Corporate Resistance to the Adoption of Linux and Open Source Applications
So just what is it about Linux and Open Source that company owners and executives don’t like? Heavy hitters like IBM, Sun, Oracle, Dell and HP don’t seem to be making huge strides toward the adoption of Linux. OpenOffice.org, MySQL, PostgreSQL, and a myriad of other applications have not pushed Open Source programs or Linux, in large numbers, to the world’s Desktops or Servers. I don’t get it.
Why do people inherently fear something for free? Is it because we have been taught that you don’t get something for nothing? Or is it deeper than that? Do we feel that by using Microsoft’s Operating Systems and applications that we are somehow better off? Have you ever tried to call Microsoft for assistance? Did you get helped for free? There are many vendors, not just Microsoft, that sell applications, and then charge you when you need help with them. Even some Open Source ones do. This seems wrong to me.
Free enterprise is a good thing, but shouldn’t users of your software be given at least a little bit of good will? I think if a large user base is going to purchase and use your software, that you are obligated to give some free support. It should be limited in some way so that you don’t burn up all of your resources assisting people who are too lazy to read instructions, manuals or online help. It seems that most people would rather pay high prices for software and unreasonably high rates for support than to opt for free software and free support.
I have yet to be disappointed by the online support communities that exist to help with issues with software that they themselves produce. Many times I get a personalized answer right from the developer of the application. And what if I need some extra button or function in Microsoft Word? I can suggest it and it may come out in a future release that I have to pay for or I can ask the developer of an Open Source application if they can add the widget for me, or guess what? I can add it myself because I have the source code for the application.
The excuse I used to hear was “we don’t want anything proprietary” or “we don’t want to have to try to get support if this person gets hit by a bus.” What is more proprietary than anything Microsoft? The bus thing is also a moot point. If one has the source code, then there are other people in the world, perhaps in the town I live in, that could support the application.
These are the major barriers to Linux and Open Source adoption that I have observed:
· Lack of a Responsible Party
· Support
· Abandonment
· Longevity
· Stability
· Training
· Certification
· Multiple Distributions
· Hidden Costs
· Compatibility
· Licensing
Companies always want a responsible party to point the finger at when things go wrong. Linux and many Open Source applications don’t have a responsible party or a legal entity associated with them that can be forced into litigation. Without a legal arm to twist, many companies shy away from such blameless entities. Redhat and Novell have attempted to change this over the past few years by offering a company that backs its products. This approach has only captured a few more converts.
One reason, I believe, is that the costs are prohibitive compared to those of Microsoft. For instance, on Novell’s website, the cost for SUSE LINUX Enterprise Server – 1 server, up to 2 CPUs, per year is $349. Redhat ranges from $349 for Redhat Enterprise Linux ES (Basic Editions) all the way up to $2,499 for Redhat Enterprise Linux AS (Premium Editions). If you need the ability to call someone for help, and have someone to blame, then perhaps one of these options will appeal to you.I have often seen the excuse for not supporting or using Open Source applications or Linux stated, in the form of a question, as: “What happens when I can’t get support?” My answer is always “you can” but this is rarely convincing. I explain about user communities, Google, Sourceforge, and others but all I get is a blank stare. Reliable, reasonably priced support is very important to corporations, large and small. The opinion is that if it isn’t a mainstream application then you can’t get reasonably priced local support. The actual fact is that, while this may have been true 5 or more years ago, it certainly isn’t true now.
I have also heard the issue of abandonment when it comes to Open Source applications. I suppose this is legitimate since people get attached to applications that end up getting discontinued for one reason or another. The nice thing about Open Source applications is that, with the source code, someone else can pick up the project and continue it indefinitely. Chances are that, if you are hooked on an application, others are too. Closed source applications either have to be released to the public domain (rare) or someone else has to purchase the application’s source code and rights to reproduce it. For example, there is an application that is a great Unix/Linux application that was abandoned by a commercial vendor a few years ago. I called them and begged them to release the application to the Open Source community or to give it to me so that I could update it and keep the software up to date. They told me that “we are considering many different options at this time.” I never heard about it again after that. The software has not been released to any other company, the Open Source community or to the public domain. This was a commercial software package that has been abandoned.
Longevity of Linux and Open Source applications is of concern to some potential adopters. They feel as if Linux and Open Source are fads that will soon be part of a “Remember When” column in some obscure nerd journal. I personally discovered Linux in 1995 and have proudly remained part of the “fad” for over 12 years now. I don’t know many fads that last for 12 years. If longevity is any indication of how viable an Operating System is, then I would say that Linux has passed the test. Not only is it still around, but it has grown from a single project to spawning many new companies such as Redhat, SuSE (now part of Novell), VA Linux, and many others. That single project has also sparked the creation of hundreds of custom distributions and solutions from embedded servers to Linux on mainframes. Linux has also put public domain and Open Source software into the mainstream.
