Which database is best? A com­par­is­on of MongoDB vs DynamoDB

Our com­par­is­on of MongoDB vs DynamoDB revealed striking dif­fer­ences in each solution’s approach to storing and managing data. Above all, DynamoDB is fully embedded in the AWS cosmos. MongoDB on the other hand, could offer more freedom.

MongoDB vs DynamoDB, two re­com­men­ded NoSQL services

While in the past there was hardly any way around the classic SQL database, the emergence of ever larger amounts of data has meant that ap­plic­a­tions and websites need a new way of handling data. The solution came in the form of NoSQL systems (Not only SQL) such as MongoDB and DynamoDB. These do not work re­la­tion­ally thus making them much more flexible and, above all, more scalable. Instead of tables, these solutions mainly rely on documents. So, they can also store un­struc­tured or semi-struc­tured data in a way that creates clarity and saves space. If a company’s re­quire­ments grow, the database grows with them without any problems arising.

What is MongoDB?

MongoDB was released in 2009 by 10gen (now MongoDB, Inc.). It has since become one of the most popular solutions in the field of SQL database man­age­ment systems. The software is written in C++ and stores data in the form of binary JSON documents (BSON). MongoDB is open source, dis­trib­uted under the Server Side Public License (SSPL) and offers optional com­mer­cial support packages. The database, which owes its name to the English term ‘humongous’ (meaning ‘gigantic’), is known for its flex­ib­il­ity and scalab­il­ity. MongoDB uses the MongoDB Query Language (MQL) for queries.

What is DynamoDB?

DynamoDB has been around since 2012. The server­less NoSQL database comes from Amazon and is provided by AWS (Amazon Web Services) as a com­pletely managed service. The ap­plic­a­tion is pro­pri­et­ary and was ori­gin­ally written in Java to manage the ever-in­creas­ing volumes of data from the company’s own e-commerce activ­it­ies. The database is document-oriented and is con­sidered one of the fastest and most scalable solutions in the NoSQL segment. Although DynamoDB also uses tables, these are not re­la­tion­al. Due to its structure, the system is not only par­tic­u­larly suitable for large companies, but also for smaller companies whose data volumes and workloads increase over time.

The purpose and scalab­il­ity of MongoDB and DynamoDB

On comparing the MongoDB and DynamoDB systems, it seems clear why the two NoSQL databases are used for similar purposes. The two systems can be used wherever large amounts of data need to be clearly stored and managed. MongoDB uses a technique called sharding for ho­ri­zont­al scaling. Large data records are auto­mat­ic­ally split and the load is dis­trib­uted across several servers. As the system can also work with un­struc­tured data without any problems arising, it is ideal for managing and providing websites and e-commerce ap­plic­a­tions. All required content, including images or videos, is stored in one document. Retrieval then takes place im­me­di­ately without any no­tice­able op­er­a­tion­al delays.

DynamoDB offers similar de­ploy­ment options. Amazon itself re­com­mends the in-house database for ap­plic­a­tions, media workloads, e-commerce and gaming platforms, among others. This claim is pretty un­der­stand­able since the system offers — at least the­or­et­ic­ally — infinite scaling options and has nothing to hide when compared with MongoDB.

However, one of the biggest dif­fer­ences between the two systems is their in­teg­ra­tion into AWS. DynamoDB does not work outside of this ecosystem and is therefore only suitable for users who can live and work with this lim­it­a­tion. But in return, you get com­pre­hens­ive ad­min­is­tra­tion and benefit from extremely fast read and write processes, even with rapidly in­creas­ing data volumes.

Func­tion­al­ity

Although both databases may have similar goals, a closer look at the MongoDB vs DynamoDB com­par­is­on quickly reveals the different ap­proaches. MongoDB com­pletely does away with the SQL-typical tables and instead stores data in documents, which in turn are bundled into col­lec­tions. These documents in BSON format consist of key-value pairs, can have different schemas and can be changed by adding or deleting a field. While keys always consist of character strings, values can also contain character strings, other documents, arrays or Boolean values. The text search in MongoDB is carried out via a text index.

DynamoDB works as a NoSQL key-value store that stores data on SSD hard disks. The data can be stored schema-free as numerical values, strings or binary. The in­form­a­tion is divided into non-re­la­tion­al tables con­sist­ing of items that must contain at least one attribute. The op­er­a­tions ‘Create’, ‘Read’, ‘Update’ and ‘Delete’ are available for writing and reading.

