by Angela Guess
A recent article looks at the differences between relational and non-relational databases and situations in which each might be ideal: “There are two types of databases you can use when hosting data in the cloud: a relational database or one that is so not so relational. For example, Amazon offers a choice of SimpleDB (not so relational) or MySQL (relational). Soon it will offer Oracle 11g (relational) as another option. Some cloud providers, such as Microsoft, offer a similar choice, while Google, for example, offers only a not-so relational approach. Typically, software-as-a-service (SaaS) offerings run on top of relational databases… With all these options, it is important to understand when you might prefer a relational approach or one that is not so relational, or when it might be appropriate to use a hybrid approach.â€
The article goes on, “To begin with, why I have called some databases ‘not so relational’? The reason is that there are two classes of these databases which confusingly are known as NoSQL, and NOSQL databases. One stands for ‘Not SQL’ and the other stands for ‘Not only SQL’. The basic fact is that these are all non-relational databases in the way that they store data. The distinction between ‘Not SQL’ and ‘Not only SQL’ is that the former can only be addressed via a suitable API and the latter can also be addressed using SQL. This distinction is not confined to the cloud, neither is it new. For the moment we need to understand the problems that they try to address.â€
photo credit: aaronparecki
