Lesson 2: Databases

SQL vs. NoSQL

SQL (Relational Databases)

• Structure: Structured data with predefined schemas (Tables, Rows, Columns).
• Query Language: SQL (Structured Query Language).
• ACID Compliance: Strong guarantees for Atomicity, Consistency, Isolation, Durability.
• Examples: MySQL, PostgreSQL, Oracle.
• Best for: Complex queries, financial transactions.

NoSQL (Non-Relational Databases)

• Structure: Flexible schemas (Key-Value, Document, Graph, Column-Family).
• Scalability: Designed for horizontal scaling.
• Examples: MongoDB (Document), Redis (Key-Value), Cassandra (Column).
• Best for: Rapidly changing data, massive scale, unstructured data.

Scaling Databases

Replication

Copying data to multiple servers.

• Master-Slave: Writes go to Master, Reads go to Slaves. Good for read-heavy systems.
• Master-Master: Writes can go to any node. Complex conflict resolution needed.

Sharding

Partitioning data across multiple servers (e.g., Users A-M on Server 1, N-Z on Server 2).

• Pros: Handles massive data volumes.
• Cons: Complex joins, rebalancing data is hard.

---

🛠️ Sruja Perspective: Modeling Databases

Sruja allows you to define the type of database and its role in the system.

import { * } from 'sruja.ai/stdlib'


UserDB = container "User Database" {
    technology "PostgreSQL"
    tags ["relational", "primary"]
    description "Stores user profiles and authentication data."
}

SessionStore = container "Session Cache" {
    technology "Redis"
    tags ["key-value", "cache"]
    description "Stores active user sessions for fast access."
}

view index {
include *
}