You would be hard-pressed to find a person in technology that hasn’t heard of Linux and Open Source. Linux is here to stay. From a business point-of-view, stability and scalability are still questions that need to be answered before Linux is widely accepted into Data Centers alongside of Solaris, HP, and AIX big iron. The much anticipated 2.6 kernel has been a boon to big iron Linux hopefuls. The 2.6 kernel is SMP ready and there have been benchmark tests on servers that have up to 64 processors and 256GB RAM. I would have to humbly say that this is considered scalable and SMP servers that support 16, 32, and 64 CPUs are known as “big iron” in the Data Center.When total cost of ownership (TCO) discussions surface, they usually include the cost of training. I have seen reports that state that Linux does not lower TCO due to the costs of training technicians and engineers in this new operating environment. This may be true if your technicians and engineers are not Unix literate. Linux is Unix. If someone literate in Solaris connects to an HP Unix server, she will still be able to issue most common Unix commands and will need no OS-specific training. The major differences are the administrative tools between vendors as is the case with Linux versus Unix. Training needs will be fairly minimal unless your support team is now composed of Windows gurus. And Linux is not the culprit in the need for training. If you switched to any version of Unix, the need for training would exist.Amongst the anti-Linux ranks, there are those who complain about the myriad distributions of Linux that are available. They state this as a negative against the adoption of the Operating System. I feel that the ability to create and use Linux for very specific purposes is an extreme positive that should speed, not slow, its adoption. The fact that I can use an entire fully-functional Operating System from a CDROM is highly desirable. And if I don’t find a distribution that works for me, I can create my own that will.
Professional certification is another area of concern for many would-be adopters of Linux and Open Source. There are now multiple avenues for certification for technicians and engineers who support Linux. Redhat has the most visible certifications with RHCT (Technician), RHCE (Engineer), and RHCA (Architect). The Linux Professional Institute (LPI) also offers a certification is considered “distribution neutral” and is well-respected in the community. CompTIA offers the Linux+ certification which is also vendor independent. There are plenty of Linux certifications available for all levels of support personnel. CompTIA is well-known for it’s A+, Network+, and Security+ certification tracks and therefore is considered to be industry standard. Redhat’s certification, though vendor and distribution specific is probably the best known and most highly sought after in the industry. LPI doesn’t have the name recognition or the long history of the other two but I believe that it is just as viable and valuable as any other for Linux certification.It is certainly true that Linux is available in just about any flavor or configuration you could possibly want or need, but that is a good thing. The fact that Linux can be modified freely to meet specific needs or any purpose makes it unique, flexible and extremely valuable to the worldwide IT community. It is very cost-effective to have a firewall set up on an old computer with a small disk, floppy disk, or CDROM, compared to purchasing a full commercial Operating System, a commercial firewall package, and still have to worry about access licenses for the server and the firewall. No thank you. Freedom and flexibility are good things. Having many, in fact hundreds, of different possibilities for an operating system is a positive attribute. Opposition to Open Source adoption will always speak of “hidden” costs associated with Linux and other allegedly “free” software. They state that while the Operating System (or application) itself is free, there are the lengthy setup/installation times, the possibly costly/commercial dependencies for the “free” software, training issues, and of course the development costs to work the application or system into the current infrastructure. And now there is the popular argument about who owns the source code to Linux and other Open Source applications. I think there is probably some truth to the statement that some setup/installations can be lengthy for Open Source applications, but certainly that is not the norm. I have never found a commercial dependency for any application that I have ever installed so I am not sure where that argument originated. Perhaps some of you readers have war stories to share about such things. There may also be some reality to the statement that working new applications or Operating Systems into current infrastructure can require some development resources, but many of the tools are web-based and therefore are fully platform independent. My argument is that, while there may be some extra time spent up front on development and configuration, the stability and reliability of the applications and Linux will offset those costs. In my opinion, downtime is far more expensive than development because once an application has been deployed; dependencies have been built on them. Is it more cost-effective for a few IT staff to be busy on setup and deployment or to have an entire company, a department or your clients experience downtime?
The issue of compatibility is often cited as the great divide in adopting anything new, not just Linux. It seems that whatever is currently owned or used is what the adoptive application has to be compatible with. When my clients were converting from WordPerfect 5.x to Microsoft Word, Word had to be compatible with WordPerfect or the conversion was not going to happen. Many people in those companies resisted and complained about the conversion, but eventually relented. The main point here is that compatibility is only as much of an issue as you want to make it. Long-term productivity, TCO and the freedom to enhance and customize an application is far more valuable than the costs and pains of conversion. The clients, by the way, couldn’t believe how much more they could do once they had converted to Microsoft Word. And they also couldn’t believe that they held on so long to the old application. People invariably resist change. As an IT professional, either internal or external to a company, you have to make the transition as painless as possible.
And finally, my favorite topic to discuss: Licensing. When a company has a licensing policy so complicated that even the employees at that company have problems understanding it; there is a problem. Licensing also probably shouldn’t have version numbers. It doesn’t bother me necessarily to pay for software, but what bothers me about paying for software is that I have to keep paying and paying for the same software over and over again. If I purchase server software from a company, I think it is a rip-off to have to then pay for access licenses. Does this mean that if I don’t buy access licenses that I can install it but I can’t use the server except at the console? That really isn’t much of a server now is it? My definition of a server is defined as a multi-user, network-accessible computer. To call an Operating System, a server, is to assume that it will be accessed by more than one person sitting at the console. If everyone agrees that servers are multi-user Operating Systems, why should I pay for licenses to access it when accessing it over a network is the sole purpose of the server? In the interest of everyone being able to make a buck, I think that server software (in the event the vendor feels compelled to charge for access licenses) should come with a number of “free” client licenses to ease the cost of technology to small companies. A server OS and 10 client access licenses would be a nice start. It is a serious financial burden for small companies to purchase server software, access licenses that can easily surpass the cost of the server software, server hardware, and desktop computers with Operating Systems. If one considers the costs for server software, access licenses, hardware, desktop software and applications, it is surprising that small companies aren’t converting by the hundreds to Linux and no cost Open Source software.
