Open Source Data Integration - Eine unabhängige Produktanalyse

Talend Open Studio v2.1 An InDetail paper by Bloor Research Author : Philip Howard Publish date : October 2007 This document is Copyright © 2007 Bloor Research In our view, Talend is the most enterprise oriented of the open source data integration vendors Philip Howard Talend Open Studio v2.1 page 1 Fast facts Talend is an open source provider of data • The product does not have parallelisation integration products or, in traditional parlance, capabilities at present. Talend Integration ETL (extract transform and load), although the Suite has the ability to parallelise across product also supports an ELT approach. The servers while version 2.2 of Talend Open company’s flagship product is Talend Open Studio (due at the end of September 2007) Studio but the product is also available as the will support multi-threading. core component within Talend On Demand, which is the company’s software as a service • Metadata management, based on a (SaaS) offering; and the Talend Integration repository that supports reuse, is provided Suite, which is a bundled product and services within Talend Open Studio though full offering that includes additional software multi-developer support with version control capabilities for multi-developer environments and check in and out is only available within and to provide high performance deployment Talend Integration Suite. options, including scheduling. • As far as we know, Talend is the only open Key findings source vendor in this market to have offices outside its home market: in this case in China In the opinion of Bloor Research the following and the United States as well as France and represent the key facts of which prospective Germany. This means that Talend can offer users should be aware: 24x7 support on all continents. • Talend Open Studio is a code generating The bottom line product. That is to say, it generates either Java or Perl (plus SQL) that you can deploy as We have long been of the view that the required (and across a grid, if appropriate). advantages inherent in a code generating This means that Talend avoids the ‘black box’ approach have been overlooked when approach of many other vendors that can compared to black box approaches—largely result in the black box becoming a thanks to the marketing machines of the major bottleneck. ETL vendors—so we are glad to see it making a comeback. In particular, the ability to generate • A major advantage of the code generating Java will be of particular appeal to companies approach is that it is easy to encapsulate a wishing to embed relevant routines within their job as a web service that can be invoked by own applications. third party applications, or to embed the generated Java into third party applications. Talend therefore has two main claims to fame: This is more of an issue for conventional first, it is open source—if you like open source approaches because they require an engine then we believe that Talend Open Studio (the black box) in order to run. represents an excellent option. Secondly, it is code generating and, again, if you are convinced • Using a conventional ETL approach you by the merits of code generation then Talend is would normally deploy your applications as likely to appeal to you. The real issue will be if close to the source as possible but ELT neither of these holds true, in which case the options allow you to deploy on the target. At question is: how well does Talend Open Studio present the support for ELT is limited to stack up when compared to the conventional MySQL and Oracle databases but support for suppliers within this market? In our opinion: more environments will be introduced in due surprisingly well, which suggests that the course. product is worth serious consideration in more or less all circumstances. • The recommended approach for development within Talend Open Studio starts with the Business Modeler before drilling down to developer functions. We particularly like this: by no means all of the major ETL vendors have introduced comparable facilities as yet. • In most respects Talend Open Studio looks pretty much like other products: a graphical user interface, dragging and dropping onto a palette, a visual SQL builder, debugging facilities and so on. It actually uses the familiar Eclipse interface. • There are data cleansing capabilities built into the product, though not data profiling. Talend Open Studio v2.1 page 2 Vendor information Background information Product availability and support Talend was established, in Paris, in 2002 and it Talend Open Studio is currently in version 2.1.1. first came to market in August 2005 with the Typically, the company issues a ‘milestone’ beta version of Talend Open Studio. The release every month (similar to beta versions in company is privately owned and backed by the proprietary world, but available to all venture capital. It is an open source vendor and developers) and a major release very quarter. the product is available simply as a free Version 2.0 was first released in April 2007. The download from the Talend web site. There is product is made available under a GPL v2 therefore the question of where the company license and the environment provided is based will make its revenues? In practice there are on Eclipse. three answers to this: the first is that Talend Open Studio is marketed through indirect The most notable feature of version 2.2, channels to partners that want to embed Talend expected at the end of September 2007, is Open Studio’s capabilities within their own support for parallelisation of data integration product sets. These partners may or may not be jobs: Talend Open Studio will support multi- open source vendors themselves, as is threading in this release. JasperSoft, for example. In any case, Talend provides chargeable support to these partners. Note that Talend Open Studio is a code (Java, Perl and SQL) generating product and there is Secondly, Talend charges for training, support therefore no requirement for a run-time server and technical expertise (consulting) via a direct to host Talend Open Studio, just a JVM for Java sales model. To this end the company has or an interpreter for Perl. For integration opened offices in the United States, Germany purposes it supports FTP, HTTP and web and China as well as its native France. services. More than 120 connectors are available out of the box in the current version. Thirdly, Talend has two additional offerings: Talend On Demand, which is a software as a As is indicated by the fact that Talend has service (SaaS) offering whereby Talend charges offices around the world, it is able to offer 24x7 for its hosting facilities; and the Talend support. Integration Suite, which is a combined services and product offering for large enterprises. This includes some features (for example, for team-based development) that are not available in the standard Talend Open Studio product— these features will be discussed in the relevant section that follows. Finally, it is worth noting that Talend has partnerships with other open source vendors, such as MySQL, Ingres and SugarCRM, as well with a number of systems integrators; and that Talend is a co-founding member of two important consortia: the Open Solutions Alliance and OW2. Talend web address: www.talend.com Talend Open Studio v2.1 page 3 Product description Introduction Source/target support Talend Open Studio is suitable for use in all The product has native support for more than sorts of data integration environments but it is 30 databases and ODBC/JDBC can be used particularly targeting operational requirements otherwise. Specific capabilities for supported such as data migration, consolidations and databases are extensive. For example, for real-time embedded capabilities in addition to MySQL there are separate components for bulk the sort of facilities required to load data loading, commits, reads and writes, rollback warehouses. This is not because Talend Open and so on. In version 2.1 there is now support Studio is inferior in any way in the data for slowly changing dimensions (types 1 and 2).