Open Source Inflection Points
Open source has been through a lot over the past 30–40 years, both technically and culturally.
Shortly after I joined WSO2 in November, Joseph Jacks asked if I’d be interested in participating in the Open Core Summit Dec. 16–18.
The resulting interview touched on multiple aspects of open source, including my experiences with it as an open source vendor (with both IONA Technologies and now WSO2), as a former participant in OSS communities, and the perspective working within the financial services world. Also how multiple inflection points have changed OSS along the way.
Most of my perspective comes from the years at IONA when I was tasked with leading the company’s open source strategy. I served on the Boards of ObjectWeb and Eclipse, and helped set up projects at ObjectWeb, Apache, and Eclipse. Some of the OSS projects we started or acquired continue to thrive at Red Hat under the Fuse brand (who acquired them from Progress after Progress acquired IONA back in 2008) — Apache CXF, Camel, Karaf, Active MQ, and ServiceMix.
I’m well aware of the Richard Stallman origin story with GNU and CopyLeft (see previous link).
Another big inflection point for me was JBoss. I know Linux and the Apache Web Server predated JBoss, but this was the first time I could see an open source product truly commoditizing (and in dramatic fashion) a category of proprietary products.
This meant the industry knew what an app server was. The proprietary model charged for the use of intellectual property developed at great expense in the R&D function of a tech company. This model was built around rewarding those who created *new* things. At one time the industry did not have databases or app servers or TP monitors.
And of course it took a while for us all to figure out what an operating system was, meaning what was the standard for an operating system. Or database, or app server, etc.
The analogy I use is an old one — initially Microsoft Word did not replace special purpose word processing machines, because it didn’t have enough features. Once it crossed that threshhold, and everyone could create documents on the general purpose PC, special purpose (and proprietary) word processing systems died out. After a while, no one really needed new features in Word, but Microsoft kept creating new versions anyway because that was how the proprietary software model worked.
Once a given product category, such as an app server or relational database, reached the point at which new features didn’t really solve new business problems — not to the extent of tipping the balance on a purchasing decision- the category was ripe for commoditization and open source took care of that by eliminating the surcharge for intellectual property (which by definition wasn’t needed after the product reached this level of maturity).
Today we are in the middle of another inflection point. When Google invented the commodity computing platform we now recognize as “cloud native,” it succeeded because it was the most cost effective IT environment ever developed. Part of the reason for this was its extensive use of open source.
The widespread adoption of the commodity platform initiated a huge innovation cycle because SW had to be engineered specifically for that environment to realize its benefits. The balance of innovation in the industry shifted from the proprietary vendors (who were already being marginalized) since their products were not engineered for this environment- were not cloud native, to use the current phrase.
Lots of innovative open source projects grew up around the cloud native platform (just look at the CNCF project lists). But then the cloud providers started competing based on the services offered by their platforms, and included many of these open source projects. This is another inflection point but I don’t think we know the end of this story yet.