Dell Terminates {code} - Architecting IT

Dell Inc has apparently pulled the funding for the {code} team, a set of projects to develop open source software within the company. As reported by The New Stack, the team is being wound up, with the work allegedly being picked up other teams. However, blogs like this from Kendrick Coleman indicate that perhaps the team are either being let go or prefer not to stay within the Dell family.

Credibility

In looking at the idea of developing open source within a company like Dell (or EMC when the project was first kicked off), we really have to look to the level of credibility these initiatives have. Remember ViPR from EMC? This was billed as the universal storage management engine, one platform for management to rule them all. Separation of data plane and control plane. In fact, the software was a rejig of the acquisition of iWave with some other components thrown in for good measure. EMC failed to get traction to expand support to other storage vendors, so opened the software up as “open source” via CoprHD in an attempt to get the community to contribute. This wasn’t really open source; it was a way to get support for non-EMC products back into a chargeable product. CoprHD seems all but dead and I never hear talk of ViPR these days.

With EMC’s history of hyperbole and history of selling commercial software, how did the company ever think that it would be easy to transition to an open source model?

Deliverables

Of course any project, open source or not, has to deliver something. What did the {code} team achieve within their 4 years of existence? At DockerCon (Austin, April 2017), I spoke to Josh Epstein, VP and leader of the {code} project about REX-Ray, a plugin for orchestrating persistent storage for containers. This seems to have been the only project of merit and has led to some of the developments seen in the Container Storage Interface (see my post here). Outside of this, every other project appears to be a piece of partially finished side work, with no real direction or strategy.

Even REX-Ray only reached version 0.4 in this time. We’re talking here about a plugin, not an actual product that even then only has limited adoption with EMC’s own products. It’s not really a lot to show.

Dell Rationalisation

Should we be surprised Dell canned the initiative? Not really. 18 months into the formal acquisition, we’re starting to see Dell gradually dismantle the old EMC empire. Many key figureheads have left and the business is being rapidly being restructured. We talked about this on a recent Storage Unpacked podcast with Chris Mellor (here’s his post from The Register). Dell (the company) certainly doesn’t need side projects that don’t contribute to the bottom line and I don’t think the company has any desire whatsoever to become an open source software vendor.

The Architect’s View

I feel a twinge of sadness for the people who worked on {code} as they look for that next opportunity. However, I do think that EMC was directionless in 2014 and looking to see what ideas would stick, rather than having a clear strategy. Open source was never going to work in the manner EMC established {code}. It looks more like PR exercise than an attempt to radically transform the business.

Where does that leave the projects? I don’t see any future value in REX-Ray as the approach of block storage for containers is a dead end. Like CoprHD before it, the code will continue to live online and perhaps at least be a guide as to where storage and open source should head (or not) in the future.

Credibility

Deliverables

Dell Rationalisation

The Architect’s View

Further Reading