Atlassian are perhaps the one that currently best leverage such a double licensing paradigm in the CM industry (to not mix with a dual licensing strategy which only concern the relicensing of the same code but based under different IP agreements).
This is perhaps because it is easier to tell to your customers that you are selling a classical proprietary software on top of an open source infrastructure rather than to explain to them that your content infrastructure projects is powered by an Open Source business model of type A while your finished ready-to-use solutions (CEVA) by another business model of type B.
Moreover you are most of the time either GPL or Apache fanatics in the OSS industry. You can not be fan of both. IMHO this is a wrong debate. My conclusion is that the usage of a business-friendly license is greater for CCII projects but that the usage of a proprietary (or even better: a dual license GPL+ commercial license) is better for CEVA applications. So let’s mix both business models.
Does this mean that 100% Open Source Content Management Vendors now better need to differentiate licenses used for their codebase according to their usage (e.g: Would Alfresco Core better leverage an Apache type of license while still keeping a GPL one for Alfresco Share (Collaboration) or Surf (WCM)?)?
What do you think?
2 comments
Regards,
Document management service

