BrainTrust Query: Is Open Source Worth It?
By Bill Bittner, President, BWH Consulting
I’ve been working on an Open Source based software project. It’s really caused me to think through the whole process that brings a company to decide on an Open Source solution vs. buying a commercially developed (and supported) software product.
My first exposure to Open Source software was about eleven years ago when my AIX systems administrator introduced us to SAMBA. SAMBA provided us a direct interface to the Windows file systems on our (wrong choice here) token ring network. It enabled us to transfer data between the AIX operating system used on our In-Store Processors to our Windows based PC environment for version control and storage “seamlessly”. There were commercial solutions, but my system administrator was adamant and the cost was literally “free”. This first experience was relatively painless and SAMBA provided all the capability we needed.
The general public has now become aware of Open Source as discussions of Linux, Firefox and other independently developed software solutions have reached the general press. But the pattern these solutions have followed is that the “core” becomes the “queen bee”. The core software sits at the center of commercially developed extensions and support services that meet various user requirements. Thus Red Hat is a commercial software vendor that has developed extensions and supports Linux, an Open Source platform.
It is very difficult to find a “totally free” Open Source solution. Either you’re going to buy extensions to the solution or use your own staff to take off the rough edges. If nothing else, someone on your staff is going to take the time to research and educate themselves on the alternatives that exist.
The core solution is like buying a furniture kit that needs to have the pieces assembled and the surfaces finished in order to create a product. If you have the carpenters and woodworkers on staff to complete the project, you may be able to support an Open Source solution internally. Otherwise, you might hire someone (did I hear “consultant”?) who has done the assembly before and can get you past the initial installation. The challenge becomes finding the ongoing support. If it’s like a piece of furniture and the functions are clearly understood and not much adaptation is involved, then ongoing support should not be much of an issue. If it’s like an electronic device that connects with many parts of the business, like a purchasing application, then it needs to accommodate change as requirements and connections evolve. This makes support and user training critical factors.
Discussion Questions: What’s the best long-term strategy? Open Source, with its low up-front costs but inherent lack of support, or commercially developed, fully supported solutions? Can you think of situations where one is better suited than the other?
[Author’s comment:] Of course, it depends on the situation and the available solutions. If it is a narrow-scope, technically well-defined environment like the SAMBA example, Open Source can make a lot of sense. It does not involve significant user training and the ongoing support issues will likely be addressed by the technical community. When it comes to application software, the answer may be completely different. Now you are dealing with solutions that have a lot of options and require significant user training and support. It may be better to have vendor support and use their experience for training and implementation to help with the inevitable resistance to change associated with a new application.
Often, companies end up buying a commercial solution and then customizing it so much that they can’t take further updates and the vendor can no longer support it. In that case, they may have been better off with the Open Source solution all along.
We’ve come a long way from the “no one in IT ever got fired for recommending IBM” mentality but I think there are more opportunities for considering Open Source in business solutions.
We don’t talk often enough about the five things that matter most in measuring the effectiveness of software:
1. Performance
2. Availabilty
3. Adaptability
4. Cost of ownership
5. User mastery
The evaluation of Open Source should be against these five measures, not the cost savings against commercial alternatives
The Open Source alternative usually requires in-house experts necessary to ensure satisfactory performance, availability and adaptability. In my experience this expertise is difficult to sustain over productive life of a software solution. Software engineers love to be involved when it is new. But they’ll jump to the next thing (or job) when the project gets running. For that reason, most people think Open Source is too risky for an in-house project.
More recently more commercial software, services and consulting companies have embraced Open Source. They are building sustainable products that mitigate the risk.
Both approaches add significantly to both the start up and ongoing costs of Open Source projects.
Once your performance, availability, adaptability at a reasonable cost, then you can get to the real challenge: are the end users getting full value out of the investment? In my experience, that’s the area where we have much work to do.
Last I checked, no software produces benefit within itself. For benefits, you need users to master the software.
The key here is to recognize what open source (and other copy left based solutions offer you). In order to properly do this, you must first attribute a dollar cost to your TCO (total cost of ownership), but most importantly you must recognize the implementation cost for the entire system. This includes the full life cycle implementation and all of the key components that are involved in this including:
1. problem, opportunity, and objective identification
2.Information requirements determination or requirements gathering
3.system needs analysis
4. system design
5. system development and documentation
6. system testing and maintenance
7. Implementation and Evaluation
A true, and accurate system implementation requires many resources, as well as continual checking and measuring of key hurdles (after they have been identified). Costs frequently escalate, and anytime you are working in a customized or semi-customized solution, these costs are even higher. Don’t forget that the most supported, stable and best written for piece of software for the enterprise environment is Linux, and it is completely open-source!
Certain technologies are more conducive to open source development than others. As Hannaford Bros. and others have shown, store-level systems like POS and LP are typically receptive to open source. Headquarters applications, many of which need to be customized, may not be as amenable. I know that’s a sweeping generalization, and that there are many exceptions, but the division may fit nicely for practitioners who often separate responsibilities along these lines.
Open source offers the retailer/customer lots of alternatives. This article talks about open source in generic terms. Generically speaking, yes you should do a lot of due diligence and look at long term support, knowledge, functionality enhancements, etc. But wouldn’t you do that due diligence on any system?
Being in the software selection business for years, this question comes up quite frequently. The benefits of open source are huge and as in commercially packaged software systems, there are good choices and bad choices. It’s knowing the difference and finding the best fit for your infrastructure. Many commercial packaged solutions charge so much for on-going maintenance (once you’re hooked) that this option is a no brainer. The best long-term strategy in my opinion is to map your requirements short term and long term and evaluate the vendors and offerings that are in the marketplace with those requirements. Technology is changing so very quickly it is essential you look underneath the covers so to speak. Many top software providers are still running on very old programming language and platforms. This handcuffs you as the technology is not flexible or user friendly as the more advanced software that is just coming on the market. I believe you’ll see the world changing when it comes to selections. No longer does the older, stable software get the prize but the newer more innovative, more flexible will win! (Much like in the olden days, having 20 years with a company was a good thing on a resume, today having more than 5 years experience people say, what’s wrong with you? You’re stale!)
Your own virus protection software is a good example. When having all sorts of problems with one of the top two software packages, I called the help desk. It seems this package doesn’t play well with other software and causes some inherent problems in each instance. Calling late at night and getting to know the technician for a couple of hours on the phone, I asked what do you use on your personal computer? The answer might surprise you, he uses a free package available on the internet and he said it’s the best on the market! Hmmmm…………
Well-run systems departments use total life cycle costing. Before buying any new system, people total the cost of the software, maintenance, custom modifications, updates, installation and configuration, and daily operations, based upon the expected years of use. If the IT folks expect to use a system for 7 years, they can estimate the cumulative 7 year cost. The open source software might be free or low-cost, but the total life cycle cost might be higher than software as a service (SaaS), or ASP solutions or higher-cost proprietary software licenses. There’s no consistent conclusion because each installation has different economics.
I think Bill answered his own question. When it comes to direct application to IT problems, the decision to opt for Open Sourcing does indeed “depend on the situation and the available solutions.” In and of itself, the idea of Open Sourcing is a powerful one that shouldn’t necessarily be confined to IT. There are several examples in my second book (The Deviant’s Advantage) of non-IT Open Source solution sets. That said, the idea that “all of us are smarter than some of us” isn’t always necessarily true and Open Sourcing can, in fact, lead you down some blind alleys on occasion. Again, as Bill correctly points out, “it all depends.”