数据库与存储选项

Every application needs to store information. Think about user profiles, product catalogs, or posts on a social media app. This data needs to be saved reliably and retrieved quickly.

In system design, choosing the right storage solution is crucial for performance, scalability, and reliability.

Persistent data is information that remains stored even after the application or computer system that created it has been shut down or restarted.

• It's different from data held in memory (like RAM), which is lost when power is off.
• Databases are the most common way to store persistent data in backend systems.

Databases are specialized systems designed for efficient storage, organization, and retrieval of persistent data. They offer several advantages:

• Organization: Structure data logically.
• Efficiency: Fast searching and updates.
• Integrity: Rules to keep data consistent and valid.
• Concurrency: Handle many users accessing data at once.

Relational databases are perhaps the most traditional type. They store data in tables, which are like spreadsheets with rows and columns.

• Each table represents an entity (e.g., "Users", "Products").
• Tables are linked together by relationships, using common columns.
• They use SQL (Structured Query Language) for managing and querying data.

Examples include MySQL, PostgreSQL, Oracle, and SQL Server.

Imagine a simple user table in a relational database:

TABLE Users:
id (PRIMARY KEY)
username (VARCHAR)
email (VARCHAR)
created_at (DATETIME)

Each row would be a user, and columns define their attributes. Relationships can link users to their orders, for instance.

NoSQL databases (short for "Not only SQL") offer more flexibility than relational databases. They don't use the traditional table-based structure.

NoSQL databases are often chosen for their:

• Scalability: Easier to scale horizontally across many servers.
• Flexibility: Handle unstructured or semi-structured data.
• Performance: Optimized for specific data models.

Examples include MongoDB, Cassandra, Redis, and Neo4j.

One common NoSQL type is a document database. It stores data in flexible, JSON-like documents. Here's how a user might be stored:

{
"_id": "user123",
"username": "coddykit_user",
"email": "user@example.com",
"profile": {
"firstName": "Coddy",
"lastName": "Kit",
"bio": "Learning backend systems"
},
"interests": ["tech", "coding", "design"]
}

While databases are primary for structured data, systems often need other storage types:

• Object Storage: Stores large, unstructured data files (images, videos, backups). Think Amazon S3.
• File Storage: Traditional file systems, often network-attached (NFS) for shared access.
• Block Storage: Provides raw storage volumes that act like a hard drive, often used by virtual machines.

Choosing the right storage depends on your specific needs:

• Data Structure: Is your data highly structured (like financial records) or flexible (like user profiles with varying fields)?
• Scalability: How much data will you store, and how many users will access it?
• Consistency: How critical is it that all users see the absolute latest data at all times?

This decision is a cornerstone of system design.

Great job! In this lesson, we explored the fundamentals of data storage in backend systems.

• We learned about persistent data and why databases are essential.
• We introduced Relational (SQL) databases, known for their structured tables and strong consistency.
• We also covered Non-Relational (NoSQL) databases, offering flexibility and horizontal scalability.
• Finally, we touched upon other storage types like object, file, and block storage.

Understanding these options is key to designing robust systems!