As we reported yesterday, IBM announced the release of DB2 9, which is the official release name for Viper, their effort to incorporate XML content into a relational database. Microsoft and Oracle have their own strategies for doing this, and once all their work settles down and starts to get seriously deployed, building enterprise applications will never be the same. It has been 20 years since the early demand for databases that could handle marked-up content (SGML back then), and there have been many products developed to manage SGML/XML repositories since then – Astoria, Berkley DB XML, Ipedo, Ixiasoft, Mark Logic, Software AG, Vasont, X-Hive, and XyEnterprise are some currrent examples.
There has been lots of debate over the years about the best approach to managing marked-up content, and it is safe to say that there is not a single answer. This means that you need to understand what the differences are between them – and it won’t be easy for those of you new to the unstructured data world – this is much trickier than the relational data world. In spite of the huge benefits of the major DB players providing serious XML support, the wide variety of content application requirements will ensure a long-term need for quite a few specialty vendors, whether they are targeting vertical applications or horizontal components. The good news is that even with many different XML application schemas, it continues to get easier to integrate all kinds of XML data.
Bill points to an article in eWeek here. Also see Dave Kellog’s comments and links to other articles.