Now that my PDX Trimet pass expired, it’s time to reflect on OSCON 2014. Unfortunately, I did not ride my unicorn home on a rainbow; this year’s event seemed to be about raising red flags.
My four key observations:
- No superstar. Past OSCONs had at least one buzzy community super star. 2013 was Docker and 2011 was OpenStack. This was not just my hallway track perception, I asked around about this specifically. There was no buzz winner in 2014.
- People were down on OpenStack (“DownStack”). Yes, we did have a dedicated “Open Cloud Day” event but there was something missing. OpenSack did not sponsor and there were no major parties or releases (compared to previous years) and little OpenStack buzz. Many people I talked to were worried about the direction of the community, fragmentation of the project and operational readiness. I would be more concerned about “DownStack” except that no open infrastructure was a superstar either (e.g.: Mesos, Kubernetes and CoreOS). Perhaps, OSCON is simply not a good venue open infrastructure projects compared to GlueCon or Velocity? Considering the rapid raise of container-friendly OpenStack alternatives; I think the answer may be that the battle lines for open infrastructure are being redrawn.
- Free vs. Open. Perhaps my perspective is more nuanced now (many in open source communities don’t distinguish between Free and Open source) but there’s a real tension between Free (do what you want) and Open (shared but governed) source. Now that open source is a commercial darling, there is a lot of grumbling in the Free community about corporate influence and heavy handedness. I suspect this will get louder as companies try to find ways to maintain control of their projects.
- Corporate upstreaming becomes Imperative. There’s an accelerating upstreaming trend for companies that write lots of code to support their primary business (Google is a primary example) to ensure that code becomes used outside their company. They open their code and start efforts to ensure its adoption. This requires a dedicated post to really explain.
There’s a clear theme here: Open source is going mainstream corporate.
We’ve been building amazing software in the open that create real value for companies. Much of that value has been created organically by well-intentioned individuals; unfortunately, that model will not scale with the arrival for corporate interests.
Open source is thriving not dying: these companies value the transparency, collaboration and innovation of open development. Instead, open source is transforming to fit within corporate investment and governance needs. It’s our job to help with that metamorphosis.
Last week, RackN announced our enterprise support for Kubernetes using nothing but upstream Ansible from the project itself. This effort represents years of effort by the RackN founders to keep platforms interoperable via open and shareable operations automation.
This week, we added Install Wizard templates to the DC/OS install automation we build in collaboration with Mesosphere last year. That makes it even easier to run DC/OS on physical infrastructure. Like our Kubernetes work, the Digital Rebar automation uses the same community dcos_install.sh that’s used in the community documentation. The difference is that we’re also driving all the underlay prep and configuration automatically.
Rob Hirschfeld 