Guy Harrison - MongoDB Performance Tuning: Optimizing MongoDB Databases and their Applications

Any source code or other supplementary material referenced by the author in this book is available to readers on GitHub via the books product page, located at www.apress.com/9781484268780 . For more detailed information, please visit http://www.apress.com/source-code .

When MongoDB emerged in 2009, database technologies were at a crossroads. For more than 20 years, relational databases such as Oracle, SQL Server, and MySQL had dominated the database market. These databases, which combined the relational data model, SQL language, and ACID transactions, had been the foundation for applications that transformed modern business and which powered the Internet revolution. But by the middle of the first decade of the new century, it was clear that the relational database was failing to meet the demands of a new breed of always-on, globally scalable, web applications. These new Web 2.0 applications demanded new breeds of database management systems.

By 2010, a plethora of non-relational NoSQL systems had emerged Hadoop, HBase, Cassandra, and many others. Of these non-relational upstarts, MongoDB has been by almost any measure the most successful. As we write this, MongoDB ranks as one of the top five database management systems. Of these top five, only MongoDB is based on 21st-century technologies. The other four (Oracle, MySQL, SQL Server, and Postgres) all have their origins in the 1980s and 1990s.

MongoDBs success can be ascribed to many factors such as alignment with object-oriented programming paradigms and compatibility with modern DevOps practices. In the main, MongoDB has thrived because it made life easier for developers. However, in the past few years, weve seen MongoDB graduate from a by developers for developers database to a platform supporting a new generation of mission-critical systems across an increasingly broad range of enterprises.

As MongoDB has matured and expanded its enterprise footprint, performance management has become increasingly important. As we know, poorly performing customer-facing applications can be fatal for todays online enterprise. For instance, when the load time for a web page increases from 1 second to 5 seconds, the probability of a user abandoning the page rises by 90% directly impacting online revenue. And because databases perform so much disk IO and data crunching, the database is often the root cause of that poor performance.

Furthermore, in the cloud, performance management is cost management: poorly performing databases consume unnecessary CPU, memory, and IO resources that cost real money. A couple of days spent tuning a large-scale MongoDB-based cloud application could potentially save hundreds of thousands of dollars in hosting fees.

Indeed, we could even argue that performance management is an environmental imperative. The electricity that powers busy database servers costs more than just money its also associated with greenhouse gas production. Reducing energy consumption in the home is a social responsibility; reducing energy consumption in the data center is as important. A badly tuned MongoDB database is like a poorly tuned car that backfires and belches smoke: it may get you from A to B, but it will cost you more in gas and exact a heavier toll on the environment.