Lecture Notes: NoSQL Databases

Introduction to NoSQL Databases

• Popularity: Widely used by large companies to manage extensive data and numerous queries per second.
• Definition: A type of database that doesn't rely on traditional relational models.

Limitations of Relational Databases

• Examples: MySQL, MariaDB, SQL Server.
• Characteristics:
  • Store relational data efficiently.
  • Tight organization linking tables (e.g., customers, orders, products).
• Scalability Issues:
  • Difficult to scale due to maintenance of relationships.
  • Capable of vertical scaling but limited in horizontal scaling.

Advantages of NoSQL Databases

• Scalability:
  • Scale both vertically and horizontally.
  • Horizontal scaling compared to adding more buildings (more flexible).
• Data Structure:
  • Dissolve costly relationships (key-value stores).
  • Simplified design allows better scalability.
  • Partitions for workload distribution; each server responsible for parts of database.
• Partitioning:
  • Items identified by primary key.
  • Uses hash function to partition data into keyspace.

NoSQL Database Characteristics

• Schemaless:
  • Allows varied structures for database items.
  • Advantageous for evolving applications and data structures.
• Eventual Consistency:
  • Data replication may cause temporary inconsistencies.
  • Trades some consistency for performance, but consistency options are available.

Comparison with Relational Databases

• Limitations:
  • Retrieval by primary key only, less efficient for complex queries.
  • Relational databases excel in complex data retrieval.
• Complementary Use: Both databases have distinct strengths and are expected to coexist.

Examples of NoSQL Databases

• Cloud Providers:
  • AWS DynamoDB, Google Cloud BigTable, Azure CosmosDB.
• Self-hosted Options:
  • Cassandra, Scylla, CouchDB, MongoDB.
• Case Study: Amazon's peak at 45 million requests/sec during Prime Day 2019.

Origin of the Term "NoSQL"

• Interpretations:
  • "Not only SQL": Some NoSQL databases understand SQL.
  • "Non-relational": Reflects inability to store relational data easily.

Conclusion

• Future Outlook: Expect continued coexistence of NoSQL and relational databases, each with unique advantages.
• Additional Note: Subscription prompt for further learning.