Andrea Ruf, Open Source Automation Development Lab (OSADL) eG
Carsten Emde, Open Source Automation Development Lab (OSADL) eG

When the Open Source Automation Development Lab (OSADL) was founded seven years ago, the founding member companies had a clear goal in mind: Create an organization that

• gives automation and other industries a voice in the various Open Source communities and takes care that any missing software components are provided and maintained,
• establishes a test center with a wide variety of embedded and other systems to ensure industrial production quality of the released Open Source software,
• provides technical support, legal advice, certification services, scientific research and networking between interested companies and academia to foster the use of Open Source software in the industry, and
• first focuses on the real-time Linux kernel but takes care of other important industrial Open Source software projects as well.

Funding was intended through a flat-rate annual service fee based on the membership of interested companies in the OSADL cooperative. The plan was to acquire about 100 member companies in the first five years and to use the resulting funds to provide, at least partially, the above services. The OSADL funding members had hoped that the availability of these services would then attract more members and, thus, keep OSADL growing.

Seven years later:

• Numerous positive feedback and favorable press coverage confirm the need for an organization such as OSADL.
• The Linux real-time kernel is generally accepted in the industry and used in a large number of industrial products. The impressive download peak after the release of a new patch version speaks for itself.
• OSADL's test center (aka OSADL QA Farm) discovered important software anomalies and helped to localize and fix a number of them.
• Connecting industry, academia and the community has shown to be fruitful.
• OSADL's various consulting activities have significantly helped build confidence in the concept of Open Source and, thus, are minimizing users' technical, procedural and legal risks.

• The real-time patches still are not part of mainline Linux. In consequence, RT compatibility cannot be made mandatory for mainline code to be merged, but incompatible code may take a lot of effort to be adapted or removed later when the RT patch is upgraded.
• Fortunately and thanks to Steven Rostedt, RT patches are now professionally maintained on top of linux-stable, but they may not be as stable as the name suggests. In addition, a number of currently off-line patches still await their inclusion into the RT patch series.

What can we do to stabilize the RT patches and increase the merge speed? One solution could be to ramp up OSADL's efforts which primarily requires a better funding. By some reason, OSADL's member base is growing considerably slower than expected but OSADL has a large number of users worldwide. Thus, a few member companies are funding a work that is used by many non-members. To cope with this situation, OSADL is discussing a number of various scenarios that this paper will propose and explain. Since all scenarios deal, at least in part, with the modalities of developing, testing and releasing real-time Linux, the Real Time Linux Workshop appears to be an adequate forum to make these discussions public.