The last year for Linux was an exciting one. The little hobby project from Finland turned 20, rolled over the 3.0 milestone, and dominated almost everything from mobile devices to supercomputers. That’s a tough year to follow, but 2012 will still be an interesting year for Linux.
Linux doesn’t really have a roadmap. What you get is the result of collaboration among hundreds of developers from many companies. Nobody is setting up a feature list and directing developers to work on them—things make it into the kernel (or are taken out) when a developer (often as part of their day job) submits a patch and gets it accepted. This isn’t always a straightforward process, and it can take months or years for some features to make it in, if at all.
But if you pay attention to discussion in the Linux community, you can get a reasonably good idea what’s going to happen in the near future. Here are a few things you can expect to see in Linux in 2012.
One of Oracle’s big contributions to the Linux kernel is Btrfs, a filesystem that adds many features that enterprises would like to see in Linux. For example, Btrfs allows for snapshots, a maximum file size in the exabytes, compression, integrated RAID features and many other features you don’t find in Ext.
However, Btrfs has been missing a few features—most notably a filesystem check (fsck) tool—that you’d want before rolling it out for production use.
We should be seeing the Btrfs fsck tool early in 2012, and with that expect to see it pushed into some of the community Linux distributions like Fedora pretty rapidly. If I had to make a prediction, I’d say that Btrfs will eventually supplant Ext4 as the default filesystem for most major Linux distributions. Don’t expect to see it in widespread use in production before mid-2013, though.
Android, ARM Alignment, Embedded Focus
Linux in embedded devices will continue to be a major focus in 2012. This includes everything from set-top boxes like the Roku to Android phones and tablets, to printers and just about anything else you can think of.
A lot of fuss has been raised in the tech media about Google’s Android being a “fork” of Linux. Here’s what you haven’t heard much about since then—the Linux kernel folks and the Android folks have been doing their level-best to sync up the mainline kernel and Android’s kernel.
With the 3.3 kernel, much of Android’s functionality should be present in the mainline kernel. Not everything, but progress is being made pretty rapidly. If all goes well, users should be able to run Android on top of a vanilla kernel by the end of the year.
Note that this sort of asynchronous inclusion is not unusual, nor a reason to panic. Xen was outside the mainline kernel for years while the Xen folks learned to work with the larger kernel community (and vice-versa). Red Hat and other major Linux distributions have always included patches for features or devices that haven’t yet made it into mainline.
At the same time, the kernel folks continue to tame the “wild west” of the ARM architecture. At one point, there were about 70 sub-architectures in the ARM tree of the kernel. Compare that to other architectures, and you see there’s a bit of a problem. You’ll see a lot more work on ARM this year in the Linux kernel and a little bit less of vendors going it completely alone.
Part of that is the long-term support kernel tree for consumer electronics vendors. The Long Term Stable Kernel Initiative (LTSI) is focused on producing a stable kernel that will be available for about the same amount of time as most consumer electronics devices (two to three years). Getting vendors to work together on a single kernel should provide a lot of benefit, similar to the way that enterprise Linux vendors have largely standardized on the same kernels in their offerings.
Canonical has also announced at CES that they’re making a push into set top boxes and DVRs. It will be interesting to see whether they get any traction with major manufacturers. I’m skeptical about their chances, but if they succeed in getting a few devices to market with major brands I think that users will like them.
Better Tuning and Provisioning
The control groups (cgroups) feature has been around in Linux for a while, but it continues to evolve and allow finer-grained control of Linux systems. For example, in Linux 3.2 (released on January 4) we got a new feature called CPU bandwidth control that allows admins to define how much CPU time a process group can use.
The 3.2 release also added “thin provisioning” to the Linux Device Mapper, which allows admins to over-provision on user storage quotas. That might sound iffy, but if you have a few hundred users on a system you need to set an upper limit on how much storage they can use. Of course, if set properly, most users won’t approach that quota. So you probably don’t need enough storage to give every user their max quota.
You can expect to see continued improvement in cgroups and other areas in the kernel to allow enterprise admins to set resource limits and otherwise fine-tune their systems even further. The improvements may not turn up in enterprise releases for a few cycles (it’ll be a while before you see a 3.x kernel in an enterprise release) but they’re on the way. I suspect that SUSE Enterprise Linux will ship with a 3.0 kernel in 2013 when they move to SLES 13.
Ascent of OpenStack
Finally,expect to see OpenStack finally seeing minor production deployments by the end of 2012 or early 2013. The project was announced in 2010 and has attracted more than 140 companies and organizations. It’s also the IaaS stack of choice for SUSE Linux and Ubuntu.
OpenStack has a lot of backing, but the releases so far haven’t quite been ready for prime time. The Essex release in April should be approaching a state where early adopters might be ready to start some test deployments, and the following release in the Fall should be ready for more widespread use. From my conversations with people in the OpenStack community, SUSE, and Rackspace, my thought is that you’ll see a fairly heavy push for OpenStack by mid-year.
Basically, Linux will continue to improve on the enterprise and mobile side. With the Kindle Fire and Barnes & Noble’s Nook, I think we’ll see more and more Android tablets finally being accepted. Unless something drastic happens, Android will continue to hold a firm lead in the phone market. The only place that Linux will continue to languish is on the desktop. But you can’t have everything, right?
Joe ‘Zonker’ Brockmeier is a freelance writer and editor with more than 10 years covering IT. Formerly the openSUSE Community Manager for Novell, Brockmeier is a longtime free and open source software advocate. He has written for many publications, including Linux Magazine, Sys Admin, Linux Pro Magazine, IBM developerWorks, Linux.com, CIO.com, Linux Weekly News, ZDNet and many others.