System Design Concepts Lecture

Introduction

• Purpose: Understanding essential system design concepts for building scalable applications and passing system design interviews.
• Topics: Networking, API patterns, databases, etc.

Scaling Servers

Vertical Scaling

• Adding resources like RAM or upgrading CPU.
• Simple but limited.

Horizontal Scaling

• Adding server replicas to handle subsets of requests.
• Benefits: Increases redundancy and fault tolerance.
• Requires load balancing to distribute requests effectively.

Load Balancers

• Acts as a reverse proxy to distribute incoming requests to servers.
• Algorithms: Round robin, hashing request IDs, location-based routing.

Content Delivery Networks (CDNs)

• Serve static files (e.g., images, videos, HTML, CSS, JavaScript) from globally distributed servers.
• Push or pull basis for copying files from origin server.

Caching

• Creating copies of data for faster refetching.
• Levels: Disk, memory, CPU cache.

Networking Basics

• IP addresses: Unique identifiers for devices on a network.
• TCP/IP Suite: Includes TCP & UDP.
• TCP: Reliable protocol ensuring packet delivery.
• Application protocols built on TCP: HTTP, Websockets.

Domain Name System (DNS)

• Maps domain names to IP addresses using DNS queries.
• DNS A record for specifying IP addresses.
• OS caching to reduce repeated DNS queries.

HTTP and API Patterns

HTTP Protocol

• Follows client-server model.
• Request: Header (metadata) + Body (content).

REST APIs

• Stateless and consistent guidelines.
• Response Codes: 200 (success), 400 (client error), 500 (server error).

GraphQL

• Single request (query) to fetch multiple resources.
• Minimizes over-fetching data.

gRPC

• RPC framework, efficient for server-to-server communication.
• Uses protocol buffers for data serialization.
• gRPC Web allows browser usage.

Websockets

• Bi-directional communication for real-time applications like chat apps.
• Immediate message delivery without polling.

Databases

Relational Databases (SQL)

• Examples: MySQL, PostgreSQL.
• Structured data in rows and tables.
• ACID compliance: Atomicity, Consistency, Isolation, Durability.

NoSQL Databases

• Types: Key-value stores, document stores, graph databases.
• Not ACID compliant, easier to scale.

Database Scaling

Sharding

• Horizontal scaling by distributing data across machines using shard keys.

Replication

• Leader-follower: Leader handles writes; followers replicate data for reads.
• Leader-leader: Every replica can read/write but may result in data inconsistency.

CAP Theorem

• Trade-offs in a replicated database: Consistency vs. Availability during network partition.
• PACELC Theorem: More complete version of CAP.

Message Queues

• Durable storage and redundancy options.
• Use cases: Handling more data than can be processed immediately, decoupling app components.

Learning Resources

• System Design for Beginners Course on neetcode.io.
• System Design Interview Course on neetcode.io.

Conclusion

• Importance: Preparing for scalable app development and system design interviews.
• Thanks for support.