The future of Linux looks very, very thin:
A host of technologies -- most prominently CoreOS -- are challenging the foundation of what Linux means and how it's sold. Has Linux become fat? According to CoreOS founder and CEO Alex Polvi, the answer is an emphatic "Yes!" As Linux workloads become ever more specialized, the need for a general purpose Linux distribution declines. Or, as one friend told me, "CoreOS is an existential threat to the Linux distributors."
The future of Linux, in short, may require an extensive diet. Fortunately, Linux's open source nature makes this kind of crash diet not only possible, but healthy. Whether it's equally healthy for the business of Linux is another matter.
Do I look fat in this server?
Over on InfoWorld, Paul Venezia has been harping on his idea for increasingly thin Linux. Venezia's idea and, indeed, a central tenet to CoreOS and its peers, is the reality that "there's a growing gulf not only between Linux desktop users and server admins, but also between Linux server use cases," as Venezia explains in a previous post.
Rather than a general purpose Linux with lots of functionality but also unnecessary cruft, Venezia envisions specialized, thin Linux distributions that do one thing well, like serving email. While not a new idea, it has gained considerable currency in the last few months with the rise of Docker and CoreOS.
Of the latter, Venezia writes:
Hubris? Maybe. But according to Gartner analyst Lydia Leong, CoreOS may not be a direct threat "so much as [it] 'expands [the Linux] market [while] tak[ing market] share in particular use case[s]'." Even so, she avers, it is "essentially" a clear and present danger to traditional Linux distributions."[CoreOS'] whole concept is an ultra-thin, bare-metal Linux build that is specifically tuned to run clusters of Docker container hosts. This, then, becomes the actual server 'distribution' in use for all computing resources. The containers running on top of that core can't really be considered servers, since they're usually running a single process and are probably more appropriately considered static processes that carry their own dependencies along with them....[T]his turns the concept of a Linux server distribution on its head."
When I asked CoreOS founder and CEO Alex Polvi about CoreOS' impact on traditional Linux distributions like Red Hat Enterprise Linux, Ubuntu, and SuSE, he said, "If we are successful, the base OS does not matter anymore."
But who will take out the trash?
Not that everyone is convinced. Justin Cormack, for example, notes that "Someone needs to still fulfill the distro functions, like security updates, although it is a smaller scope."
It's actually security that may be CoreOS' biggest selling point, according to Polvi. As he told me over email, CoreOS' update model "is in complete contrast to Red Hat, which every three to five years ships a big update which becomes a major migration effort within every IT org that runs it. Our model is in complete contrast, [with] lots of tiny updates that are automatically applied."
Polvi borrowed the idea from the browser world he lived in for a time at Mozilla, where he saw web security dramatically improve once Google Chrome started to automatically update browsers. Could the same thing be done for servers? If so, he told me, "CoreOS will be the only way to effectively manage [server security] going forward."
Right or wrong, it's definitely a change to how IT has traditionally managed Linux. But that may not matter, as IT may not be driving this Linux bus anymore.
Objections to thin Linux
At this week's Gartner Symposium, Gartner noted that 38% of total IT spending now comes from outside IT, a number expected to explode to 50% by 2017. Or, as Redmonk analyst Donnie Berkholz finds, "Departmental budgets coming from marketing and from lines of business are leaving IT, and over the course of a few years, this will transition to subtractions directly from IT's budget."
While developers can't claim all the credit for this change -- so-called Shadow IT, of which developers play a key role, makes up much of this shift away from IT -- developers are the ones tasked by their lines of business to build applications regardless of IT inertia.
And developers are driving Linux to get thinner and thinner.
"There's nothing new under the sun"
Some suggest that CoreOS isn't necessarily novel, that "CoreOS is just another Linux distribution with a strong opinion," as New Relic system administrator Kelsey Hightower phrases it. Others, like Chuck Short, speculate that CoreOS is "just another way of doing just enough OS."
Still others, like Twitter's Chris Aniszczyk, think that, even if new and different, it would "not [be] that hard for someone like Fedora to clone what they did with a 'spin'." He has a point, but as Simon Wardley intimates, it wouldn't be very easy for the big Linux vendors to change their business models overnight.
And yet something must change.
We've grown up with the idea that Linux can be all things to all people, running everything from mobile devices to supercomputers. While there are distributions optimized for these different hardware platforms, increasingly we need special-purpose Linux distributions for specific services, which aren't necessarily "distributions" at all.
Linux, in short, becomes a very thin service for modern developers. CoreOS is worth watching.
