There was a recent article in Information Week which I found largely
inaccurate and poorly researched, regarding the NoSQL vs SQL debate.
The link to the article is:
http://www.informationweek.com/news/software/info_management/232901328
My response to the editor follows:
====================================
Hello.
I read with interest Mr. Kajeepeta’s article on the need for a more
thoughtful approach to selecting data management technologies, rather
than succumbing to the latest fad/trend. While I agree with much of
what he states, particularly the need for a mixed/multi-database
approach to modern application design, there are a few flawed elements
of the article that I felt merited correction. First of all, graph
databases such as Neo4J, Infinigraph, and AllegroGraph are most
definitely *not* “binary large object databases”. In fact, they are
graph databases with a fundamentally different data storage and
retrieval model than all of the other types of databases mentioned,
and in general, are actually quite poor at BLOB data.
In fact, graph databases are a critical piece of a modern application
strategy, in that they allow data models that are often far more
reflective of the “real world” than our attempts to fit the world’s
information neatly into rows, columns, and documents. His example of
product and customer master data as an “inappropriate” use of NoSQL
demonstrates ignorance of the capabilities of graph databases. Also,
he made a blanket assumption that NoSQL databases do not support ACID
transactions. That is simply untrue, and is one of the
differentiating features of Neo4J, for example. It is very clear from
the article that Mr. Kajeepeta has very little experience with most of
the products he has declared inferior to his beloved relational
databases. I also agreed with his summary statement – that it isn’t
about NoSQL, it’s about “not only SQL”. A phrase which, ironically
enough, is attributed to Emil Eifrem, CEO of Neo4J. I should also
note that we leverage both relational and graph database models in our
ThingWorx platform.
While it was a worthy topic for an article, the factual errors
unfortunately undermined the credibility of the overall content.
Sincerely,
Rick Bullotta
CTO/Co-Founder
ThingWorx