Dr.

Koushik Dutta
Associate Professor
Dept. of CSE, IEM, Kolkata

05-02-2026 1
 NoSQL!
• NoSQL databases are
currently a hot topic in
some parts of
computing, with over a
hundred
different NoSQL
databases.
• MongoDB, Cassandra,
Neo4j, Redis,
TimescaleDB etc.

05-02-2026 2
 Introduction
• A NoSQL database (sometimes called as Not Only
SQL) is a non-relational database that provides a
mechanism to store and retrieve data other than
the tabular relations used in relational databases.

• NoSQL databases
o Schema-free
o Support easy replication
o Support Sharding Replication
o Have simple API
o Eventually consistent
o Handle huge amounts of data

05-02-2026
Sharding 3
 Real Life Application
• Social Media:
o Social media platforms use NoSQL databases to manage their
enormous workloads and large datasets of user profiles, feeds,
and content.
• E-commerce:
o These platforms benefit from NoSQL's scalability and flexibility to
handle large and varied product catalogs.
• Content Management:
o NoSQL databases are ideal for managing diverse forms of
content, such as text, images, and videos.
• IoT and Real-time Applications:
o NoSQL databases are excellent for handling time-series data
generated by IoT devices and for providing real-time analytics.
• Recommendation Systems:
o Netflix uses Apache Cassandra, a column-family store NoSQL
database, to handle the constant stream of real-time data
required for its personalized movie and show recommendations

05-02-2026 4
 Objective of NoSQL
• The primary objective of a NoSQL database is to
have
o Simplicity of design,
o Horizontal scaling,
o Finer control over availability
o Handle Structured or Semi-structured or Unstructured Data

Types
of Data

Process of Scaling
05-02-2026 5
 Where does NoSQL come from?
• Non-relational DBMSs are not new, Before relational
databases, companies used hierarchical (IMS by IBM) and
network databases (CODASYL model)

• But NoSQL represents a new incarnation

o Handle massively scalable Internet applications -
ecommerce, and social media created massive,
unstructured, and fast-changing data.
o Working on distributed and parallel computing
environment

• RDBMS struggled to handle:

o Horizontal scaling (across many servers)
o Schema flexibility
o High-speed reads/writes
05-02-2026 6
 Continue…
Development
• 2000: Carlo Strozzi coined the term “NoSQL”, originally describing a
lightweight relational DB without SQL interface.
• 2006–2007:
o Google’s Bigtable (paper published in 2006) – inspired column-family
stores.
o Amazon’s Dynamo (2007) – inspired key-value stores.
• 2007-2009: Many open-source NoSQL projects emerged:
o Neo4j (2007) → Graph database
o Cassandra (2008, from Facebook) → Wide-column store
o MongoDB (2009) → Document store
o Redis (2009) → Key-value store
• 2010 onwards: NoSQL Adoption in Big Data & Cloud
o Tech giants like Google, Facebook, Amazon, Twitter, LinkedIn adopted
NoSQL for large-scale systems.
o Hadoop ecosystem and NoSQL databases became central to Big Data
analytics.
o Cloud providers (AWS, Azure, GCP) started offering managed NoSQL
services (e.g., Amazon DynamoDB, Azure Cosmos DB).
05-02-2026 7
 SQL vs. NoSQL
SQL NoSQL
Stands for Structured Query Language Stands for Not Only SQL

Relational database management system (RDBMS) Non-relational database management system

Suitable for structured data with predefined Suitable for unstructured and semi-structured
schema data
Data is stored in tables with columns and rows Data is stored in collections or documents
Follows ACID properties (Atomicity, Consistency,
Does not necessarily follow ACID properties
Isolation, Durability) for transaction management

Supports JOIN and complex queries Does not support JOIN and complex queries

Uses normalized data structure Uses denormalized data structure

Requires vertical scaling to handle large volumes of Horizontal scaling is possible to handle large
data volumes of data
Examples: MySQL, PostgreSQL, Oracle, SQL Examples: MongoDB, Cassandra, Couchbase,
Server, Microsoft SQL Server Amazon DynamoDB, Redis
05-02-2026 8
 Why Not ACID in NoSQL?

• In distributed, large-scale
systems (like social
networks, e-commerce,
IoT), strict ACID slows
performance.
• These systems need high
availability, consistency,
and partition tolerance
(from the CAP theorem).
• Follow BASE principle

05-02-2026 9
Types of NoSQL
• Document-Oriented Databases
o Store data in JSON/BSON/XML documents
o Ex: MongoDB, CouchDB
• Key-Value Stores
o Store data as key → value pairs
o Example: Redis, DynamoDB, Riak
• Column-Family Stores (Wide-Column)
o Store data in rows with flexible columns grouped into families
o Example: Cassandra, HBase, ScyllaDB
• Graph Databases
o Store entities as nodes and relationships as edges
o Example: Neo4j, Amazon Neptune, OrientDB
• Time-Series Databases (specialized)
o Optimized for time-stamped data (IoT, logs, metrics)
o Example: InfluxDB, TimescaleDB
05-02-2026 10
 Document
-oriented

Time Series

Key-value
store

Graph
Database

Column-Family Stores
05-02-2026 11
 Advantages & Disadvantages
of NoSQL
Pros:
• High scalability and availability
• Flexible schema (easy to modify data structures)
• Fast data retrieval for specific use cases
• Handles huge datasets and distributed computing

Cons:
• Lack of standardization (each system is different)
• Limited support for ACID transactions (though improving)
• Not always the best fit for applications needing complex
queries and joins (where RDBMS is better)
05-02-2026 12
 Which is better SQL or
NoSQL?
• The decision of which type of database to use - SQL or
NoSQL - will depend on the particular needs and
requirements of the project.
• For example, if you need a fast, scalable, and reliable
database for web applications then a NoSQL system
may be preferable.
• On the other hand, if your application requires complex
data queries and transactional support then an SQL
system may be the better choice.
• Ultimately, there is no one-size-fits-all solution - it all
comes down to what you need from your database and
which type of system can provide that in the most
efficient manner. It's best to research both options
thoroughly before making a decision.
05-02-2026 13
05-02-2026 14
 ACID in SQL
• ACID properties (Atomicity, Consistency, Isolation,
Durability) in a DBMS ensure that transactions are
processed reliably.
• Atomicity: A transaction is treated as a single,
indivisible unit. It must be completed in its entirety,
or not at all.
• Consistency: A transaction must bring the
database from one valid state to another.
• Isolation: The effect of concurrent transactions
should be the same as if they were executed one
after another (serially).
• Durability: Once a transaction has been
committed, its changes are permanent.

BACK 05-02-2026 15