This is an exciting time for enterprise software. Never before has there been so much flexibility when choosing enterprise software. However, we also see a convergence in the market between commercial and open source solutions. Both models have their place in today’s world, but open source is intensifying the business environment commercial vendors are facing by increasing potential competition and accelerating innovation.
The rising relevance of open source software results from more than its zero-cost nature. Larger businesses also gain nonmonetary but valuable benefits such as source code control and modifiability. The market’s acceptance of the open source business model, where companies pay for support instead of the code, pushes the market to focus on open source product development and requires respectable service levels from supporting companies. The answers to most problems related to open source may be freely available through community forums, but many businesses that do not have in-house expertise will want some experts on tap, especially for mission-critical applications such as business intelligence (BI).
Early Linux and open source marketing focused on the total cost of ownership (TCO) and security arguments, but the increased focus on these claims proves that benefits are actually deployment specific. Some companies find the early availability of patches through open source more valuable than the much more measured release cycles of a commercial vendor. Conversely, other companies need some guarantee of stability and supportability, even if it exposes them to software vulnerabilities for an extended period of time.
Deciding whether to implement open source is not a single dimensional decision. Should a company move from a proprietary to an open source operating system? How much risk does moving to an open source database present? Will the application tier be more flexible if you have the source code, or will a mature commercial software development kit (SDK) suffice?
Business Intelligence Landscape
BI is one of the fastest growing enterprise software segments, and the number of options available to a company is exploding. With the increase in both commercial and open source business models, most prospective customers do not have a strategy to evaluate them. Perhaps not surprisingly, many buyers do not consider their own deployment strategy or future roadmaps but rely on much more subjective information.
Commercial BI companies have been making a move toward less proprietary technologies such as Linux, Apache, and Java, which provide more platform freedom than more integrated or monolithic technologies. Due to the high level of development and customization many customers make to their commercial BI deployments, the open source developer market has made a significant impact on technology roadmaps. While monolithic architectures such as .NET are still prevalent in the enterprise, the rise of Java application servers is a testament to the desire for open, standards-based architectures.
Commercial or Open Source: The Customer Still Wins
Customers ultimately benefit from this convergence of historically opposed business models. Good open source eventually commoditizes technology by making it freely available. Commercial companies are understandably worried about their present business models because their newfound competitors are giving away a product for free. There is a strong and consistent pressure for commercial companies to continually innovate and provide greater value, either through a better product, more responsive support, or superior customization services. Complacent companies will quickly find their value propositions vanish, along with their customers.
However, there is also an equally strong and consistent pressure on open source to match the enterprise-class requirements of scalability, reliability, and performance. Commercial software vendors provide the “one throat to choke” that customers demand and are on the hook if their product fails. The money that customers pay software vendors not only funds the development of the product but also enables an “insurance policy” in the form of product support behind it. Successful software vendors invest as many, if not more, resources on support and R&D to ensure their products meet the complex technical requirements set by their customers. Unsuccessful companies are simply “choked” right out of the market.
This convergence of the open source and commercial BI markets has accelerated the state of enterprise software as we know it. The sheer number of developers in the open source community almost guarantees a stable flow of innovative ideas, and the nature of the community itself provides a mechanism to have the best ideas float to the top. These ideas are freely incorporated in open source projects and provide a large sample set from which to find the best. Finally, the end product is freely available for those with a sense of adventure, or at a relatively low cost for those who want commercial support.
The benefits of open source also ensure that commercial companies must continuously innovate to stay one step ahead and provide value on a product-to-product basis. Those customers with more moderate support needs must still find commercial software attractive on its own. Because successful commercial companies have sizeable installed bases, they have not only a closer feedback loop to translate customer needs into product features, but also the considerable resources of alliance partners that can exponentially improve the TCO of a solution through better product integration. This last point is important because the value of an enterprise solution is directly related to the value chain supporting it.
That is not to say commercial companies cannot also produce open source or participate in the open source community – in fact; there are strong value propositions for doing precisely that. In the BI market, we are already seeing many companies enter the community either as a participant or a contributor. The rising popularity of the Eclipse framework for report development is one example. We need only look at what the Eclipse framework has done to the developer integrated development environment (IDE) market: its market dominance has prompted Borland to move from its own JBuilder IDE to a value-add commercial component within Eclipse to adapt their business model to changing market dynamics. Eclipse’s founder, IBM, bundles six value-add components, including reporting components, to the Eclipse framework and offers the integration as the IBM RAD environment.
Understanding Your Options
Back to the goal for many BI implementations: “Improve decision-making within the enterprise by providing access to relevant data that is easily visualized and is presented in a timely fashion.” Note that while a goal may include, “must be implemented on Linux” as a constraint, this refers to the operating system and not what we typically mean by “open source BI.”
