Linux's Future: Observations by a Linux user
Having been inspired by the Neowin original two-part article, "Microsoft's Future" by Max Majewski (part one and part two), I decided that a look at where Linux stands now, and what the future may hold for it.
In 1991, Linux was created with a announcement on a Minix Usenet group by Finnish student Linus Torvalds, stating that he was looking for feedback on a free operating system he was developing. A following started, and people began to submit code back for inclusion into the project. The hobby grew legs of its own and, over the years, turned into the powerhouse community that it is today. Large corporations rely on Linux. Businesses make money selling Linux services. Linux is starting to make an appearance almost everywhere.
A bird's eye view of the penguin.
Let's start by looking at where Linux is strong, and where it needs improvement. I am going to exclude boring technical aspects from this list. This is a list of things that affect marketability and suitability. While the advantages of ext4 over ext3 may be a topic some find interesting, if it reads/writes a FAT removable media, the world is happy and could not care less about ext filesystems.
- The GNU Public License (GPL) is second only to BSD-style license in giving developers freedom; it is first for keeping code free for all.
- Scalable from embedded http://www.linuxfordevices.com/c/a/News/Li...rnet-connector/ to supercomputing where it is hard to find any in the Top500 list that aren't running Linux.
- Customizibility. One word says a lot. It is not only designed to be modular, but the source code is freely available to be modified by anyone to meet their needs. Of the "big three" OSes out there, Linux has this one hands-down.
- Linux is not tied to any specific architecture. If a piece of hardware has a CPU, chances are it can run Linux. A manufacturer isn't going to be tied into a specific subset of hardware in making their product.
- The license. For all its strengths, the GPL can cause headaches for those uncertain of developer responsibilities. If a developer chooses to leverage and use GPL code in their projects, they can find themselves surprised that they have run afoul of the GPL terms. The FSF actively pursues those who violate the GPL to enforce their copyright, and requires them to change their code to remove the GPL violations or to comply with the license and release any GPL-licensed code they used and/or modified.
- An abundance of variety tends to confuse people who don't know what they are after, and what the different options provide. That is, most people new to Linux can find the choices overwhelming.
- No real "marketing" or "P.R." department. Red Hat speaks for Red Hat. The FSF speaks only of what concerns them. While IBM had aired some "Linux" commercials on television a few years ago, it is no longer creating ads that solely focus on Linux, as the "Prodigy" ad did.
- There is a lot of "trend following" in Linux. While there are good ideas and implementations in Linux, it was first created as a clone. To this day, it still attempts to implement features that exist elsewhere. As Eric S. Raymond (often known as 'esr', his initials) stated, it is about scratching that personal itch. Developers code what helps or interests them.
Where are the markets today?
Supercomputing is currently a solid Linux win. There aren't many competitors at all, and if you pick out 30 random entries, chances are they will all be Linux. It will take a supercomputing paradigm shift to knock Linux out of this niche market.
Embedded devices such as GPS units, routers and Digital TV Recorders often use Linux as their OS. This market is much more levelled, with Linux competing actively with Microsoft Windows, VxWorks and others. It is hard to find good reliable numbers, as it seems that different survey groups have different criteria they use when they classify something as "embedded" or not. As this market is generally populated with low-cost devices, licensing costs are an increasingly large percentage of OEM expense in making the product. Any OS that has a zero for cost and license tracking requirements holds an advantage here. Google's Chrome OS is a browser/OS platform that seems to be designed for a cloud or thin client environment. More on that a little later. For embedded devices, Linux is in this market to stay for the foreseeable future, even if it isn't the dominant player.
Mobile (communications/phone) devices are a specialized (and growing!) form of embedded devices. The LiMo Foundation formed in 2007 so various mobile phone OEMs could work together for a common goal. However, a dark horse has put the LiMo Foundation on shaky ground, and may push it to the irrelevancy of an also-ran. And that specter is Google with their Android platform. Also based on Linux, but from outside LiMo, it is surrounded by buzz, and seems to be growing at an impressive pace. The pizazz and polish that has been applied to the Linux GUI by Google seems to be just what the market is seeking. Google is allowing others to use the Android platform, while also releasing their own Google-branded phone using Android. This direction is potentially walking a line that could irritate the Android licensees, as was brought up by Microsoft. As long as Google isn't seen by the the other OEMs as taking unfair advantage of their position as developer of Android, I think that there should not be the dire fallout. Nokia may be warming up to Linux, but they are not yet abandoning their leading Symbian, with newest roadmaps including multi-touch and UI revamping.
In the server market, Linux has been strong. Yet, much of the gains here have been in takeovers from other Unix installations. Even several of Microsoft's now-defunct "Get the Facts" (renamed to "windowsserver/compare" ) case studies are not people switching, but shops that have already run Windows deciding to keep running Windows. Unix was the server leader. Microsoft is currently in the marketshare lead, but I see Linux continuing to erode the Unix market more than Windows does, and being a major contender in the market. It is so easily deployed by small and medium businesses, and it isn't something that can be tracked with certainty, due to the licensing freedoms to install "at will". The sales of Red Hat Linux subscriptions have risen, especially during the economic downturn, and their earnings beat expectations again.
