April 4, 2011

LF Collaboration Summit Preview: TI's Bill Mills on Yocto Project


The Linux Foundation Collaboration Summit is just two days away, and we were excited to be able to reach Texas Instrument's Bill Mills, chief technologist for open Linux solutions. Mills is participating in a highly-anticiapted panel on Wednesday titled, "Introducing the Yocto Project: What it Means for the Embedded Linux Industry," and shared a few thoughts with us before he prepares to arrive at Hotel Kabuki, including his idea of the state of embedded Linux.

You're participating at The Linux Foundation Collaboration Summit on a panel about the Yocto Project. Can you tell us more about that project and why it's important?


Mills: Texas Instruments (TI) has been a user of OpenEmbedded (OE) for several years now. We find the power it provides the integrator is key for our customers to maximize the benefits of embedded Linux and TI silicon. However, TI discovered various aspects that could be improved to make OE better, such as enabling OE support in commercial products and making it easy to use by every engineer. To me, Yocto is about working collectively on those needs.‚Ä®‚Ä®I think the attention and effort on Yocto has already led to improvements in OE itself with the formation of the OE-core concept.  This is an ideal example of a Yocto expressed use-case being addressed in up-stream OE.‚Ä®‚Ä®

Your panel follows Jim Zemlin's opening keynote on the state of the Linux union. What would you say is the state of embedded Linux union today?

Mills: ‚Ä®‚Ä®Embedded Linux is very healthy. It is finding uses in almost every type of mainstream and vertical market.  Knowledge of how to develop with embedded Linux is the new entry-level requirement for many companies.‚Ä®‚Ä®I still think there is room for improvement. If you draw an analogy to desktop Linux, I think embedded Linux is at Slackware 1 or 2 level, and people are looking forward to the live CD generation. Like the desktop evolution, I think much of the change needs to come from tooling and distribution technology, with only a few key enablers coming from the kernel.‚Ä®‚Ä®

Linux will celebrate its 20th anniversary later this year. What's your prediction for its future and what needs to happen to see your prediction realized?

Mills: I see Linux as the flagship of a mesh of open source technologies going into the future.  I think we will see more diversity in user space solutions appear but then settle down to two or three. Some deployments of these technologies in embedded spaces may not even use the Linux kernel, but even then Linux, will shape how the mesh is structured and interacts.‚Ä®‚Ä®I think two things are important for Linux to be successful in the future: using open source cloud services and developing good, cross-platform, open source applications.

How does collaborative development, like the work the Yocto Project is doing, benefit TI's overall business and technology strategy?

Mills: TI's job is to make the right silicon for every customer's product. The easier it is to use the hardware, the more chips we sell.  From the software side, our goal is to make Linux easy to use, develop, and ship in volume. Cooperating with people with the same goal just makes good business sense.  Motivated self interest is the engine that drives open source for individuals, and it is no different for companies.