Bloor Research - Datenintegrationsplattformen Marktübersicht

Data Integration Platforms - Talend Author : Philip Howard Publish date : July 2008 page 1 Introduction Talend is an open source provider of data integration products. However, while many open source suppliers are built on little more than a man and his dog, Talend has a business model and organisation that more closely resembles that of a conventional software vendor, while still leveraging the open source model. Thus, for example, it is the only open source vendor in this market to have offices around the world (in China and the United States as well as France and Germany) so that it can offer 24 x 7 support to its corporate customers. Similarly, the company has an extensive partner network, both with other open source and proprietary vendors, as well as with a number of system integrators. The company’s flagship product is Talend Open Studio. This is a code (Java and Perl plus SQL) generating tool that supports both ETL (extract, transform and load) and ELT (extract, load and transform) approaches to data integration. A notable feature of the product is Business Modeler which, as its name implies, makes it easier for business analysts to become involved in the development process and to enable collaboration. It is worth noting this in a broader context because Talend introduced this capability before either IBM or Informatica introduced anything similar. This is a good indicator of Talend’s intention to be treated seriously as an enterprise-class vendor, as are the product’s support for multi-developer environments (in the Talend Integration Suite, which bundles in value added services as well), metadata management, high-performance deployment options and the recently introduced support for change data capture to enable real-time operating environments. The company also offers its capabilities through Talend on Demand, which is the company’s software as a service (SaaS) offering. By April 2008 there had been over 250,000 “core product” downloads of Talend Open Studio and, while this is not necessarily indicative of usage (and still less of customers paying for support or other services) it is still a substantial number. Moreover, Talend has some significant accounts that it can name, such as Virgin Mobile, Alliance Healthcare, SNCF French Railways, US Cellular, Swiss Life and Shopping.com amongst others. In any market, open source represents a potentially disruptive approach. However, relatively few vendors are effectively organised to take advantage of such disruption. Talend is an exception. © 2008 Bloor Research Data Integration Platforms page 2 Introduction Champion At an enterprise level, the market for data integration is now less focused on traditional ETL (extract, transform and load) and more centred on data integration as a broader concept. A number of vendors SAP (Business Objects) Informatica have been preaching that message for some time but it is only recently Ab Initio SAS IBM that this idea had achieved broad acceptance within the marketplace. Oracle (Data Integrator) Sybase This is not just a question of the ETI Talend major suppliers getting their marketing messages across but it can also be seen when you look Pervasive at the spectrum of offerings Cha Pitney Bowes Open Text (Genio) or available from different vendors. iWay Today, the position is that vat llen there are a number of providers ETL Solutions CloverETL no of data integration platforms on ge the one hand while, on the other, In r we are seeing the emergence of specialist products that focus on just a single aspect of data integration such as data migration, content ETL, web ETL, replication Figure 1: Bullseye assessment – the nearer the centre, the better and synchronisation, data federation, and data services amongst others. In particular, there is a significant emerging sub-market for be packaged as a web service. Further, the data migration and a Market Update on this product needs to be able to support both subject is being published alongside this one. batch and real-time operations and have We expect that data services will emerge as scheduling, reporting, debugging, auditing, Market Update a further sub-market in due course but this data lineage and similar facilities built in. remains at an early stage. Ideally, there should never be a requirement to leave the environment to code complex We are also aware of new vendors planning to transformations as you then lose data lineage, enter the market with innovative, potentially impact analysis and where-used capabilities. disruptive, approaches to data integration. There is a general trend towards introducing more business friendly capabilities into data In this Update we are focusing on data integration offerings. integration platforms suitable for deployment on an enterprise-wide basis. We are not, Products should also be able to handle a therefore, considering any of these specialised wide range of data sources and targets solutions, nor broader integration solutions and will need to be able to handle not just that include ETL capability (unless ETL is a structured data but also XML, semi-structured specific focus), nor solutions focused on a sources such as EDI and SWIFT messages, specific target database or use case. and unstructured data in the form of Word documents, Excel spreadsheets, PowerPoint Key market issues presentations, and so on. It will also be helpful if the software has an understanding of the To genuinely offer an enterprise-ready date metadata that different content management integration platform implies a long list of vendors wrap around the content that they requirements. To be comprehensive the store so that the platform can be used to product (suite) needs to offer not merely migrate from one content environment to ETL but data federation, replication and another and also to support federated queries synchronisation as well as data services that across those environments. Data connectors can plug data retrieved by the platform into that have specialised knowledge of source an SOA environment. Transformation services environments, especially leading application and similar facilities should also be able to products, will be an advantage. We do not regard data quality provision (both Also of interest are the moves by other page 3 data cleansing and profiling) as a necessary suppliers, notably Business Objects (now requirement for a data integration platform. SAP) and SAS, to join these leading two. Both However, if it is not provided by the vendor of these companies, of course, are saddled then a close integration with appropriate third with the subjective impression that they are party tools will be required: simply having a merely in the BI/ETL business (as is/was the partnership is not enough. case with Cognos) but both are keen to break out of this mould and have the products to We do not have any particular view about the do so. Business Objects, we believe, is better benefits or otherwise of a “black-box” solution placed to do this, at least in the short term, as opposed to code generating approaches; in part thanks to its recent acquisition by there are arguments on both sides, though SAP. Business Objects also has advanced it will clearly be an advantage, in some capabilities aimed at business analysts. instances, to be able to generate code (such as Java) that can be embedded within third Ab Initio is also a significant vendor at the party applications. We are aware that certain high-end of the market but, as the company government agencies do not like the idea of has a policy of not talking to analysts, we can black box.