XML at IBM Information on Demand Conference
November 11, 2008
I thought I’d take a few moments and provide a quick report on the Information on Demand conference in Las Vegas. Like the past couple of years, it was a very interesting conference from a native XML storage point-of-view. There were all sorts of interesting sessions, including sessions about:
- Innovative uses of hybrid relation/XML modeling
- Native XML storage in SOA environments
- Native XML storage for electronic medical records
- Native XML storage for government information sharing
- Native XML storage for electronic forms
- And much, much more
For me, one of the highlights of the show was a session delivered by Li Cui and Charlie Wang from UCLA Health System. They described the different implementations of electronic medical records at UCLA Health System, and the benefits they are realizing from the adoption of native XML storage. Their presentation materials were simply outstanding, offering a very clear before and after indication of the impact of native XML storage. In their words, DB2 pureXML “significantly simplifies database design, web services development and robustly provides capability to handle XML data”. We heard about how it now takes hours instead of weeks to set up new types of electronic medical records, as well as a number of other quantifiable benefits. But it was the impact on the patients that left the greatest impression on me. It was great to hear how–thanks to their new system–patient records are now available much sooner, which obviously has a very real and significant impact on patient care.
Across all of the native XML storage sessions, some themes did emerge. One theme is that many organizations are using native XML storage to optimize their Service Oriented Architecture (SOA) environments. XML is the de facto language for information exchange, and there are a variety of use cases where organizations are persisting that XML in native format. These include the storage of transactional data, the implementation of a service bus cache, the logging of user events, and more. Another theme is that many organizations are using XML to augment traditional relational types. There are situations where modeling data using traditional relational types is challenging. One of those situations is where the structure of data changes often. In such cases, storing the fixed data in traditional relational columns and storing the data whose format changes often in XML format is an ideal solution. It takes advantage of the relative strengths of traditional relational types, as well as taking advantage of the flexible nature of XML for data whose format changes often.
I’m looking forward to the next conference… IOD Europe in Berlin next June.