Virtualization is becoming increasingly relevant. Once found only in the domain of high end server rooms, it is now becoming a common tool for home enthusiasts on desktop PCs. I think it is safe to say that the popularity of virtualization will continue to grow. The question is, who delivers? Well, both Linux and Windows do. And once again, the advantage that Linux has in licensing has caused Microsoft to relax their restrictions to make virtualizing Windows more appealing. Microsoft will not give up this sub-market easily. If Linux continues a strong presence, as I expect, this will drive a more intense competition.
When it comes to desktops the question is less "Should I use Linux?", and more, "What is Linux?". Until recently, Linux users had to install it themselves. A few PC OEMs toyed with Linux pre-installs, most of them making the pages hard to find. A disruptive force arrived recently with the advent of the "netbook". A diminutive PC with meager specs and a lower price tag. Like the "embedded" market, above, OS costs are a more significant portion of product cost. Linux took a strong lead, until Microsoft made their older XP product available at reduced pricing. Despite Steve Ballmer's spinning of netbook marketshare as by eliminating whole markets and distribution methods to claim a over 90% share, the truth is closer to 32% of netbooks have Linux. However, Windows is so entrenched into the OEM PC vendors, that even the Dell Mini page with the Ubuntu OS prominently displays "Dell recommends Windows 7" at the top of their page. Or it could be the discounts that Microsoft provides by contractually requiring them to display that message on their PC sales pages. This shows that the entry path for home Linux is rather steep. One could say that Linux desktop marketshare has doubled from 2005 to today, but the reality is that the doubling was from 0.5% to 1.0% of the web-browsing market (which is about the only real way to measure how much of a marketshare Linux may have, since it isn't acquired or licensed from any sort of centralized organization. Even if it doubles again in 5 years, that will only put it at 2%.
However, desktop computing, itself, may be changing. Computers aren't all "desktops" any more. They are mobile. Very mobile. And they are often taken mobile! So-called "cloud" computing is casting a few shadows on the traditional desktop paradigm. And, like shadows, they may be seen as somewhat lacking solid substance at this point. One can consider this a "what was old is new again" situation, as it is similar to the "thin client" of the past. The client computers aren't given the computing horsepower that is typical in a desktop, as the action is taking place remotely in the more powerful servers, and displayed on the clients. The current "cloud" implementation does give the clients more power, as they can run apps locally, and the cloud is more about data accessibility from multiple locations and PCs. The apps may be (but are not required to be) local on the client, but the data is remote. As an anecdote, my kids play online games. The games are typically Flash or Java based. They are often single-player, but some are MMO. When they have friends over, sometimes they use my PC. My PC is Linux. Yet, the kids that come over don't need training. They navigate Xubuntu's interface and open Firefox and go to their sites and the app (game) is served to the computer and runs locally. They log in and play with each other. The Flash and Java apps are small, and are executed locally, but the data is saved on the servers. The same concept is being used for photo sharing, photo editing, office applications, and more. These are the applications that comprise the "everyday use" for a large number of PC users, and here is where Google is going to try to enter and encourage market growth in this direction with their Chrome OS based devices. Accessing email through a browser is already common, and it shows how convenient it is to use the cloud for your data. Away from home and using someone else's computer? No problem! Just login. Get a new PC? Again, your data is as easily accessed from your new PC as it was from your old. And it doesn't get lost if your harddrive heads crash into the platter, peeling spirals of magnetic media off the surface. However, the cloud is still vulnerable to the data being lost from the server, as in the T-Mobile contact loss. Incidents like these should be more infrequent, if companies look at this as a learning experience. T-Mobile didn't have a good backup before making a system upgrade.
What has been unsaid.
Note that three items I did not bring up are cost, stability or security. While Linux is obtainable for free, that has never been the feature that the FSF promotes (though many users and individuals will bring this up as an important point in their opinion). In fact, the GPL specifically provides for charging for distributing and copying Linux, and calls it out as a protected right under the GPL.
Stability and security are two other items often promoted by various users. Indeed Linux is stable and secure. But so is a properly configured Windows or OSX install. And improper administration, or adding in buggy (beta or just poorly-written) code will compromise any OS. Quite frankly, I don't see these items as strengths to any particular OS at all.
Linux is strong in so many markets. Yet, the market most visible to the average consumer, the desktop, has eluded a Linux presence, for the most part. Is it because Linux is complex? No, not really. Using Linux is much like using Windows. My kids' friends have no problems using my PC. Administering Linux is quite a bit different, though. And that is what people are really saying when they talk about "using Linux is complex". But being different or new does not make it more complex. Alas, that is likely a topic for a different time and place.
If cloud computing becomes more popular, it will make the OS running a bit less relevant. Apps can run remotely. Some are delivered to the client from the server. These Java, Flash and even, to some extent, Silverlight web-based apps are cross-platform. This is a good environment for Linux.
Linux is strong and is safe from extinction.
This is a guest post by Neowin Veteran Mark Jensen