The number and scope of business intelligence (BI) applications that leverage open source BI components is increasing slowly but steadily. The promise of a lower total cost of ownership, due in large part to the relative absence of expensive licensing and maintenance/support fees, has been the primary instigator for companies to forge ahead with their open source BI projects.
With open source BI functionality catching up to that of commercial BI software in many areas, the attractiveness of open source BI has increased markedly. Having a non-fee-based platform that can be used for analyzing, transforming, integrating and manipulating BI data—and to build and code data visualization and KPI dashboards—holds unlimited attractiveness for today’s cash-strapped IT departments. The entire corporate intelligence lifecycle, from analysis to development is becoming more socialist in nature; with both BI source code and object code made freely available to licensees for their customization, the cost of entry into establishing an internal business intelligence practice is reduced drastically.
In contrast to open source practices, commercial BI vendors make their object code available to their customers but heavily shelter and protect their source code from alteration or reverse engineering--both illegal practices under existing global intellectual property laws. This lack of transparency subjects their clients to expensive fees for software, maintenance and consulting services from “experts.” Yet on closer examination, open source BI is not necessarily the ultimate panacea for reducing BI costs, regardless of what some industry experts would have us believe. Not being coupled to a specific BI vendor comes with number of risks, despite the front-loaded cost savings and freedom from proprietary BI platforms. Perhaps the biggest tradeoff is that while the open source BI model greatly reduces software costs for businesses, the traditional models of software support and maintenance that they have grown accustomed to with commercial software providers will not exist.
Until recently, much of the open source BI software on the market evolved from a specific commercial endeavor, where a corporation developed a solution to meet one of their specific business needs or a need in a specific market sector. However, open source BI software has expanded from being a stop-gap effort to a much more strategic endeavor, with increased visibility and stakeholders. With this in mind, companies starting their first foray into open source BI should be looking for products that offer support for the entire BI lifecycle—from initial sourcing to final executive analysis, with web-based and mobile reporting, mining, data integration/ETL, and dashboarding.
The best open source BI solutions will offer several vital particulars:
- A multitude of common reporting formats (Excel, XML, PDF) must be supported, from flat rows and columns to more sophisticated matrix layouts. In addition, there must be a mechanism for porting data cubes to other OLAP front ends or dashboards.
- A clear and proven path to achieving full integration with your legacy applications, including large CRM and ERP suites. Data access to RDMS and XML sources will be secure and easy.
- There must be a wide breath of online support. Some of this information will evolve in a grassroots fashion and some of it will be created from commercial software vendors and backed by their financial muscle. Community forums and knowledge bases on the internet will offer a front-line means of support for your BI developers and system administrators.
- Enhancements and patch releases must be socialized to the user community in an expedient manner.
- Hybrid models of support will be available, i.e., the software will be free, but paid support (email, phone, consulting) will be available in parallel from a third party and readily codified into SLAs, SOWs, etc.
- Nobody wants to have to build their BI dashboards from scratch, thus, a good number of dashboard templates should be available for quick customization.
- Supported ETL functionality must have transparency, i.e., the ability to see if there are table scans, nested iterations, or onerous joins occurring when data is being scrubbed or aggregated. Many open source BI-focused ETL tools are lacking in this area; however, some of the more robust tools are able to some pretty cool stuff—e.g., take advantage of data movement parallelism, issue asynchronous ETL tasks, or incorporate techniques involving such things as server clustering.
- A forward-thinking BI dashboard will employ data caching and in-memory functionality; this will be especially important as mobile BI becomes much more ubiquitous in the next few years.
- When weighing the benefits of gravitating towards open source BI, make sure there is a clear map of where the open source standards in question are headed so that you maintain maximum scalability and immunity from having a portfolio of brittle BI applications. Keep aligned with products that support and integrate with service oriented architectures (SOA) and will complement remote BI instantiations in the future.
- End users should be able to build (at least in a rudimentary sense) their own dashboards by picking and choosing various display fields, distribution lists, indicators of performance, rollups, calculations and beyond. The key is to achieve a degree of data visualization that is richly intuitive and interactive. Adobe Flash templates are now being successfully used to this effect, providing open source BI with powerful ally and pointing the way toward the future of web-enabled BI and mobile BI.
