Contributed Article.  Author: 10gen – The MongoDB Company

 

The debate over relational databases and non-relational databases has been ongoing for quite some time. A few years ago, the argument was mostly one sided in favor of relational databases (SQL) because most of the applications developers were creating were focused on transactional data and the social and mobile explosion could be treated more like a trend than a reality.  However, as we’re beginning to see from companies such as Amazon, Netflix, and MTV Networks, modern applications need the flexibility and scalability of a non-relational database (NoSQL) that rely on bigger data and faster write performance. Modern applications with social components and product recommendations based on user data don’t fit RDBMS.
 

This shift has led some to question if we’re witnessing the end of relational database systems. In a world where cloud computing is really beginning to take over and blossom, NoSQL databases simply make more sense for several reasons. But is saying that relational databases are coming to an end an overstatement?

Big Data 
 

The new trend of “Big Data” is what some believe makes NoSQL databases necessary. The sheer amount of data that’s being demanded of databases today and what will be demanded in the future, call for new types of databases and technology. The days of just transactional data being stored are pretty much gone or quickly fading away. Big data is comprised of the different variety of data, speed in which data changes, and of course the amount of data. When social media can be used to improve a user’s online shopping experience through customized recommendations and provide real time data for analytic applications, developers and DBAs need a solution that can handle “Big Data” and the consensus seems to be that non-relational databases are the answer.
 

If the troubles of migrating to a NoSQL database your only deal breaker, this article by Matt Asay from MongoDb goes into detail about how “Big Data” is increasing the need for non-relational database systems and how migration should be the last think you should be worried about. 
 

Scalability
 

Another issue that you have to consider when discussing relational vs. non-relational databases is scalability. This is a financial issue as much as it is a technical issue. Relational databases are generally stored on a single server. In order for you to scale up, you must add more processors and memory in order to handle the growth of data. If you spread your relational databases across several different servers, you will have to synchronize the databases through replication. This entire process can become unnecessarily complex and time consuming. Non-relational databases allow you much more freedom for scalability because they employ the use of cloud computing. For IT managers looking for a DBMS that will give them more scalability, non-relational databases are becoming the more popular option. Having a viable scaling option in place will help reduce the likelihood that a company will have “Big Data” issues.

Post-Transaction Era
 

Without risk of sounding like a broken record, the fact is times are changing. We’re approaching a time where more and more of the applications being developed are post-transaction. As a result, NoSQL systems are becoming even more relevant and useful. This fact is exemplified by some of the leading tech related companies today. One of the most popular examples is Amazon and their focus on improving search results and the ability to recommend products that a customer might be interested in based on the information stored in their database. This has been a very important facet of their business and represents the direction that many companies are heading in.
 

While some people may never make the switch from relational to non-relational databases, many already have or plan to in the future. These NoSQL white papers will help give you more insight about the growth and development of non-relational databases and why a growing number of developers and DBAs are opting for NoSQL instead of SQL to focus on applications improvements and not database limitations.