LESSON - Faster, Higher, Farther: Open Source in Business Intelligence: Bringing Open Source to Enterprise Business Intelligence
By Yves de Montcheuil, Vice President of Marketing, Talend
Open source brings important advantages to business intelligence (BI) projects, especially when compared with proprietary solutions.
Just at the time when nobody argues anymore that open source databases are ready for the deployment of enterprise BI projects, a new layer of open source products is becoming prevalent in many of these same projects: open source ETL/data integration and open source BI tools. The advantages of such products are clear and cannot be ignored.
Even for large organizations, extracting data from multiple databases and systems is always a complex proposition. Fundamentally, proprietary solutions consist of “black box” tools, which either run or don’t—but if they don’t, nobody has any way to figure out why, unless the reason is obvious: the vendor did not find it useful to support such and such source system, because the market for it is too small. And if they do run, the acquisition costs, coupled with ongoing maintenance and deployment costs, will always end up making it an extremely expensive project.
Among the high costs of proprietary data integration, one typically finds (in no particular order): license and training; rarity of skills for proprietary environments; per-source, per-target, per-CPU, per-core development and runtime licenses; need for dedicated high-performance hardware, etc. Indeed, these solutions are deployed in monolithic architectures: the software is installed on a physical server, and all the transformation processes are centralized on this server. With fast-increasing data volumes, performance of the data server will be degrading, and the only response of proprietary vendors is to recommend buying additional servers—with additional licenses of their tool.
Compared to these constraints, open source brings numerous advantages to the data integration table. They start with a zero-programming interface and a fast learning curve. Then comes the use of industry standards (Java, Perl, SQL, SOA, etc.) that makes finding trained consultants a lot easier. The code generation approach brings a huge advantage compared to proprietary engines—both from the standpoint of performance and the ease of deployment and ramp-up.
What is holding companies from transitioning to open source data integration? Not much, actually.
But where open source brings maybe the most value is in the breadth of its technological coverage. Proprietary vendors focus only on the subset of technologies that bring them the most revenue and hence provide only a few dozen connectors to the mainstream databases and ERP. Conversely, the contributions of the open source community— carefully reviewed and vouched for by the vendor’s development team—allow such solutions to cover databases (proprietary and open source), data warehouse appliances, ERP and CRM (conventional and SaaS), files, Web Services, and many more technologies. Connectors are available for all users to download and use, at no extra charge.
Finally, the cost is a critical factor. Not that open source is free—this is a myth. Open source vendors are well aware that the cost of the project goes well beyond license costs—and furthermore they are not charitable organizations. Commercial open source vendors are companies backed by reputable venture capitalists who expect a return on their investment. However, open source data integration is free to adopt through a downloadable version, licensed under GPL and not restricted in use. Commercial subscriptions to additional features and enterprise-grade support cost only a fraction of what proprietary vendors charge. And since there is no charge per-source, per-target, or per-CPU, incremental deployments are not hindered by budgetary considerations.
What is holding companies from transitioning to open source data integration? Not much, actually. Already many large and small organizations have made the choice, including leading financial institutions, telecommunications operators, retailers, entertainment companies, manufacturers, institutions of higher learning, and government agencies.