reports on the launch of Pitney Bowes MDM Software which now has at its core a NoSQL graph database.

This new MDM product has taken a bold step by using as its database platform, not the obvious choice, a relational database (which is what the competition use), but a NoSQL database. As ever with a different database choice, this has pros and cons. This graph database has no need for pesky database indices and, due to its structure, is very well suited to certain types of analysis, such as exploring relationships between data, that is awkward to do in regular SQL. As it happens that is handy for a number of common MDM use cases, where for example you may want to explore the links between customers (such as seeing whether they are related in some way e.g. are in the same household, or perhaps live in the same street or even belong to the same club, assuming you have access to such data). Other examples would be in fraud detection or looking for non-obvious patterns of behaviour as some secretive government agencies need to. As an MDM hub, the database does not need to worry about update processing or transaction integrity as an all-purpose relational database has to. Moreover Pitney Bowes’ existing and well-established data quality tools means that there is no issue with having to build in links to data quality technology, since this is already a core part of the Spectrum platform. There are also opportunities, such as providing pre-built links to social media feeds and, in the B2B space, potentially building links to business information such as D&B or Bloomberg data.
Read the full article.