Skip to content
Noroff Case Studies

MongoDB

Introduction to MongoDB

MongoDB is a open source NoSQL database management program that has focus on being scalable, flexible, allowing for the storage of large amounts of data in a fast and agile database. MongoDB was created by the owners of a internet advertisement company, called DoubleClick, they struggled with the storage databases not being scalable and agile enough. Instead of using tables and rows, like the SQL databases, MongoDB gathers information and data in collections and documents.

Brief History

In the early days, 1970s, NoSQL was just a supplement to the SQL ways of storing data. But as the internet grew and evolved, so did the number of users, the records and the size of data being stored. The limitations of SQL became more apparent. Adopting a more flexible and scalable way to store data was a priority. This began the evolution of NoSQL, from that MongoDB arose.

In the list below, we will look at important years and releases for MongoDB.

  • 2007 -MongoDB was founded by Dwight Merriman, Eliot Horrowitz and Kevin Ryan. It was initially developed as a component of a planning platform as a service product.
  • 2009 - launched and shifted to a open-source development platform.
  • 2015 - In the early 2015 MongoDB released their version 3.0 - With WiredTiger Storage Engine, Ops Manager and 50-member replica sets.
  • 2015 - Later the same year the released their version 3.2 - With Encrypted Storage Engine, In-Memory Storage Engine. They also released the MongoDB Compass and MongoDB Connector for BI.
  • 2016 - Released the version 3.4 - MongoDB Atlas released, same with Zoned sharding and MongoDB Connector for Apache Spark.
  • 2017 - Released the version 3.6 - Fully managed MongoDB Atlas. change streams to build always-on, real time, reactive applications.
  • 2018 - Released the version 4.0 - Offers multi-document ACID transactions. MongoDB Atlas Global Clusters and enterprise security controls, and MongoDB Charts.
  • 2019 - Released the version 4.2 - MongoDB Atlas Search and MongoDB Atlas Data Lake. And MongoDB Operator for Kubernetes and MongoDB Connector for Apache Kafka.
  • 2020 - Released the version 4.4 - MongoDB Atlas Online Archive, Realm & Sync, MongoDB Atlas multi-cloud clusters.
  • 2021-2022 - Released the version 5.0, and Rapid released - MongoDB 5.x, MongoDB Stable API, Atlas Serverless instances, MongoDB Atlas Data API
  • 2022 - Released the version 6.0 - Atlas Serverless instances, Atlas Data API, Atlas CLI and Flexible Sync. Atlas Data Lase, Atlas Data Federation and Atlas SQL Interface. And Cluster-to-cluster sync.

MongoDBs Main Features

Indexing

  • The search through the database becomes less time consuming and easier to do.
  • This means that every field in the documents are indexed with primary and secondary indices.

Replication

  • This means the database makes copies of the data and sends it to multiple servers. Which make the data more available and redundant, if one server fails, the data can be retrieved from another server.
  • A replica set can contain of two or more copies of the data, each set may act in the role of primary or secondary replica.
  • The primary replica serves as the main server and interacts with the client and performs the read/write operations.
  • The secondary replica is a copy of the data in the primary using built-in replication.
  • If the primary replica should fail, it automatically switches to the secondary, that then serves as the primary server.

Aggregation

  • Makes it possible to perform operations on the grouped data, and then receive a single or computed result.
  • This is thanks to the three different aggregation methods;
    • aggregation pipeline
    • map-reduce function
    • a single-purpose aggregation methods.

Scalability

  • Through sharding you distribute the data to different servers, where the data is partitioned into chunks using a shard key.
  • The chunks are evenly distributed across shards thats placed on different servers. This way MongoDB provides horizontal scalability, which balances the load to keep the system up and running, and gives less chance of hardware failure.

Document-oriented

  • Each database contains a collection which contains documents. And each document can be different, with variations in number of fields and size.
  • The documents structure can be familiar to developers, as they are constructed like objects.

Ad hoc queries

  • Supports search by field, range queries and by regular expression searches.

ACID (Atomicity, Consistency, Isolation, Durability)

  • This is one of the biggest weaknesses of NoSQL databases.
  • MongoDB data model allows data to be stored in a single document, which don’t require a multi-document transaction.
  • After their version 4.0 came out they introduced transactions.
  • A transaction is a group of database operations that will only succeeds if all the operations within the transactions succeeds. This can impact a single record or multiple records.
  • The ACID transactions gives a guarantee that the database will be in a consistent state after running a set of operations.
  • There are some weaknesses to ACID in MongoDB:
    • One is that the database needs to “lock” the resources involved to prevent writes from interfering with one another. So if more than one tries to write in the same data, they can be stuck and have to wait for the current edit to be finished.
    • There is a performance cost to multi-document transactions.
    • Limit transactions to 1 000 document modifications.

Strengths

  • It is schema-less - you don’t need to design any schema.
  • It’s flexible.
  • High-performing
  • Available and Scalable.
  • It replications makes it less prone to hardware failures.
  • Programming languages - they have official drives for multiple programming languages, both major and a large number of smaller once.

Weaknesses

  • Uses a lot of memory for data storage.
  • Security may not be their strongest suite as there has been reports of MongoDB installations has been stolen, and servers has been held for ransome. Creators claim to have taken measures to upscale the security.
  • Joining - there isn’t a simple, automated way to join documents. This can be done, but has to be done manually
  • Limited data size - a document can’t be more than 16MB.
  • Limitation to nesting - it can nest, but there is no nesting above 100 levels.

MongoDB compared with CouchDB

CouchDB is also a NoSQL document oriented database. The design favors availability and consistency. The documents are JSON based and are stored in databases. The software can be installed as single instance or a cluster of instances, this makes it durable and gives it give availability.

CouchDB and MongoDB compared on some key features:

Data Model

  • MongoDB: stores the data in BSON format - documents gets grouped in databases and collections and are returned as JSON. They support a large amount of data types, including: strings, numbers, geo data, dates, arrays, decimal, nested objects, binary data.
  • CouchDB: stores JSON documents. They only support data types like: string, numbers, arrays, objects and booleans.

Indexing

  • MongoDB: indexes are supported in many different types: Compound, Text, Geo, Wild Card, TTL, and Partial.
  • CouchDB: indexing is limited, and for some workloads it requires a “design document” definition.

Query Language

  • MongoDB: Uses their own query language, MongoDB Query Language (MQL). Designed for easy use for developers.
  • CouchDB: Uses HTTP protocol for API, javascript.

Availability and Scalability

  • MongoDB: Using replica and sharding mechanisms. Replica in short: Makes copies of the document so when primary fails, secondary will become primary. Makes it fault tolerant. Sharding in short: Makes it easy to scale collections over replica sets, geo-zone sharding makes it easy to manage data.
  • CouchDB: Also uses a concept of replica and sharding. Replica in short: Makes copies of the documents, which makes this fault tolerant, but each “node” is also the shard, the sharding is a automated mechanism making it not possible for the user to define keys or groupings.

This comparison is created with the help of MongoDBs own comparison, and CouchDBs homepage .

Summary

Overall MongoDB is a strong NoSQL database. In the author’s view the strengths trump the weaknesses. MongoDB is easy to use, it has many features many developers look for, it supports many different development languages, which makes it more accessible for different types of developers. It can contain big amounts of data and the data is easy accessed.

Credits

  • Jenny Marie Feragen (Jenfer93)

References