Straight Dope About Open-Source BIStraight Dope About Open-Source BI

Open-source business intelligence has changed nothing, yet it is making all the difference in the world. What conclusion other than the first can you draw from analyst indifference and the yawns of established BI and enterprise application vendors? Six years after the launch of the open-source Mondrian OLAP server - widely used in the open-source (OS) world - there's no trace of open-source BI (OSBI) anywhere in Gartner's 2007 BI magic quadrant analysis. Yet Mondrian is the analytical core of at least four OSBI suites - JasperAnalysis, OpenI, Pentaho, and SpagoBI - and the packagers claim to be "selling" software like hotcakes. Clearly the explanation for the seeming contradiction is that we have two or even three different BI markets at work.The first market is conventional BI users, typically business and financial analysts, often folks who work with enterprise applications. No one knows this world better than Nigel Pendse, author of The OLAP Report and a former analyst who has been following BI for decades. According to Nigel, "the proprietary BI software vendors seem to be genuinely unconcerned by OSBI. They never mention it to me, and they seem quite surprised if I ask them about it. A few have looked at products like Pentaho and seem totally unimpressed/unconcerned. I guess they don't sell into [the open-source world] anyway, and therefore aren't losing any business to OSBI that they are aware of."

Nigel sees a parallel universe of OSBI adopters who have "decided for idealistic, economic or technical reasons that they must have an open-source solution, and don't even consider any proprietary options." That's a second BI market, yet I don't believe that assessment fully captures the situation. Much of present-day software innovation is based on the J2EE stack and on ecosystems that may mix open-source and conventionally licensed tools. My Jolt Awards participation - my own review of contest entries and my reading of my fellow judges' evaluations - confirms this reading. And open source seems to be the BI option of choice for Java developers. OS is free and easy, it integrates with the rest of the J2EE stack, it plays nice with popular developer tools such as Eclipse, and best of all, there are (usually) no nasty licensing conditions or costs to complicate sales and deployment to end users.

Web services may change the picture a bit - the ability to invoke on-demand BI functions in a service-oriented architecture (SOA) - but the mainstream BI vendors have been slow to promote SOA. Why should they? They long-ago refocused from technology to higher value solutions. Microsoft's 1998 SQL Server 7.0 OLAP Services entry into the BI space commoditized BI; it accelerated the BI pure-plays' move into performance management, financial forecasting, industry verticals, enterprise-application integration, and the like. Oracle isn't paying $3.3 billion for Hyperion to get Essbase. They're buying a credible performance-management story, value that wasn't created by selling BI tools to developers.

The hidden world of J2EE developers: that's where OSBI is making all the difference. Those developers are building BI functions into line-of-business applications for the Web and the enterprise. They're bringing analytics to places and people who don't know they're doing BI or who can't match the SAPistas' fat wallets. We experts have been predicting BI for the Masses for years. I'm beginning to suspect that it will be open-source business intelligence that delivers.

P.S. You've just read the gist of my presentation to last month's LinuxWorld Solutions Summit.Six years after the launch of the widely-used open-source Mondrian OLAP server, there's no trace of open-source BI (OSBI) in Gartner's 2007 BI magic quadrant. Yet Mondrian is the core of at least four OSBI suites - JasperAnalysis, OpenI, Pentaho and SpagoBI - and the packagers claim to be "selling" software like hotcakes.