Around three decades Oracle, IBM, Microsoft relational databases were consistent leaders for Operational Database Management Systems. However, currently all of these vendors offer NoSQL platform parallel to their traditional database platform e.g. Microsoft got Azure Cosmos DB as NoSQL database and so on.
The world of NoSQL emerged in 1998 and it's gained popularity in 2010s. In 2015, the popular NoSQL database provider MongoDB appeared in Gartners magic Quardent as a leader for operational database management system.
Fig 1:Gartners Magic Quadrant in 2015 |
Though recent trends shows that NoSQL is getting more popular in Big data landscape than relational database management system.
What is NoSql and Why?
NoSQL stands for 'Not only SQL'. NoSQL database is for storing both structure and semi structure data.
NoSQL stores data in one of four categories:
- Key-Value storage
- Document storage
- Wide Column storage
- Graph database
The most popular NoSQL database MongoDB is document storage, however, it can be used as a key/value store, which is just a subset of features.
Today data become more broader in terms of shape and size. Data is spread around documents, social media, complex web application, IoT sources. The traditional SQL database can't handle these data due to the continuous changing behavior. Traditional database need to know shape of the data (schema) beforehand to store those, hence it failed to capture continuous evolving data. And thus, NoSQL emerge and conquer the world!!!
Fixed schema vs. No Schema
For structure database (traditional SQL database) you need to have schema ready before you insert data to a table but for NoSQL you don't need to have schema created at first rather you can directly insert the data. Though, it's better to say 'Schema agnostice' than 'Schemaless' e.g. Microsoft's NoSQL database Azure Cosmos DB known as schema agnostic database, which is not bound by schemas but are aware of the schemas.
Relationship vs. No Relationship
Traditional SQL require to maintain relationship to perform, that's why normalization is there. Whereas, NoSQL doesn't require to maintain the relationship. You can embed a few tables into one table in NoSQL database.
For structure database (traditional SQL database) you need to have schema ready before you insert data to a table but for NoSQL you don't need to have schema created at first rather you can directly insert the data. Though, it's better to say 'Schema agnostice' than 'Schemaless' e.g. Microsoft's NoSQL database Azure Cosmos DB known as schema agnostic database, which is not bound by schemas but are aware of the schemas.
Relationship vs. No Relationship
Traditional SQL require to maintain relationship to perform, that's why normalization is there. Whereas, NoSQL doesn't require to maintain the relationship. You can embed a few tables into one table in NoSQL database.
Below table (fig:2) shows basic differences between SQL and NoSQL database:
Fig 2: Comparison between SQL and NoSQL |
Conclusion: NoSQL denotes 'Not only SQL', it means NoSQL database can perform what traditional relational database can do as well as do more. But these two types of Database have different expertise, as per business requirements you can choose from them. A way to find out which fit best for you could be asking the question. Do you know the shape of data well advance? If answer is 'Yes' which means you can define schema earlier and can build the relationship before data arrive then it's good to choose traditional SQL. On the other hand, when you don't know the shape of data and behavior changes continuously then go with NoSQL. Nevertheless, some complex enterprise solution can be built by using both SQL and NoSQL database to leverage best of each.
1 comment:
I think this is a very crucial post related to SQL and REST. Apart from this I also have some interesting topics to share.
SQL Server Load Rest Api
Post a Comment