By now, it should be clear that, just as in life, there is no silver bullet to choose between commercial and open source BI. We cannot merely look at the ongoing Linux versus Windows war, as it is based on theoretical cost advantages, conflicting security claims, and artificial benchmarks. We also cannot only look at feature claims between them because not even two commercial BI offerings can be compared apples to apples.
To illustrate a typical set of options in a more straightforward example, let’s consider another goal you might have: “Take a vacation in Costa Rica and spend some time at the beach.” You have several options on how to reach this goal, all of which have different trade-offs in cost, risk, and customizability:
You could contact a travel agent and book the entire vacation by merely specifying that you want to go for seven days, only stay in four-star hotels and want to spend at least three of those days on the beach. In this case, you are paying a knowledgeable person who has the objective of using their own experience to ensure you have the best possible vacation with as little work as possible. The travel agent also takes responsibility to help you if something goes wrong with your itinerary, such as if your reservations are not honored at a hotel. In the event of a disaster on your trip, you also have a 24×7 emergency number to reach your agent who can change your itinerary or even arrange to have you return early.
Alternatively, you could contact a travel company that specializes in charter vacation packages for the specific location you want to visit. The agent will provide an itinerary that includes your flights, hotels, and transfers, and while you may get an excellent deal, there will be little to no flexibility in making changes or special requests. The agent may have little to no power to help you if you need to make a change or have special needs.
Finally, you could go to an online travel site and create the whole trip yourself. Here you pick your own flights, hotels, and car rentals, but you need to do a lot of research on the destination to decide what you want to do when you get there. Expedia’s search engine will do its best to recommend excursions and find you the best rate, but it is no substitute for an experienced travel agent who knows the ins and outs to make your trip truly special. While this gives you the maximum flexibility in choice, you take all responsibility yourself, and you must expend considerable time and energy learning about your destination. By the end, you may be able to plan a completely unique vacation, but you will have spent the time that could have been used doing more important things such as working or spending time with family.
Implications on Your Business
We are all very familiar with each of these three models. The travel agent route may be initially more expensive, but it is a known cost and has a guaranteed result. You have help and some form of recourse if something goes wrong, and you can focus on having fun instead of planning mundane things such as which hotel to stay at. Companies that want to spend their resources on their core business may want to consider this option to ensure their mission-critical BI solution is optimal so as to increase their operational efficiency instead of expending resources on the implementation itself. If the return on investment (ROI) of an optimal BI solution is higher than the cost of the commercial solution, this approach provides the best value as it enables a company to focus on its core business, which is most likely not BI implementation.
The charter company or hybrid/bundled open source approach increases the project’s risk at perhaps an initial cost savings, but you may find the ROI calculation lower than expected because of the added risk to the project and, ultimately, the company. However, this is still a very viable solution for companies that have shorter-term goals and have the maneuverability to take evasive action if their BI needs change.
Finally, the do-it-yourself or best-of-breed open source approach provides the ultimate in flexibility, but you must really understand how the pieces fit together. There is also an ongoing cost of maintenance because there are inherent risks over the life of the project. The additional cost of resources to do this must also be weighed in the ROI calculation against the cost saved by integrating your own pieces. A few open source companies are now trying to bridge this gap by pre-integrating or gluing other open source projects together and providing support on top of the resulting suite. Decision-makers would be wise to ensure that the vendor they choose is a right integrator and not masquerading as a hybrid/bundling company.
Making the Choice
While there is not a right answer that suits everybody, there is a right answer for your business. Understanding the real monetary value of your BI solution is the first step. What are your goals and the size of the return you are expecting from this implementation? Removing the cumulative annual cost of software, maintenance, and labor is the next step. A solution’s value quickly decreases if the costs to support it (whether by an ongoing support contract or required in-house technical personnel) are very high or unpredictable. Third, understanding the long-term roadmap of your potential solution and how clearly it maps to your own long-term goals are necessary to ensure you actually meet your target.
Believe it or not, this decision is actually getting easier. With more commercial companies joining the open source movement, either by contributing code or by providing better tools and APIs for enterprise developers, we will start seeing increased convergence between open source and commercially available BI. Determining which solution meets your strategic goals will then become an exercise in finding the best tool for the job (improving your bottom line) instead of whether you should choose a commercial or open source solution (which is not a business decision anyway).
It will be interesting to see how this phenomenon of convergence manifests itself in the future. Open source has changed the BI landscape forever, and we are likely to see more companies follow in IBM’s and Borland’s footsteps by moving to a more mature business model that leverages the open source community while still offering the high-value commercial software provides. You, the customer, will ultimately benefit, but it is dependent on you making decisions focused on long-term company goals and not on the specific technology or software model used to achieve them.