What’s the Difference Between Cassandra and MongoDB?

Apache Cassandra and MongoDB both belong to the NoSQL database group. NoSQL databases can store structured, unstructured, and semi-structured data without a database schema.

Data storage in NoSQL databases isn’t constrained by the tabular format and relationships between tables, unlike in a traditional relational database management system. You can freely partition and replicate data across multiple nodes to scale efficiently. 

Additionally, Cassandra and early versions of MongoDB are open source. This means you can download the source code of these NoSQL databases and configure them as you like.

Facebook developed Cassandra and then released it as an open-source project with Apache. MongoDB was developed by a small group of developers under MongoDB, Inc. All versions of MongoDB released before October 16, 2018 are available under the GNU Affero General Public License.

Read about NoSQL »

Read about open source »

Due to differences in their data models, Cassandra and MongoDB implement several database features differently.

Basic unit of storage

In Cassandra, Sorted String Tables (SSTables) are the basic unit of storage used to persist data on disk. An SSTable is a file that contains a sorted set of key-value pairs for a particular column family (table) and partition. SSTables are immutable, which means that once they’re written, they can’t be modified. 

In MongoDB, the basic unit of storage is a document. A document is a set of key-value pairs where the keys are strings and the values can be of various types. For example, values can be other documents, arrays, strings, numbers, dates, and Boolean values. Documents are stored in collections.

Query language

A query language is the statements you use to insert and retrieve data from the database.

Cassandra Query Language (CQL) is the query language that you use on Cassandra. While it has a similar syntax and structure to SQL, Apache has developed CQL to work with the column-family data model.

On the other hand, MongoDB uses MongoDB Query Language (MQL), which has similar commands as Node.js. MQL supports create, read, update, and delete (CRUD) operations. You can write MQL commands on the MongoDB Shell.

Indexing

Indexing is a technique used in databases to improve the speed and efficiency of data retrieval operations. It involves creating a data structure that maps the values of one or more columns in a database table to the physical location of the corresponding data on disk.

Cassandra supports two types of indexes:

• Secondary indexes on individual columns
• SSTable-Attached Secondary Indexes (SASI) indexes on multiple columns

SASI indexes store index data directly in the SSTables. They support complex queries like range, prefix, and full-text search on columns with large numbers of unique values.

In contrast, MongoDB supports indexing at the collection level and field level. It provides multiple index types like single field, compound, and multikey. It also offers these indexes:

• specialized geospatial index for geographically distributed data
• a text search index for large volumes of text data
• hashed and clustered indexes for numerical data

Concurrency

In databases, concurrency refers to when multiple users or processes can access and perform database transactions simultaneously without interfering with each other. 

Cassandra achieves concurrency through turntable consistency and row-level atomicity. Only a single user can operate on a single row at a time.

In turntable consistency, each replica node maintains a vector clock, which is a data structure that tracks the version history of associated data. When a write operation is performed, the vector clock updates to reflect the new version. When a read operation is performed, Cassandra returns the version with the highest timestamp across all replicas, which ensures that the most recent version of the data is always returned.

In contrast, MongoDB supports mechanisms for multi-version concurrency control (MVCC). MVCC allows multiple versions of the same data document to exist simultaneously. Each document has a unique revision ID that is incremented on each update. Document-level locking and MVCC provide a more robust concurrency strategy.

Availability

Availability means you ensure there’s no data downtime, even during server outage. Both Cassandra and MongoDB ensure availability by replicating data across multiple sever nodes.

In Apache Cassandra, each node in the cluster holds data replicas for other nodes. Every node coordinates reads to the correct node to write or pull data. Simultaneously, it also repairs data that got out of consistency across the nodes. This can impact performance at scale.

In contrast, MongoDB uses single primary node replication to offer high data availability. MongoDB replicates data into replica sets. Only one primary node receives the writes, and the other nodes simply replicate data from the primary node. However, the primary node creates a single point of failure.

Scalability

Both Cassandra and MongoDB allow sharding—a technique to horizontally partition data across multiple nodes in a cluster. For example, if you have thousands of customers, you partition the data so every node only has the information of a few hundred customers. That way you can scale your database without causing bottlenecks.

Cassandra uses a distributed hash algorithm called consistent hashing to determine which node is responsible for a particular data value. Cassandra also supports virtual nodes (vnodes), which allow a single physical node to have multiple data ranges.

In contrast, MongoDB uses sharding keys to identify where the data value can go. Database administrators can define sharding keys to partition the data. You can divide the data based on factors like geographical location, alphabetical order, or any other system that is most efficient for your dataset.

	Apache Cassandra	MongoDB
Data model	Cassandra uses a wide-column data model more closely related to relational databases.	MongoDB moves completely away from the relational model by storing data as documents.
Basic storage unit	Sorted string tables.	Serialized JSON documents.
Indexing	Cassandra supports secondary indexes and SASI to index by column or columns.	MongoDB indexes at a collection level and field level and offers multiple indexing options.
Query language	Cassandra uses CQL.	MongoDB uses MQL.
Concurrency	Cassandra achieves concurrency with row-level atomicity and turntable consistency.	MongoDB uses MVCC and document-level locking to ensure concurrency.
Availability	Cassandra has multiple master nodes, node partitioning, and key replication to offer high availability.	MongoDB uses a single primary node and multiple replica nodes. Combined with sharding, MongoDB provides high availability and scalability.
Partitioning	Consistent hashing algorithm, less control to users.	Users define sharding keys and have more control over partitioning.