Last year, I interviewed the head of the Linux Foundation, Jim Zemlin, about his own career, and about his organisation. That interview took place at the first European LinuxCon, which was held in Prague. This year, it took place in Barcelona, and I took the opportunity to catch up with Zemlin on what had happened in the intervening time (disclosure: the Linux Foundation paid for my travelling and accommodation while I was there.)
It seems it has been a good year for the Foundation, with a number of major companies joining up for the first time, or upgrading their membership. Zemlin says that income as a result has gone up by "double digit percentages".
The involvement of one company in particular has caused a few raised eyebrows. Microsoft's appearance as a "gold sponsor" of the European meeting has led some to suggest that the Linux Foundation had sold out by allowing Microsoft to speak at one session. But Zemlin says that the speaking session did not come as part of the sponsorship deal:
"There ain't an event sponsorship that they can pay us that's big enough for us to compromise principles. Part of our organisation's strength is the legitimacy that we've earned over a decade servicing this community in a humble, helpful way. We're not going to give that for a two-bit events sponsorship."
In the past, Microsoft has been uninterested in sponsorship; the fact that it has decided to participate this year is a further sign the initiative is shifting to open source. As Zemlin explains:
"They have to support Linux in their cloud, in Hyper-V; they genuinely need to participate in our community to service market demand."
Moreover, as he points out:
"What people should consider is just how dramatically Linux's role in the world versus Microsoft's role in the world has changed. The future of computing is what people are broadly calling the cloud. If you're a hosted software provider, having access to the source code so that you can modify it for large-hyperscale computing, is essential; being able to iteratively collaborate on the infrastructure that you use, is essential; being able to get it at a very low cost so that your service margins are higher, is essential.
In that world, Microsoft's problem is not whether or not they make good software; their problem is that fundamentally their business model and development model don't fit at all. That's how dramatically the world has changed."
And yet Microsoft's worst problem is not in the cloud, but at the other end of the spectrum, in small-scale computing systems. That's because the once-powerful lock-in to its Windows platform is not only absent, but has been replaced by a new lock-in – to Linux:
"The whole thing is flipped on its head. Microsoft had locked in the computing industry when the PC desktop was the critical form of computing, when the WinAPI was the de facto standard. The hardware used to be just something that MS would use to support the real lock-in, which was at the upper end of the stack, in the application APIs. Now application APIs are becoming less important. More developers are writing OS-agnostic apps, particularly in enterprise computing.
The new lock-in is at the bottom of the stack – that's what creating the network externalities that are really making Linux the de facto standard. It supports every architecture known to mankind: now every SoC [system on a chip] native development platform is Linux, all that driver support, all that hardware support [is for Linux]."
Board enablement is the big thing in embedded systems, where you have all these different chip makers that build what are called BSPs [Board Support Packages] - they are board software enablement for the hardware SoCs – it's real work. That translates into someone who's making a car navigation system or people who are manufacturing a Toys R Us children's tablet, they don't have any software engineers who are even remotely close to being able to do that kind of complex porting. What they need is to just grab an SoC that already has an open source BSP, just whack on an Android-based system or Linux-based system, and go."
This has led to an invisible revolution, whereby Linux is quietly but ineluctably taking over the world of embedded systems. Zemlin also thinks that Android has played an important role here:
"That's what's Android has given Linux, it's done all this board-enablement work. The reason is that all these SoC guys spend real money and time and super-smart people doing board enablement on their low-power SoCs. But once they have done it, all of a sudden they discover, hey – we can use this in refrigerators and in a car. Now it's there, and it ain't going backwards. That is the gift in addition to the user base that Android has given to Linux, which is a really big deal."
It's also another reason why Android's rise is inevitable, as I suggested earlier this week. However attractive Apple's iOS may be, you won't find it in refrigerators any time soon. That means the Android ecosystem will continue to grow unhindered into something vastly larger than one based around smartphones and tablets, with all that this means for Linux and open source.