I have been considering why many companies do not consider No-SQL or Document based storage systems for their applications. Instead they continue to churn out object to relational database mappings with lots of custom code in order to get adequate performance.
One of my most recent experiences brought the reasoning to the surface. The individuals making the decision were quite clear on why this kind of move wasn’t the best implementation at this time.
One reason was the number of reports that still needed to be supported for performance metrics. These reports were built using a reporting tool connecting to database stored procedures. Introducing document storage would hamper some of the existing reports, and make it more difficult to generate reports on the new data.
Another reason that also made a lot of sense was that they did not have any document management systems in place today. Servers would have to be requisitioned. Operations people would have to be trained. Security would have to be implemented. Sure, it would shorten the development cycle, but it would have a big impact on a release date.
The final reason was that they already had an infrastructure and support team in place for traditional Relational Databases. There was nothing to purchase or deploy. Deployment was fast and seamless.
This experience has helped me to see things from a different perspective. It is more likely the case that a company has an investment they need to utilize rather than a fear of something new. Would you agree? Why not leave a comment with your perspective, or drop an Email to btaylor@sswug.org.
Cheers,
Ben