Steigern Sie Ihren Return on investment (ROI) mit Open Source Datenintegration

The Return on Investment of Open Source Data Integration WHITE PAPER Talend White Paper The Return on Investment of Open Source Data Integration Table of Contents Table of Contents ............................................................. 2 Open source, a model that benefits all parties .......................... 3 The different alternatives open source data integration is replacing........................................................................ 5 Elements of the ROI........................................................... 6 ROI calculations ............................................................... 8 Cost Comparison and ROI ................................................... 17 Summary & Conclusion ...................................................... 19 Page 2 of 19 Talend White Paper The Return on Investment of Open Source Data Integration Open source, a model that benefits all parties Cost is often presented as one of the major reasons for adopting open source solutions for IT projects. However, cost savings must be deter- mined carefully, as the actual cost of a project goes well beyond the ob- vious initial costs – the license costs, which can be viewed as the tip of the iceberg – and include a much greater component, the hidden part of the iceberg. It is not the purpose of this white paper to provide a tutorial on the cost calculation of an IT project. Rather, this white paper will explain the different alternatives to open source data integration, and will present the costs – visible and hidden – of these alternatives, on several types of projects. Before getting into the core subject of this white paper, it is worth men- tioning a few facts about the business model of open source vendors — and killing several false ideas at the same time. • The free part of open source is not related to the absence of costs but to the freedom to use the solution (to use a common analogy, it's "free" as in "free speech", not as in "free beer"). • Open source vendors are for-profit organizations, often backed by reputable financial firms, who expect a return on their investment. • Although there are variants to this model, open source vendors often provide a no-cost solution to the community and charge for services (support, training, expertise, etc.) and value-added features, needed only by enterprises for large-scale projects. • However, even the services and value-added features are signifi- cantly less expensive than what vendors of comparable proprietary solutions would charge for them. There are many reasons for this lower cost, among which one can mention: Page 3 of 19 Talend White Paper The Return on Investment of Open Source Data Integration o A sharply decreased cost of sales, due to the broad com- munity adoption (oftentimes, projects start small with the free version, and grow). o The involvement of the community in the design, devel- opment and testing of the solution decreases the cost of development. o Open source vendors, being usually newer companies, don't drag along years and years of legacy compatibility that they have to maintain. • Open source vendors are not "vampires" who take advantage of the free work of the community. Vendors do give back a lot of features to community users, at no charge. And by the way, one should never assume open source community members are naive and are taken advantage of — they are not! The open source model is actually beneficial to all parts involved: community members, paying customers, and vendors. Page 4 of 19 Talend White Paper The Return on Investment of Open Source Data Integration The different alternatives open source data integration is replacing Open source is not the only approach to data integration. Other ap- proaches used by many organizations include on-demand custom devel- opment, and proprietary data integration solutions. On-demand custom development, a.k.a. "manual coding" On-demand custom development, also called manual coding, is perhaps the most widely used approach to data integration. Simply put, it con- sists of writing programs in third generation programming languages "We only had three months to develop our (C++, Java, etc.), or in scripting languages (Perl, Python, etc.), to ad- integration solution - actually, more like 5 to 6 dress data integration needs. These programs would typically extract weeks when accounting for tests and deploy- data from one or several sources, perform any number of data process- ment. Custom scripts development could not ing and transformations, and then write the data to its target(s). While possibly have worked, both in terms of timing offering a great degree of flexibility, the manual coding approach is and in terms of features." time-consuming, requires high levels of programming skills, and poses -- Medialab great technical difficulties for any process that goes beyond the simplest ones. In addition, manual coding creates maintenance headaches, fa- miliar to anyone who has had to maintain code written by other developers. Proprietary data integration solutions Unlike manual coding, proprietary data integration solutions provide an "We have quickly been pushed to abandon [Ora- abstraction layer that decreases the complexity of the developments, cle's] solution (...): our growth model turned out and offer much higher productivity. However, proprietary data integra- to be incompatible with the new pricing model of tion solutions are plagued with high costs (both visible and hidden), long Oracle, based on a cost learning curves, and do not offer a high degree of flexibility. per target CPU." -- Eurofins Page 5 of 19 Talend White Paper The Return on Investment of Open Source Data Integration Elements of the ROI The cost studies presented here take into account a number of ele- ments. Some of them can easily be quantified, while others are difficult to evaluate and are mentioned here as intangibles. • Development licenses costs include the acquisition costs of devel- opment licenses. They apply only to proprietary data integration solutions. • Training curve costs include both quantifiable elements — the costs of training classes and the salaries of the staff during their training and rampup phases — and intangible elements: the time to market of the project, which is delayed by lengthy training. They apply both to proprietary and open source data integration solutions (for the sake of argument, this white paper considers that developers performing manual coding are already proficient in the language they use). • Development time includes both quantifiable elements — the sala- "The open source model limited the need for an ries of the staff during the development phase of the project — and upfront investment, an important factor in a intangible elements: the time to market of the project. They apply fiscally responsible or- ganization, financed by to all approaches. taxpayers' money.