Relational DB

Relational databases are the default choices for structured data storage. One of the most extraordinary powers of the relational database is its abstractions of ACID transactions and related programming semantics.


  1. Data Consistency: SQL databases ensure data consistency across multiple tables through the use of transactions, which ensure that changes made to one table are reflected in all related tables[1].
  2. Backup and Recovery: SQL databases have built-in backup and recovery tools that help recover data in case of system crashes or other disasters[1].
  3. High-Speed Query Processing: SQL databases are known for their high-speed concurrent query processing capabilities[7].
  4. Data Integrity: SQL databases abide by ACID principles, ensuring data integrity[10].
  5. Support: All SQL databases receive great support from their vendors. Also, many independent consultants can help you with SQL databases on a large scale [3].
  6. Reduced redundancy: Data related to an entity is organized in one table, while relevant information is stored in other tables linked through foreign keys. This is called normalization and helps eliminate inconsistent dependencies.


  1. Lack of Real-Time Analytics: SQL databases are designed for batch processing and do not support real-time analytics, which can disadvantage applications requiring real-time data processing[1].
  2. Limited Query Performance: SQL databases may have limited query performance when dealing with large datasets, as queries may take longer to process than in-memory databases[1].
  3. Complexity: SQL databases can be complex to set up and manage, requiring skilled database administrators to ensure optimal performance and maintain data integrity[1].
  4. Limited Flexibility: SQL databases are less flexible than NoSQL databases when handling unstructured or semi-structured data[1]. However, popular SQL DBs like MySQL and PostgreSQL support JSON datatype columns, which can store unstructured data up to a certain extent.
  5. Resource-Intensive Scaling: In SQL databases, vertical scaling involves investing in more powerful hardware, a costly and time-intensive process[7].


There are various types of NoSQL databases. These databases are used in applications that require a large volume of semi-structured and unstructured data, low latency, and flexible data models. Relaxing some of the data consistency restrictions of other databases can achieve this. Types of NoSQL DBs:

  1. Key-Value store -> DynamoDB, Redis
    1. Key-value databases are efficient for session-oriented web applications that store users' data during a session in memory or a database.
    2. User data is stored with a unique ID in a key-value database for easy access and storage.
  2. Document store -> MongoDB
    1. Document databases are ideal for storing unstructured catalog data, such as JSON files or other complex hierarchical data structures.
    2. In e-commerce, a product has too many attributes to store in a relational database, impacting the reading performance.
    3. A document database efficiently stores attributes in a single file for easy management and faster reading. It's ideal for content management apps like blogs and video platforms.
  3. Columnar store -> Cassandra
    1. Columnar databases are efficient for analytics queries, reducing disk I/O and data load requirements.
    2. Financial applications often require summing transactions over time. Columnar databases speed up the process by reading only the monetary column, ignoring customer attributes.
  4. Graph store -> NeoDB
    1. Graph databases aid social apps by providing user activity insights.
    2. Graph databases are a type of database that stores data in a way that enables analyses and decisions based on the relationships between different entities.
    3. Graph databases are versatile and can be used for various applications, such as data regulation, machine learning research, and financial services.


  1. Scalability: NoSQL databases are highly scalable and can be modified to meet your applications' unique scaling needs. They can be scaled horizontally instead of vertically, which provides a clear advantage over SQL databases[2].
  2. Simplicity: NoSQL databases require less structuring than SQL, so every item is self-contained and independent[2].
  3. Flexible Data Model: NoSQL databases easily store and combine any data, both structured and unstructured[8].
  4. Elastic Scalability: NoSQL databases scale on low-cost, commodity hardware, allowing for almost unlimited growth[8].
  5. High Performance: NoSQL databases are built for outstanding performance, measured in terms of both throughput and latency[8].


  1. Less Mature: NoSQL is less mature than SQL, disadvantaging projects requiring experience[5].
  2. Requires Multiple Databases: If you’re using NoSQL, you’ll probably end up using multiple types of databases and data models to fill all the niches and needs of your application[5].
  3. Lack of Complex Query Functionality: NoSQL databases are more flexible when storing a wide variety of data structures, but they need more complex query functionality found in SQL[2].

In conclusion, the choice between SQL and NoSQL depends on the specific requirements of your application. SQL is generally a good choice for applications that require complex queries, transaction reliability, and structured data. At the same time, NoSQL is better suited for applications that require scalability, flexibility, and high performance.

Thoughts 🤔 by Soumendra Kumar Sahoo is licensed under CC BY 4.0