Data structure

There are also dif­fer­ences in the com­par­is­on of MongoDB vs DynamoDB with regard to the data structure and maximum data size. MongoDB’s storage format is called BSON and is a binary form of JavaS­cript Object Notation (JSON) that does not require a schema. If you want to create a new document, you do not need to create your own col­lec­tion in advance. Instead, you create the document with a simple command, which is then either inserted into an existing col­lec­tion or created auto­mat­ic­ally. The system allows data with a maximum size of 16 MB.

The data structure of DynamoDB, on the other hand, is much more re­min­is­cent of the table form of classic SQL solutions. Here, too, there are tables that contain elements with certain prop­er­ties, although the number of these prop­er­ties may vary. Before you can use a table, you must first create it and define a primary key that par­ti­tions the data. Each of these par­ti­tions contains three nodes with a copy of the data. As the database is com­pletely managed by the system, this process is fully automated. There is also a no­tice­able dif­fer­ence between MongoDB and DynamoDB in the size of the stored documents. With Amazon’s service, the limit is maximum 400 KB.

Avail­ab­il­ity of the data

By default, DynamoDB divides the three nodes mentioned into one primary node (leader node) and two sub­or­din­ate nodes. All read and write op­er­a­tions are performed on the primary node first. Only then is the rep­lic­a­tion carried out on the two sub­or­din­ate nodes. However, it is also possible to dis­trib­ute the read op­er­a­tions to all three nodes in order to prevent per­form­ance losses.

The somewhat older solution in the MongoDB vs DynamoDB com­par­is­on also dis­trib­utes the data across several nodes and thus prevents problems which could be caused by a server failure. Similar to the com­pet­i­tion from Amazon, there is a primary node that receives the write op­er­a­tions and then makes the data available to the sub­or­din­ate nodes. If the main node fails, one of the other nodes auto­mat­ic­ally takes its place and becomes the lead node, which, thanks to rep­lic­a­tion, already contains all of the data. The process can take up to 60 seconds.

In­fra­struc­ture, com­pat­ib­il­ity and pro­gram­ming languages

One of the clearest dif­fer­ences between MongoDB and DynamoDB is the de­ploy­ment and in­fra­struc­ture of the two services. MongoDB is a com­pletely free open source solution with pro­pri­et­ary support offerings. The database man­age­ment system is com­pat­ible with most cloud solutions, operating systems and con­tain­ers. How much control and ad­min­is­trat­ive effort is necessary and possible is entirely up to the user. The database must be con­figured and main­tained in­de­pend­ently. On the one hand, this can be demanding and time-consuming. On the other, it can offer a great deal of freedom.

DynamoDB takes a different approach. The database is part of the AWS ecosystem and does not work outside of this en­vir­on­ment. This means that although DynamoDB works extremely well with other tools from Amazon, the ap­plic­a­tion is only com­pat­ible with other solutions to a limited extent. The fully managed database is im­me­di­ately ready for use and handles updates, scaling and other tasks auto­mat­ic­ally. This does mean that users only have limited insight into the work ‘behind the scenes’.

The following table sum­mar­ises which pro­gram­ming languages the two systems support:

Security

DynamoDB performs a number of tasks for its users, including all security activ­it­ies. As part of the AWS cosmos and the IAM model (Identity and Access Man­age­ment), the database is well protected and embedded. There is no direct con­nec­tion to the internet, as requests are first routed through an API gateway.

Compared to DynamoDB, the workload when using MongoDB is sig­ni­fic­antly higher if you want to ensure the security of your database. Access au­thor­isa­tions, roles and firewalls must be im­ple­men­ted and main­tained in­de­pend­ently, at least in the standard versions of the system. Managed solutions that provide all the necessary security features can resolve this issue.

Which companies rely on MongoDB and DynamoDB?

Companies that use MongoDB in whole or in part include the following:

• Adobe
• Amadeus
• AppScale
• Craftbase
• Disney
• Etsy
• Foursquare
• Lyft
• MTV
• The New York Times
• Via Varejo

Companies and platforms that manage their data with DynamoDB include the following:

• Airbnb
• Amazon
• Comcast
• Disney
• Dropbox
• Netflix
• Nike
• Redfin
• Samsung
• Tinder
• Zoom

Which service suits your re­quire­ments?

MongoDB and DynamoDB are two very powerful database man­age­ment systems that remain flexible and scalable thanks to their NoSQL approach. However, the approach of the two solutions differs from each other. If you are looking for a system that is perfectly embedded in the AWS cosmos, has a strong security ar­chi­tec­ture and takes care of most ad­min­is­trat­ive tasks for you, DynamoDB is an excellent choice. If, on the other hand, you want freedom of choice in terms of your cloud en­vir­on­ments and con­fig­ur­a­tion, MongoDB is the optimal service.