Evaluating Open Source Software for Your Business

by Dietrich Kappe Chief Technology Officer, Pathfinder Development

This month we will look at the strange world of open source IT decision making that can get in the way of good decisions. The exuberant IT manager sees open source like found money or a snow day; if it’s free and good, why bother with the evaluation? Let’s just use it. The cynical IT manager sees open source as a toy or a joke; you get what you pay for, so why bother looking at open source? Both of these attitudes miss the mark. When it comes to open source, you should evaluate it like any other software you are considering for your company.

If your company is like most, you probably have five criteria for evaluating software. First, is it solving the business problem for you? Second, is the software usable? In other words, can your employees use it error free and efficiently without extensive training? Third, is the software reliable? How often does it break down or produce unwanted results? Is it secure and bug free? Fourth, can the software scale and perform to required capacities? In other words, can it handle ten times as many users or transactions, or will it blow up or thrash at some point? Fifth, how supportable is the software? Can I get the developers I need to work on the software? Will the skill sets required for working with this software carry over to other products my company is using? Will new versions and bug fixes come out in a timely fashion? Will my preferred database be supported as an option in future versions and will my customizations carry forward to future versions?

Finally, after answering all of the above questions and deciding between the “nice to have” and “got to have” features of the products, you come to the question of cost. How much does it cost to buy the software, install, configure, customize, integrate, upgrade, train, support, etc.? Once you have the answer to all of these questions, you can make your decision: which software choice meets my needs while not breaking the bank? That sounds easier than it is. For instance, vendors can go out of business, dead-end a product, or change their licensing costs and terms. New developments in the software industry can leave you holding technology that is out of favor and hard to support.

Any software evaluation is based on best guesses on future technology and business developments, but open source software can provide answers to the above questions that take some of the cost and risk out of these decisions. First, open source software doesn’t cost anything. There are no licensing costs, no upgrade fees. Second, open source software has several advantages in the area of supportability. For one, the source code is available. That means that if you want a customization, bug fix or improvement, you don’t have to wait for a vendor to do it, you can do it yourself. Further, open source development decisions are generally based on a desire to see the software used as widely as possible, rather than on maximizing a vendor’s revenue or raising switching costs to lock in customers. That means that open source software is based on open standards and widely used and accepted technologies. This makes finding developers and consultants with the skill sets to support your product easy. In fact, many companies are making a business out of supporting open source software, from IBM and Red Hat to the many small consulting companies that support what they implement.

Keeping the above advantages in mind, let’s look at a sample of open source software products and see how they stack up against proprietary alternatives.

We mentioned in our last article that spending on Linux servers was the fastest increasing item in corporate IT budgets. That doesn’t seem surprising when we look at how a Linux server with Samba stacks up against a Windows 2000 file and print server. The products both have the same functionality; Linux with Samba is a drop in replacement for Windows file servers. Usability is a non-issue—workers use the file shares and printers in exactly the same way whether the server on the back end is a Win2k server or a Linux box. Independent tests have shown that Linux and Windows 2000 have comparable levels of reliability and uptime with a slight edge going to Linux. Linux and Samba, however, tend not to be vulnerable to the same worm attacks that periodically affect Windows servers. Performance of the OS is again comparable, as both Linux and Windows 2000 can run on the same hardware; in tests conducted by PC Magazine, however, Linux/Samba outperformed Windows 2000 on file serving tasks by about 100%. Linux has both plusses and minuses when it comes to supportability. On the minus side, Linux requires a different administrative skill set, but if you already have Unix administrators in house this shouldn’t be a problem. On the plus side, configuring and maintaining a Linux/Samba server is dead simple and requires very little time. Also, point-and-click tools like Webmin can simplify common administrative tasks to the point where Windows administrators can handle them without much training. As a result, the total cost of ownership for Linux/Samba is lower than for Windows 2000: no per server and per client licenses, no upgrade fees, fewer servers to service the same number of clients, and lower overall administration costs.

Open source products that can act as drop in replacements for proprietary products make the decision making process easier. All things being equal, which one has a lower total cost of ownership. Other open source products that can act as drop in replacements are JBoss, a Java application server that replaces BEA WebLogic or IBM Websphere. In the area of databases, mysql is a free replacement for SQL Server and Oracle in low end applications. In fact, many software vendors now support mysql in their products. With transaction support and replication currently in beta, mysql promises to challenge the established players in higher end applications as well.

For other products, the analysis is a little bit more complicated. Often the tradeoff is 80% of the functionality at 0% of the cost. For example, in the Web content management space, OpenCMS, an open source product, competes against Microsoft’s Content Management Server, and to some extent against Vignette’s Story Server and Interwoven’s TeamSite. OpenCMS doesn’t offer some of the bells and whistles of the higher end systems, like behavioral tracking, but it does a creditable job with basic content management tasks, such as publishing, workflow and access control. When compared against other mid level systems, like Microsoft’s, it comes out quite well on the functionality front. Usability is also quite high. Ironically, OpenCMS makes use of some of Microsoft’s own ActiveX controls built into Internet Explorer to deliver a very polished interface. OpenCMS is implemented in Java, using J2EE, and gains the performance, scalability, reliability and supportability advantages of that platform. Finally, on the subject of cost, OpenCMS is free, while a typical Microsoft CMS implementation will run around $200,000 or more in licensing costs.

We’ve looked at just a handful of enterprise ready open source products out of the hundreds available, but the same principles can be applied in deciding whether those other products are right for your business. We hope we’ve given you some food for thought. Choosing open source software is neither a no-brainer nor a deal with the devil. Instead, it is a decision that should be based on business needs and sound business judgment.