While a wish list of BI features may be long and complex for some business organizations, others may be seeking massively smaller solutions. The ability of open source BI to scale down should not be overlooked either. There are still far too many IT departments overwhelmed with the cost of enterprise BI and OLAP tools, yet they are using these products to do nothing more than generic SQL reporting.
Is Open Source BI Really A Good Thing?
There is a clear and present danger in being too tightly coupled to a specific BI software provider. The fortunes of software companies can change quickly—a corporate merger, buyout, restructuring or bankruptcy of a software provider can indirectly serve as a catalyst that pushes your dependant applications closer to unwanted legacy status.
The players in the business intelligence space continually jockey for the top positions in various Gartner Group Magic Quadrants and flip-flop in terms of who has the most cutting-edge BI solutions in their portfolio. With such a mutating cast of market leaders, increasing your organization’s disposition for open source BI is a good way to help future-proof BI platforms and hedge risks to future scalability. Open source BI application development can be executed more freely across organizational boundaries, without having to worry about running out of developer licenses or seat licensees, and the corresponding chore of keeping close track of what departments are in danger of exceeding their allocations.
But while on the surface there are tremendous up-front costs savings in equipping the enterprise with open source software, there are many pitfalls to transparently delivering and maintaining successful open source BI applications. With open source, one is sacrificing the indemnification, support and warranties associated with proprietary software; everything exists in a perpetual “as is” state, unless a hybrid model is pursued, where you pay for support and professional services to better ensure an air-tight return on your open source investment. Many additional issues abound: standards tend to evolve organically and slowly over time; compatibility with legacy software can be spotty at best; supporting tools for version control, centralized administration and governance, and change management may be lean; documentation on functions, classes, methods and other components of the open source platform can be contradictory and unwieldy.
These days, the workplace is more or less saturated with developers and architects who are experts (with quantifiable skills and certifications) in multiple BI tools; however, finding a resource with a good familiarity and grounding in open source BI can be a challenge. Companies dabbling in open source BI may have to get used to garnering product expertise from internet forums or open source discussion boards because there is often no single trusted source to turn to for help or quality control. Regularly relying on such expertise—leveraging the advice and brainpower of people who not only exist outside one’s corporate boundaries, but may be employed by a direct competitor—will be impossibly frightening for some organizations. What is liberating for some (having free codebases and code libraries supported by a large community of affiliation-agnostic users) will be seen by others as an unnecessary risk.
Conclusion
Although open source BI accounts for only a small piece of total BI tool sales today, it is continuing to exert pressure on the big name software companies and gain market visibility. It seems almost inevitable that five years from now, open source BI will move from single digits to double digits in terms of market share. Open source BI, with its drastically reduced cost-of-entry into BI, offers the greatest hope to IT managers whom have been seeking a less-expensive delivery model for their enterprise-class business intelligence supply chain. If done right, open source BI applications can provide IT executives with better future scalability (mutual support of SOA and legacy systems) and less overt risk than implementations proposed by many of the big-name BI vendors (they know who they are).
There is no such thing as “standard” open source licensing, so close attention must be paid to the terms and conditions of use for all open source BI elements. As such, your organization may be required to share certain code or template modifications (without receiving monetary royalties) with the rest of the world in order to remain an upstanding member of an open source consortium or community. Frankly however, with the increasing commoditization and openness of software development as a whole, this trend shows no sign of reversing. What may have been seen as threats to intellectual property a few years ago, have been muted by the open source juggernaut.
About the Author
A Contributing Editor for Dashboard Insight, William Laurent is one of the world's leading experts in information strategy and governance. For more than 18 years, he has advised numerous companies and governments on technology strategy, methodologies, and best practices. William currently serves on the faculty of Baruch College, runs an independent consulting company that bears his name, and lectures frequently on various technology and business topics worldwide. He would enjoy your comments at wlaurent@williamlaurent.com
Copyright 2009 - Dashboard Insight - All Rights Reserved.