Most developers are familiar with relational databases, where tables are used for defining the structure of a dataset and where relationships are established by linking tables together via shared columns/data.
By contrast, NoSQL ("not only SQL") databases such as MongoDB store data using flexible models. But just because a NoSQL database isn't traditional doesn't mean it can't store relational data — it stores the same data differently and, in some cases, even more efficiently.
Because NoSQL databases were built to tackle the complex data problems of the internet era, many can handle relational data while offering more modern functionalities such as:
Table of Contents:
Many developers, accustomed to using relational database management systems (RDBMS) for decades, may not be aware of the advantages that a NoSQL database brings to applications built for today's internet. The advantages of a NoSQL database include:
The ability to easily accommodate evolving schemas and data models as applications grow and change.
Scaling for massive amounts of data spread out across multiple data stores.
High-speed performance for querying that associates complex and related data.
Support for analytics operations through complex queries.
Flexibility in deployment options — on-premise, in a private cloud, or in the public cloud.
There are four main types in use today:
Document databases have a very flexible data model — the document — that can vary from record to record. They store data in JSON (JavaScript Object Notation) or BSON (Binary JSON) data formats, which provide flexibility in working with data of all types. These databases serve a wide variety of use cases and can employ complex querying using multikey, geospatial, and full-text search. A document database like MongoDB is deployed on a scale-out architecture, giving it the freedom to run anywhere — on-premise or in the cloud.
Key-value stores employ a simple schema, with data stored as a simple collection of key-value pairs. Keys are unique, and the value associated with a key can range from simple primitives to complex objects. Because of this simple model, key-value stores can be incredibly fast for both reads and writes, and they are easily scalable.
Wide column stores capture huge volumes of data in a row-and-column format. While they are considered NoSQL databases, their format makes them similar to relational databases. They differ from relational databases in that every row is not required to have the same number of columns. Wide column stores are often built to handle big data use cases that require aggregation for queries.
Graph databases have a fundamentally different structure, in that data elements and their relationships are stored as a graph (as in graph theory, not a line graph). Complex queries can be used to determine multiple kinds of relationships between data points and among clusters of data points.
NoSQL databases are best for modern applications where data models often evolve and scalability is essential.
NoSQL databases are often used in agile projects because schema changes — which are common — don't require a lot of overhead. This allows developers to focus on business logic and algorithms instead of dealing with schema updates. If you anticipate that your application's data model needs to remain flexible to accommodate changes over time, the flexible schema approach of NoSQL databases — specifically document databases — may fit well for your needs.
Finally, scalability is a major reason to choose a NoSQL database. NoSQL databases have been designed for easy horizontal scaling without the need for specialized third-party components. MongoDB, for example, supports sharding out of the box. Whether you need to support hundreds of users or hundreds of millions of users, a NoSQL database can scale seamlessly.
The kind of NoSQL database that best fits your needs ultimately centers around a few main concerns: transactions, data models, query performance, scaling, and deployment. The following table lists the distinctive approaches for each of these concerns or when it would be advantageous to use a given type of NoSQL database over another.
| Database Type | Document database | Key-value store | Wide column store | Graph database |
|---|---|---|---|---|
| Example Offering | MongoDB | Redis | Apache HBase | Neo4J |
| Transactions | Supports ACID transactions. | Supports ACID transactions. | Does not support transactions, but guarantees atomicity at the row level. | Supports ACID transactions. |
| Data Models | Best when data is modeled by a set of interrelated objects, with its flexibility toward data structure making it a good general purpose database. Documents can contain nested structures for capturing complex data. | Excellent for frequent high-speed access to the same chunks of data, even if those chunks of data are large. | Best for extremely large sets of data, where querying patterns are predictable, often for supporting aggregation and analytics. | Suitable when there is a need to store and query data about the connections between related data, such as in social network contexts. |
| Query Performance | Fast retrieval of individual documents, much like key-value stores, but can also support complex queries with aggregation pipelines. | Best for instantaneous retrieval when queries are not complex; useful for caching use cases. | Excellent when aggregating values on particular columns, with strong compression and in-memory support. | Fast for simple queries, but also efficient when answering questions about connections between data; excels when profiling data and determining clusters of related objects. |
All four types of NoSQL databases have a similar approach: They support horizontal scaling, achieved through a combination of in-memory support and distributed systems that serve the database from a cluster of nodes. Similarly, all four NoSQL databases are deployable as self-managed systems on-premises, in a private cloud, or as a DBaaS in the public cloud.
With all types of NoSQL databases offering similar advantages over RDBMS, we see that the question of which kind of NoSQL database to choose depends mostly on your data model and querying requirements.
Each NoSQL database type has its own pros and cons, and your choice should depend on your actual use cases. Yet, for most modern websites, the ideal NoSQL databases to use are the document stores, as they support most general use cases.
A document database is particularly suited for websites thanks to its ability to support complex data structures and parent-child relationships within a single document. There is no need to incorporate object-relational mappers (ORMs) or object-document mappers (ODM) because document databases store the raw objects in JSON/BSON or similar formats.
As the most efficient NoSQL database option for websites and API endpoints, MongoDB is easy to get started with, and it can grow and evolve with your website. MongoDB already has excellent integration with popular web programming languages like Python, PHP, Node, Java, Golang, and many others. Additionally, several libraries help with using MongoDB with various web frameworks like Laravel or Ruby on Rails.
The MongoDB PHP Driver connects to deployments hosted in MongoDB Atlas, Enterprise, and Community. It consists of two components: the extension and the library. This was one of our first supported languages!
MongoDB has two Kotlin drivers: the Kotlin Coroutine Driver for Kotlin applications using coroutines, and the Kotlin Sync Driver for synchronous Kotlin applications. This makes MongoDB a popular NoSQL database for Kotlin.
When you decide to use MongoDB for your website or web application, it may seem daunting to deal with setup and infrastructure. If you simply want to focus on building your app, you can offload the infrastructure concerns by using a database as a service (DBaaS). You can let MongoDB handle the database management while your development team focuses on what they do best.
Atlas is the fully-managed DBaaS offering of MongoDB. That means you can sign up and start using a MongoDB database without deploying and managing a cluster yourself. You can decide how to scale and grow your cluster with a few simple clicks. MongoDB Atlas also offers a free tier, allowing you to learn and experiment with the platform, without even needing to provide credit card details.
You can also choose which cloud provider to use for your cluster. So, if your application already runs on Amazon AWS, Microsoft Azure, or Google Cloud, then MongoDB Atlas can be deployed seamlessly in the same cloud and connect securely with your application.
NoSQL databases have enabled us to create applications that can scale to handle large volumes of data and a large number of users. They support a wide variety of use cases across industries. However, with so many different NoSQL databases out there, deciding which one to use is not easy.
When choosing a database for a general-purpose web application, for example, document databases are the ideal choice. They support a variety of data structures without the overhead of implementing a wide column store. In addition, they enjoy the benefits of complex querying capabilities and simple scaling.
MongoDB Atlas allows developers to build their applications without managing database infrastructure, coupled with the flexibility of using whichever programming language they prefer.
The best NoSQL database for your project will depend on your needs. To select the right NoSQL database for you, take transactions, data models, query performance, scaling, and deployment into consideration. MongoDB is easy to set up and will evolve with your website.
It integrates with popular languages, including Python, PHP, Node, Java, Kotlin, and Golang.