NoSql Database: Cassandra, Mongo, Redis Database Comparison

1. What is NoSql database？

NoSql（not only Sql）database can Extended non-relational database that persistently stores semi-structured or unstructured data and has a flexible schema. These databases support multiple data models – such as key-value, document, column family, graph-based, in-memory, etc. – for management and access data. NoSql databases are best used for building modern applications that require big data and high volume of requests Highly scalable Low latency High performance and flexible data model High performance and flexible data model ;To provide an excellent customer experience.

In this article，We will discuss about NoSql database，Cassandra，Introduction of Mongo and Redis database，and when to use these NoSql databases for better performance .

2. Features of NoSql database

2.1 Multiple model support

NoSql database is ideal for persistent management and access to semi-structured and unstructured ized data.

2.2 Open Source

Most NoSql databases are open source. Most cloud providers offer these databases as a managed service by managing automatic scaling, patch updates, etc. in the background.

2.3 Scalability

NoSql databases can be scaled horizontally by adding more servers to the cluster and are unlike relational databases which can be scaled vertically. xff0c;The cluster can be spread across multiple geographical locations “regions”.

2.4 Low Latency

Latency is very low since data is replicated to multiple nodes on the database cluster， The trade-off between latency and consistency is important for web and mobile software applications. No matter which replication method is adopted – there will be a trade-off between consistency and latency.

2.5 Flexible Schema

NoSql database supports a flexible data model with eventual consistency ， and is inherently schema-less. This makes NoSql database suitable for efficient storage of semi-structured and unstructured data.

2.6 High Performance ，Available and Fault Tolerant

In NoSql database，data is replicated to multiple nodes in the cluster as well as to other regions in the cluster node. This feature makes NoSql database highly available and fault-tolerant. No Sql database is highly optimized for data models like document ，key-value，column-family，graph and their access patterns，these models provide higher performance.

3. Cassandra

Apache Cassandra is an open source ， distributed ， horizontally scalable ， highly available ， fault-tolerant and Wide column NoSql database. It is written in Java and is a column family storage database. All nodes in a Cassandra cluster are peers，and there is no master-slave paradigm in Cassandra. This makes cassandra highly available ， fault tolerant and has no single point of failure. Cassandra clusters can scale horizontally and be distributed across multiple data centers.

Writing in cassandra is very fast – because it doesn’t search first and then write. The data is first written to the commit log ， and then the data is carefully reflected into the table using the cassandra algorithm. In cassandra ‘model the data around the query’ i.e. first identify the application query’ and then model the data around it.

4. MongoDB

MongoDB is an open source, cross-platform, highly document-oriented document written in C Available – Scalable and flexible NoSql database. It works on collections and documents and provides high availability via replica sets.

MongoDB uses JSON-like documents – these documents can have a variety of structures. Since there are fewer schemas, there is no need to create the document structure before creating the document. MongoDB uses MongoDB QL “query language” to access data stored in MongoDB. MongoDB has very powerful aggregation capabilities and an expressive aggregation framework.

5. Redis（Remote Dictionary Server）

Redis is an open source and scalable data store，that can be used as a database，cache and message broker. It is written in ANSI C. Redis is an in-memory data store that can persist its state to disk and can restore its state even if the Redis node is restarted. Its memory storage makes it super fast.

6. Cassandra vs MongoDB vs Redis

• Cassandra stores data in Column-Family structure ， while MongoDB stores data in JSON document format.
• In Cassandra ， it is not recommended to use secondary indexes as they can reduce performance. In MongoDB it’s better to use indexes to avoid searching all documents to find the required document and get better performance.
• Cassandra is ideal for high write throughput – but if your application requires high read concurrency – use MongoDB.
• Cassandra has no master node，all nodes are peers，whereas in MongoDB，there is only one master node.
• Cassandra finally�The data written is replicated to the number of nodes in the cluster specified in the replication factor ， and the number of nodes in the cluster in different regions. MongoDB requires some setup for replication. You can set up a secondary database that will be automatically selected if the primary database fails. In MongoDB ‘reads are first committed to the primary replica’ and then replicated to the secondary replica.
• In these databases you can set a TTL time-to-live on each record so records can be automatically evicted after the TTL expires.
• Redis is a key-value data store ， very effectively used as a cache to improve application performance.
• Extending Cassandra and Mongo is much simpler than Redis.
• In Redis, the size of data storage cannot exceed the total memory space on the system, that is, RAM plus swap space. There is no inherent limit on the size of a Mongo database.
• Cassandra, MongoDB and Redis databases can be clustered to achieve high availability, backups and increase the overall size of the data store.
• If your application requires aggregation, use MongoDB. If your application requires key-value temporary storage – use Redis. If your application requires a wide column store with high write throughput that scales easily – use Cassandra.

7. When to use which NoSql database？

Following are some use cases，where different NoSql databases are most suitable and provide better performance.

Select the following use cases for Cassandra：

• Linearly scalable ， High availability ， Fault tolerance
• Multi-data center deployment
• High write throughput，but fewer reads.
• You want to build a responsive Swift reporting system on top of stored data
• Real-time data analysis
• Your application does not require ACID properties in the database
• Your application needs to integrate with Hadoop，HBase，Spark

Choose MongoDB for the following use cases：

• Instant scaling
• Document-based storage
• High read concurrency
• Cached real-time analytics
• Content management
• High write payload，i.e. large document size（maximum 16MB）
• Very useful in rapid prototyping
• Suitable for storing large text ，videos，images，media files etc.

Select Redis DB：

• Cache larger payloads to improve application performance
• Permanently Cache to disk ， needs to be restored after reboot
• Key-value store
• Requires very high performance
• Temporary data store ， e.g. User sessions
• can be used as messaging queues using their publish/subscribe model

8. Cassandra vs Mongo vs Redis DB – Summary

Understand Different NoSql databases are crucial to choosing the one that suits your application needs. Choose the right NoSql database based on your application use case. If your data has too many relationships and requires ACID properties, a NoSql database is not a good choice. To improve application performance ， use Redis as a cache as it has in-memory storage. Use MongoDB for content management and document type storage needs. Use Cassandra for highly available wide column storage bins. Choosing the right database directly affects the performance of your application.

Translated from: https://www.javacodegeeks.com/2019/02/nosql-databases-cassandra-vs-mongo-vs-redis-db-comparison.html