In the last year, major RDBMS (Relational Database Management Service) vendors have launched new and low-cost versions of data integration/ELT (Extract, Load, and Transform) software that generate SQL scripts. Yet, many major market analysts do not consider RDBMS vendorsâ data integration tools to be viable for enterprise wide solutions and in situations where the sources and target RDBMS are heterogeneous.
So, against this background, what is the right approach for the data integration industry? Yves de Montcheuil, Director of Product Marketing at Sunopsis, comments: "For at least 10 years, most ETL software tools were based on proprietary engines, and traditional ETL vendors brought a strong value proposition to the market. In the 90s, the most widely used versions of the major RDBMSs, such as Oracle, Ingres, Informix, SQL Server, Sybase or DB2, provided a reliable means for storing and accessing large volumes of data but not a means to process and transform massive amounts of data. This situation has changed dramatically over the last few years â now SQL and RDBMS can be the engines to perform transformations."
As is typical in a growing market, proprietary solutions dominate the market at first, generally at a high price tag, until an industry standard emerges that can achieve what the majority of the customers need. Over the next few years, SQL is in a strong position to become the industry standard for data integration.
De Montcheuil remarks: "The main obstacle for an IT team to move from hand-coding to an ETL tool is the flexibility that the developer looses when using a tool. Only an ETL tool based on a native SQL code generator technology provides the flexibility required to move from hand-coding to automatic code generation without throwing away all previous developments."
He continues: "Since the mid 90s, at the end of the marketing battle â fought mainly by Oracle and Sybase - RDBMS vendors have focused their engineering efforts on enhancing the functionality of their SQL languages and improving the performance and reliability of their engines. Some vendors have even added advanced analytic features â capabilities that traditional ETL tools do not offer within their proprietary languages.
Indeed, most RDBMS vendors have started to leverage the power of the enhanced SQL and now include data movement software in the server package, seriously challenging the market owned by traditional (independent software) ETL vendors. Todayâs SQL is sufficiently powerful to perform all of the work needed to integrate data even when complex transformations are involved."
Facing these new competitors, traditional ETL vendors have started to include some SQL as a possible development language choice for performing the required data integration. However many have limited the SQL to the capabilities of their proprietary engines and created their own syntax. In contrast, a new generation of powerful third-party SQL code generators has emerged. With the power and reliability of todayâs RDBMSs, IT teams strongly believe it makes sense to use the RDBMS in place to perform the data integration work.
Now that SQL is rich enough to perform ETL work, vendors of SQL code generator software offer compelling reasons for adopting their innovative approach:
- The learning curve for this approach is very short.
- The use of a SQL generator suggests that there is no need to add any software or hardware to the production IT environment.
- Performance is unbeatable: SQL and RDBMS allow a bulk approach of data processing. Bulk is up to 100 times faster than the row by row processing that proprietary ETL engines require.
- The code generator produces native SQL code for the RDBMS without the time consuming process of manual coding.
- The cost of code generators is lower than that of traditional ETL software that is based on a proprietary language and an engine, most often installed on a dedicated server.
- There is no need to acquire and maintain a new server to perform the ETL processes.
De Montcheuil concludes: "It is the flexibility of SQL generators that delivers tremendous flexibility in the ETL process â it doesnât matter whether transformations need to be performed on sources or on targets. The availability of RDBMS data integration products has definitively changed the landscape of the data integration market, providing a strong push to SQL as the industry standard language for data integration. Such SQL-based products actually pave the way for the much more powerful third-party SQL code generator products which allow complex data integration jobs to be performed in batch or real-time, without any coding. Over the next few years, all the signs are that SQL will become the industry standard language for data integration and ELT and the right architecture for enterprise data warehousing."