LESSON - Open Source Business Intelligence and Data Integration: Eight Reasons to Jump In
By Yves de Montcheuil, Vice President of Marketing, Talend
and Mike Boyarski, Senior Product Marketing Manager, Jaspersoft
Your organization may be living in the past—and paying dearly for it.
Due largely to opinions formed 5 to 10 years ago, many IT teams still shy away from open source business intelligence (BI) and data integration (DI) solutions—at least for mission-critical applications. Even at this late date, we’re hearing vague assertions that open source BI and DI don’t meet the needs of large enterprises.
As with all things technological, however, it pays to periodically reexamine your assumptions. In that spirit, let’s examine eight popular misconceptions about open source BI and DI—and explain why they no longer apply.
1. It’s not reliable enough. Some IT professionals still connect “open source” with “buggy,” but two things have happened since open source software was labeled this way. First, enthusiastic communities and installed bases have grown—open source BI and DI communities now include hundreds of thousands. The stability of the software has also increased, so much so that it’s used to power well-known, Web-based applications used globally by millions.
Even more important is the rise of commercial open source software solutions, which have become a “best of both worlds” answer for more and more enterprises. IT organizations get the flexibility, easy customization, and large, supportive communities associated with open source projects, as well as the stability born of scheduled, qualityassured releases; key project developers; training; documentation; and other advantages once confined to proprietary solutions.
2. It won’t scale. The notion that open source BI can only work with small data stores is obsolete. In fact, both commercial and project-based open source BI solutions regularly report directly against multi-terabyte databases. Support for clustering and other strategies enables the performance needed to meet these large-enterprise requirements.
Open source DI has suffered from a similarly unfair reputation, but modern, commercial versions are currently performing DI tasks involving very large databases, complex data transformations, and numerous data sources and targets. One leading solution already natively supports cloud-based architectures and the Apache Hadoop software framework used for distributed computing requirements, enabling high-performance and highly scalable data integration. The same solution sports an industry-leading 460 pre-built connectors, allowing for rapid integration among legacy systems, packaged applications, and more.
3. Features are basic and limited. A quick look at the customer lists for leading commercial open source BI and DI vendors tells the tale here. Open source BI and DI are now in use on multimillion-dollar projects for reporting, analysis, and data integration—because the functionality is there.
On the BI side, open source customers enjoy sophisticated, graphical, and highly interactive report presentation, dashboarding, point-and-click report design, plus in-memory and OLAP-based analytics. Additionally, many customers rate the ease of use of these solutions higher than that of proprietary BI products.
On the DI side, where connectivity is king and complexity is the big hurdle, commercial open source solutions also hold up well (remember the industry-leading 460 pre-built connectors example).
4. The price is wrong. Here, two conflicting misconceptions coexist. The first asserts that “It’s free—and you get what you pay for,” and the other says “Commercial open source costs just as much as proprietary—it’s commercial, after all—so why bother?”
Both assertions are wrong. Commercial open source BI and DI solutions are most certainly not free—the amenities discussed elsewhere must come at a price. But that price is dramatically lower than what traditional closed source vendors charge, with documented total costs as much as 82 percent lower. This is possible because the vendors get significant R&D help from their communities—not to mention a lot of effective Web-based marketing.
Commercial open source DI and BI are not free—they’re just highly cost effective.
5. The support’s not there. BI and DI projects get a lot of visibility now, and deployment teams don’t want to be left twisting in the wind. Who will be accountable when help is needed quickly? And if community support isn’t as “official,” can it really help in a jam?
Most developers who have worked with open source solutions—commercial or project-based—place considerable value on the role of the community, regardless of whether they have a support plan. When commercial open source solutions are in play, the vendor assumes accountability for software quality and support, offering services equal and even superior to those of proprietary vendors.
Today, a steadily increasing number of Global 200 organizations—across financial services, healthcare, government, telecom, and other industries—are using open source BI and DI tools for mission-critical applications.
While the commercial open source vendor’s service-level commitment establishes a high comfort level, the open source communities— with their convenient online forums and forges—are powerful collections of backup resources. And documentation from commercial open source vendors is every bit as strong as that from similarly sized proprietary vendors.
6. My company won’t let me use it (for good reason). Some companies specifically prohibit the use of software licensed under GPN/GPL or similar constructs, citing liability issues. Sometimes there’s also a safety concern that illicit developers might smuggle malicious functionality into source code.
Commercial open source BI and DI vendors eliminate these worries by providing commercial licenses—and by adding their internal quality assurance processes to the inevitably extensive stress-testing effect of their communities. Code contributed from communities offers double protection: the contributor’s assurance of “free and clear” intellectual property, plus the rigorous testing completed by the vendor’s internal team. The commercial open source vendor assumes all related legal liabilities, so enterprises need not fret.
7. Open source BI and DI are only used by tiny enterprises or corporate renegades. The misconception here exists on several levels. First, surveys have shown that despite the general belief that open source solutions aren’t in use at a respondent’s company, they usually are—often widely so. Next, there’s the belief that open source BI and DI are best suited for small, proof-of-concept projects or developer tools, where production data and functionality won’t be placed at risk.
The realities differ. Open source powers some of the world’s leading data environments (think Facebook, Google, and Yahoo!). Large IT organizations like open source for its flexible and modern commodity hardware–friendly architecture—and they like the strength of open source communities, which include large IT groups from other leading enterprises.
8. We’ll be left behind. Some organizations think working with a proprietary vendor assures ongoing innovation—but recent experience shows the opposite. Commercial open source BI today enables rapid ad hoc analysis using highly responsive, in-memory analytics, while proprietary BI vendors are still holding back or just now in their first attempts. The same holds true for multitenancy, cloud computing support, and so-called mashboard support (multi-data feed integration). First-anywhere support for Hadoop illustrates the same principle on the DI side.
Also, open source DI and BI solutions tend to be based on newer, more modern architectures, making them easier to evolve, integrate, and customize.
The pleasant reality: open source BI and DI work.
Today, a steadily increasing number of Global 200 organizations—across financial services, healthcare, government, telecom, and other industries—are using open source BI and DI tools for mission-critical applications. Their large IT organizations believe open source to be the best way to introduce new capabilities to their organizations without stressing budgets.
So next time you hear “We’d better not” in your organization, make sure everyone knows the facts. There’s no reason to hold your data initiatives back because of a few ancient myths.
This article originally appeared in the issue of .