In-Depth
Open Source BI Shines
Companies that have used open source solutions for a decade or more are now looking at that distribution model for their BI tools.
After numerous fits and starts, the outlook for open source business intelligence (BI) now seems very bright, despite (or perhaps because of) a chilling global economic outlook.
Market research tallies from Forrester, Gartner Inc., and IDC paint a composite picture of a teeming enterprise free and open source software (F/OSS) segment with substantial penetration in virtually all technology domains.
Open source is getting the attention of the BI and data warehousing (DW) segment, where luminaries such as Claudia Imhoff (who's penned a whitepaper on the rise of OSS BI) and Wayne Eckerson (who considers the pros and cons of OSS BI and DW in a post --http://portals.tdwi.org/Blogs/WayneEckerson/2009/07/Enterprise-Software.aspx -- on his Wayne's World blog) recently tackled the subject, both of whom were positive about the subject.
Interest in open source BI is strong. A recent survey from Aberdeen Group found that fully 25 percent of respondents plan to adopt open source BI tools of some kind over the next two years. If you want to take a longer-term view, consider a recent market research prospectus from IDC, which predicts strong compound growth of 22.4 percent for OSS software through 2013. That's considerably higher than previous projections,
OSS is increasingly a part of the enterprise software strategy of leading businesses and is seeing mainstream adoption at a strong pace," noted IDC group vice-president Michael Fauscette, in a statement.
The Price Isn't Necessarily Right
OSS advocates -- be they vendors, industry watchers, or enthusiastic third parties -- tend to downplay the cost factor, stressing that open source software isn't necessarily free -- or (more to the point) that if you buy into OSS with the expectation that it's free (or if you're motivated primarily by a desire to cut costs), you're missing the point.
Instead, vendors (and many non-vendor-affiliated advocates) like to position OSS adoption as a function of both features and cost; or -- as author, consultant, and veteran data warehouse architect Mark Madsen likes to put it -- of open source software solutions having matured to a point where they're now perceived by most customers as "good enough."
There's another wrinkle here, too. Madsen, for example, likes to contrast maturation in the OSS business intelligence and data warehousing space with consolidation in the broader (closed source) BI and DW markets. The OSS BI and DW segments teem with vendors: there are a handful of OSS data integration (DI) efforts (Talend and the Kettle Project among them); several OSS reporting entries (including robust offerings from JasperSoft and the BI Reporting Tool, or BIRT, project); several putative full-blown BI Suites (JasperSoft and Pentaho among them); dozens of open source offerings such as CRM (including SugarCRM Inc. and Concursive Corp.); and even an OSS data mining and statistical analysis effort (RStat, which has received attention from closed source players like Information Builders Inc).
Contrast that with a broader BI marketplace in which three of the biggest best-of-breed players (Hyperion Solutions Corp., Business Objects SA, and Cognos Inc.) were acquired by larger database or applications players (Oracle Corp., SAP AG, and IBM Corp., respectively). In addition, the erstwhile BI best-of-breeds acquired other specialty or niche vendors (e.g., FirstLogic, Cartesis, Celequest, and Applix, just to name a few), while their stewards have likewise nabbed more than a score of additional boutique or niche players.
As TDWI's Eckerson has noted (http://www.tdwi.org/News/display.aspx?ID=8704), the acquisitions of Hyperion, Business Objects, and Cognos didn't simply involve three (prominent) players: among them, the pure play trio -- along with competitors-cum-stewards IBM, Oracle, and SAP -- had acquired dozens of separate BI or DI vendors.
The consolidation craze has continued apace since late 2007: since then, BI, DI, or DW pure-plays have continued to get snapped up, with venerable names like GoldenGate Software and SPSS Inc. ceasing to exist (as independent entities) in just the last month. Viewing the traditional BI segment as an ecosystem, an evolutionary biologist would doubtless remark the loss of a lot of genetic diversity in the acquisition activity of these and other companies. Although the BI ecosystem still abounds with new or maturing players, F/OSS offerings account for much of what might be described as the freshest alleles in the BI gene pool.
