The growing availability of open source offerings, covering all aspects of business intelligence from ETL solutions through to data mining, is beginning to fundamentally change the way that business intelligence projects are costed, approved and evaluated. This article provides an overview of how the evaluation and implementation is being changed by open source BI offerings.
How Does Open Source Differ From Commercial?
Basically, open source is available free to download and use by anyone. However, open source usually comes as a community-supported offering: getting support quickly, or indemnification of the product's intellectual property cannot be had. There are a number of organizations that publish open source software and offer a subscription model to provide organizations with guaranteed service levels and indemnification, usually along with some additional features and functionality, such as automated software updates and free additional utilities not available to the community.
Typically, under the terms of an open source license, incorporation of an open source product into a developer's solution will result in it being considered open source as well, unless a separate OEM agreement is obtained from the open source software publisher.
Since open source software (OSS) vendors don't employ a sales force and significant marketing effort - relying instead on word of mouth, user forums and internet networking to "spread the word" - they don't need to charge an upfront licensing fee (most of the initial fee paid for a commercial license goes to fund the sales and marketing of the product).
...And Why Is This Important To Business Intelligence Applications?
The path to getting an effective business intelligence solution up and running can be a difficult one. BI projects are designed to get information into the hands of business decision makers so that they may guide the ship, so to speak. Typically, their value is measured by the improvements brought to the business as a result of better decision making. If the data collected isn't what is needed (a rare occurrence these days, but one that does happen occasionally), or is presented in a way that the business user can't use the data, the solution is discarded or left to wither and die.
To avoid this scenario, a lot of time and effort is spent to confirm requirements in order to justify the typically very high cost of databases, ETL tools and reporting and analytic offerings - not to discount the effort to design and build the solution!
Open source solutions offer a number of benefits to organizations looking at new business intelligence applications:
- They are free to download and install
The solutions are available immediately, allowing a development or assessment team to bypass an often cumbersome and (for some) annoying sales experience, the effort of obtaining trial licenses, and renewing them if they expire. Software can be installed and downloaded without involving the software organization and decent support can often be had on the software forums.
- They work well with other components
Since, by definition, open source provides the source code to all, one typically finds a great deal of support for complementary products. For example, Infobright, a column-based database designed for analytic queries, offers very good integration to Pentaho and Jasper Reports, the two leading open source reporting and analytics suites. As well, virtually all open source solutions will work with the existing commercial offerings.
Typically OSS publishers are able to develop integration into other OSS offering in short time frames: Case in point, an open source vendor spent well over one year trying to develop and test an API with a commercial software vendor - all of this time was spent in negotiating and obtaining access to the required code base and none was actually spent developing an interface! Subsequently, the same OSS vendor was able to develop an integration to a similar OSS offering in under six weeks (and this is fully tested and certified).
OSS developers focus on ensuring that their products support and work with any published standards. Since they don't really have a vested interest in keeping their product in use via "software leashes," OSS vendors can focus on fully supporting published standards.
- All significant offerings provide support and indemnification services for support of production systems
OSS vendors typically make their money by providing post-implementation services, usually via consulting services and support agreements. All significant offerings - be they DBMS solutions from Infobright, Sun/MySQL or PostgreSQL, ETL solutions from Pentaho and Talend, or BI reporting tools from SQL Power Group, Pentaho and Jaspersoft - offer support and indemnification services consistent with commercial vendors.
- OSS offerings support abest-of-breed BI strategy
Because OSS offerings typically focus on one aspect of the BI architecture, combined with their high interoperability with other products, they are ideal for supporting a best-of-breed strategy. Organizations are often "roped in" to selecting a single vendor's suite of products, since this offers the lowest total cost of ownership (commercial software vendors typically offer heavy discounts on the upfront licenses, but will base their ongoing support fee on the list price). OSS vendors only charge for ongoing support and indemnification and often at rates equivalent to or slightly lower than the fees charged by commercial vendors.
- OSS BI vendors offer functionality equivalent to commercial offerings
The major OSS BI vendors offer functionality that is equivalent to or exceeds that offered by commercial vendors and, in many cases, address niche functional requirements that the mainstream vendors cannot. Jaspersoft, Talend, SQL Power Group, Infobright, Pentaho, MySQL, EnterpriseDB and others among them offer all of the components that are required for a BI solution and have been offering these components for five or more years now.
Open source is a viable alternative to commercial software business intelligence offerings. The rapid availability of software for download and implementation, combined with interoperability of OSS with other software offerings, provides organizations with a platform for the rapid prototyping and implementation of a business intelligence solution. The provision of support by established companies ensures software viability. And, possibly most important, the "try before you buy" approach with OSS allows organizations to utilize whatever tools they need, discarding those that don't work, without risking significant upfront investments in license fees.
Open source software is changing the approach to evaluating and selecting BI software, pushing the initial vendor interaction until well after solution selection, and often only at the point that a solution is being moved into production. The flexibility of OSS will allow BI users to more readily select the software that best suits them, and to discard and replace software easily as the business needs change.
About the Author
John Kemp is a Principal Consultant with SQL Power Group, one of Canada's leading open source business intelligence implementators, and the publisher of Wabit, an open source reporting and analysis tool. John works with organizations who are selecting and implementing open source BI solutions. Prior to joining SQL Power Group, John was part of the management team that transitioned Infobright's flagship column-based intelligent database for analytics from a commercial model to an open source one.