I run a company that is defining a new category of data management software, known as an XML content server. I’ve decided to create this blog to record and communicate my experiences in watching this technology / market develop.
It’ s a fascinating market to watch, because XML content servers live at the intersection of several existing markets: DBMS, search, and content management.
While I’ve been at Mark Logic for year, I’ve just gotten around to starting this blog in August, 2005. I’ll summarize the past year quickly in this post. Here’s what happened:
1. XML content servers have continued to be successful in the publishing and government vertical markets. Major publishers are realizing that they should not be storing their content on the file system and accessing it with a search engine. They also know that today’s various RDBMS extensions (e.g., BLOBs, CLOBs, XML shredding) do not work for their content. Thus, they are open to trying a new approach, and one that handles XML content natively. On the government front, the government has a pre-attachment to XML because government loves open standards. In commercial markets, today people say “why XML?” In government, people say “why not XML?” — it’s open, it’s not controlled by any vendor, it’s self-documenting. In short, it’s the perfect format for archiving and searching public content.
2. XML content servers have started to make in-roads in other markets. For example, aerospace, financial services, life sciences, legal have all start to show interest in the category for application such as flight manuals, financial publishing, research analysis, and e-discovery.
3. XML databases have separated into two sub-categories. One group of XML database vendors has focused on managing persistent data on a middle tier for data integration applications, now known as EII applications (e.g., Ipedo). Another, smaller group has focused on managing XML content or documents (e.g., Mark Logic).
Personally, I think the first, data-oriented strategy will lead those vendors to the exact same destination as the object database vendors (i.e., nowhere). That’s because the big relational database vendors will always fight hard to control data. On the flip side, the big relational vendors have always paid only lip service to content, which is helping to create the market for XML content servers.
4. Industry analysts, consultants, and other pundits and gurus have suddenly seemed to change their collective view. For years, they simply said that virtually any new database problem can be handled by (yet another) extension to the relational model. For years, they cited the carnage in the ODBMS market as the example of how the relational vendors stomped an emerging category out of existence through signaling strategies (i.e., “wait for Oracle 8”) and RDBMS extensions. For years, they seemed to overlook the great success of OLAP servers (nee MD-DBMSs) and the focused success of other special purpose DBMS and indexing products (e.g., bitmap indexing a la Sybase IQ).
For example, in talking with one of the top database analysts at Gartner (Donald Feinberg) a few months back, I was both surprised and pleased to see that he was not apparently of the relational-uber-alles camp, but indeed questioned the use of the R in RDBMS (arguing that it’s been so extended that it’s not relational anymore) and indeed said he saw a market for special-purpose DBMSs for a variety of different applications.
Another example would be a recent article by long-time database and search pundit, Curt Monash, where he argues similarly, across a number of areas including OLAP and documents.