In another sense, OSS BI and DW tools might be said to offer customers choice -- in addition to a "good-enough" feature set and significantly lower costs. The danger of consolidation-fueled vendor lock-in has been raised OSS vendors themselves (see http://www.tdwi.org/News/display.aspx?id=9207).
The Price Isn't Necessarily Wrong, Either
Even if adoption isn't primarily a function of OSS' reported low cost, pricing is a factor. As Claudia Imhoff noted in a recent research paper, pricing concerns drive OSS adoption at any of several levels, with departmental downsizing, IT budget cuts, and a great emphasis on rapid ROI accounting for much new interest.
"Building a BI environment is still critical to … companies [that have downsized their IT departments], but the ability to build these critical applications is increasingly difficult. They need all the help and technology they can get," Imhoff writes, with a nod to the community-oriented origins of most OSS BI projects. "While BI still remains the number one initiative for most enterprises, CIOs are struggling to find the funding for these projects.
"Outsourcing, budget cutbacks, [and] spending freezes are all in play," she continues. "The ability to purchase necessary BI technology is becoming increasingly difficult. Again, the availability of free BI software is quite attractive."
Another important driver, she suggests, is that cash- and resource-strapped enterprises have little room for error. Although this might sound counterintuitive -- in light of the history of many OSS BI technologies -- that's no longer the case.
"[There's] minimal tolerance for IT failures or trial-and-error projects," notes Imhoff, who concedes that while "testing BI technologies or potential applications was a useful way for IT to learn what worked and what didn't, … it was also fairly expensive and led to disillusionment … [among] business users."
Enter OSS BI tools, which -- thanks to the availability of no-cost, no-commitment Community Edition (CE) packages -- typically give prospects an easy means to try before they commit to (much less buy) a technology. This is one respect in which the OSS model clearly outshines the inflexible and (to a degree) reactive licensing models that still characterize the BI status quo.
"The ability to test whether a technology might work was quite limited; implementers either had to buy the technology outright or get a limited-time license which may have been too short for a proper trial," Imhoff points out.
Imhoff, like Eckerson and other impartial observers, doesn't view OSS BI as a slam dunk, of course. She cites a number of barriers -- including a lack of formal training or support and a high rate (or "velocity") of change -- in addition to a perceived lack of features.
Eckerson, for his part, cites concerns about the upstart nature of many OSS BI efforts, which -- in the current clime, especially -- raises uncomfortable questions about vendor viability. Open source BI players, not surprisingly, tend to cite both the breadth of their community support and the security of the OSS model: even if a start-up vendor should disappear, its source code will live on forever -- or for as long as anyone is still using it, they claim (see http://www.tdwi.org/News/display.aspx?id=9207).
"[In Talend], you have someone who's giving you not only a solution that's flexible and diversified, but you have access to the source code, too," said Vincent Pineau, general manager for the Americas with Talend, in an interview last year. "You don't have to sign an escrow [agreement] with us. You're not worried about what's going to happen if we get acquired or if we disappear."
There's also the 800-pound elephant in the room: the specter of intellectual property (IP) infringement, which -- emphasized by the The SCO Group Inc. and its notorious pursuit of both IBM Corp. (which it accused of misappropriation by open-sourcing its Unix software patents) and enterprise users of Linux, such as Autozone and the former Daimler Chrysler.
In Imhoff's view, IP concerns haven't so much worked to constrain as to discipline the adoption of OSS BI tools, which -- on balance -- is probably a good thing. "While justification of these fears is often anecdotal, they have led to demands for more audits of open source code and legal indemnification from open source software vendors," she notes.
More to the point, the demonstrable benefits of OSS BI tools outweigh both its real and theoretically drawbacks. "In a recent survey of BI implementers, 84 percent said software license costs were a major or moderate inhibitor to expanding their use of BI," Imhoff writes, citing a thriving OSS BI ecosystem, the proven reliability, scalability, and security of OSS offerings, and the unabated growth of the broader OSS community as other important benefits.
Imhoff allows that the price is, in fact, nice: "The fact that OSS is free is perhaps the most compelling benefit -- you don't need to justify significant up-front fees before you even get started with a BI project. Instead, you download the software, begin building a prototype of a BI application, and demonstrate it to the business community, long before you have to fork over a single dime to a software company. It is a nice way to try before you buy."