Microsoft Employee is Leading Linux Kernel 3.0 Contributor
Posted on July 18, 2011 10:10 AM by Rob Williams
As strange as that title might look, don’t worry, it’s much too late in the year for April Fools’. K. Y. Srinivasan, a Microsoft employee, have landed themselves at the top of the Linux 3.0 kernel developer contribution list with a total of 343 changes. The information was first unveiled at Linux development-tracking site LWN, though the article currently requires a subscription (it will become free on July 21, 2011).
In all regards, Microsoft being at the top of any Linux development list is impressive, but it’s important to note that many of these 343 changes made by K. Y. Srinivasan are considered to be minor, thus the reason the number was able to get so high. It seems that most, if not all of the changes had to do with Microsoft’s Hyper-V driver, aiming to improve Linux hypervisor performance under Windows.
Although comparing the total number of lines-of-code contributed isn’t an accurate way to gauge importance, we can gain a little bit of an idea of Microsoft’s actual contribution when looking at it from that angle. With its 11,564 lines, Microsoft contributed a mere 1.3% of the total commits to the Linux 3.0 kernel, while Intel contributed a staggering 163,232, or 18.1%.
Nonetheless, why exactly does Microsoft care to contribute to the Linux kernel, when Linux is one of its leading competitors? The reason boils down to money. In order for a product that supports Linux to have true potential for adoption, companies must do whatever they can to get into the Linux kernel itself, and not have to rely on a patch. According to one blog, this is essentially life or death. If a solution of yours is built into the official Linux kernel, it can be taken a lot more seriously. Plus, it’s also easier to use that way.
The same blog notes that virtualization project Xen fell victim to the importance of this by allowing another competing project, KVM, to push ahead because their developers did focus on getting better-implemented into the Linux kernel. It’s not easy to get in there, either, given the stringent requirements. Sometimes many, many tweaks and re-writes will need to be done before the Linux devs feel comfortable letting something new into the kernel. A very good thing.