Architecture & System Design Guide

Master software architecture, system design, and building scalable distributed systems.

Quick Start

Monolithic Architecture

┌─────────────────────────────────────┐
│          Application Layer          │
├─────────────────────────────────────┤
│  Authentication │ API │ Business    │
├─────────────────────────────────────┤
│          Database Layer             │
└─────────────────────────────────────┘

Microservices Architecture

┌──────────────┐  ┌──────────────┐  ┌──────────────┐
│ Auth Service │  │ Order Service │  │ User Service │
└──────────────┘  └──────────────┘  └──────────────┘
        │                │                  │
        └────────────────┴──────────────────┘
                    │
            ┌───────────────────┐
            │  API Gateway      │
            └───────────────────┘
            │
        ┌───┴───┐
        │Client │
        └───────┘

Design Patterns

Creational Patterns

// Singleton Pattern
class Database {
  static instance = null;

  static getInstance() {
    if (!Database.instance) {
      Database.instance = new Database();
    }
    return Database.instance;
  }
}

// Factory Pattern
class UserFactory {
  createUser(type) {
    if (type === 'admin') return new AdminUser();
    if (type === 'regular') return new RegularUser();
  }
}

// Builder Pattern
class UserBuilder {
  constructor(name) {
    this.name = name;
  }

  setEmail(email) {
    this.email = email;
    return this;
  }

  setRole(role) {
    this.role = role;
    return this;
  }

  build() {
    return new User(this);
  }
}

Structural Patterns

// Adapter Pattern
class OldAPI {
  getData() { return { data: [] }; }
}

class APIAdapter {
  constructor(oldAPI) {
    this.oldAPI = oldAPI;
  }

  getAPIv2Format() {
    return { response: this.oldAPI.getData() };
  }
}

// Decorator Pattern
class Coffee {
  cost() { return 5; }
}

class MilkDecorator {
  constructor(coffee) {
    this.coffee = coffee;
  }

  cost() { return this.coffee.cost() + 2; }
}

Behavioral Patterns

// Observer Pattern
class EventEmitter {
  constructor() {
    this.listeners = {};
  }

  on(event, callback) {
    this.listeners[event] = this.listeners[event] || [];
    this.listeners[event].push(callback);
  }

  emit(event, data) {
    if (this.listeners[event]) {
      this.listeners[event].forEach(cb => cb(data));
    }
  }
}

// Strategy Pattern
class PaymentProcessor {
  constructor(strategy) {
    this.strategy = strategy;
  }

  process(amount) {
    return this.strategy.process(amount);
  }
}

class CreditCardStrategy {
  process(amount) { return `Processing credit card: $${amount}`; }
}

class PayPalStrategy {
  process(amount) { return `Processing PayPal: $${amount}`; }
}

Architectural Patterns

Layered Architecture (N-Tier)

┌─────────────────────┐
│  Presentation Layer │
├─────────────────────┤
│  Business Logic     │
├─────────────────────┤
│  Persistence Layer  │
├─────────────────────┤
│  Database           │
└─────────────────────┘

• Simple, familiar structure
• Easy to organize code
• Can become monolithic

Event-Driven Architecture

┌──────────────────┐
│  Event Producer  │
└────────┬─────────┘
         │ (Message Broker)
     ┌───┴────┬────────┐
┌────▼────┐ ┌▼─────┐ ┌▼──────┐
│Consumer1│ │Cons2 │ │Cons3  │
└─────────┘ └──────┘ └───────┘

• Decoupled systems
• Real-time processing
• Complexity in event management

Microservices Architecture

• Advantages: Independent scaling, language flexibility
• Challenges: Distributed systems, testing, DevOps
• Best For: Large teams, complex domains

Scalability Concepts

Horizontal vs Vertical Scaling

Vertical Scaling:          Horizontal Scaling:
┌──────────┐              ┌──────────┐
│Server    │              │Server 1  │
│16GB RAM  │              │4GB RAM   │
│4 CPUs    │   ──────>    ├──────────┤
└──────────┘              │Server 2  │
                          │4GB RAM   │
                          ├──────────┤
                          │Server 3  │
                          │4GB RAM   │
                          └──────────┘

Load Balancing

• Round-robin: Distribute equally
• Least connections: Route to least busy
• IP hash: Sticky sessions
• Weighted: Prefer certain servers

Caching Strategies

Application -> L1 Cache (In-memory)
            -> L2 Cache (Redis)
            -> L3 Cache (CDN)
            -> Database

Database Scaling

• Read Replicas: Multiple read-only copies
• Sharding: Horizontal partitioning
• Replication: Master-slave, multi-master

Distributed Systems Concepts

CAP Theorem

Choose 2 of 3:

• Consistency: All nodes see same data
• Availability: System always responds
• Partition Tolerance: Survives network failures

Consensus Algorithms

• Paxos: Complex but proven
• RAFT: More understandable
• Gossip: Eventual consistency

Distributed Transactions

• Saga Pattern: Long-running transactions
• Two-Phase Commit: ACID across systems
• Event Sourcing: Immutable event log

API Design

REST Principles

GET    /api/users              # List
GET    /api/users/:id          # Read
POST   /api/users              # Create
PUT    /api/users/:id          # Update
DELETE /api/users/:id          # Delete

GraphQL Basics

query {
  user(id: 1) {
    name
    email
    posts {
      title
      content
    }
  }
}

gRPC

• Protocol buffers for serialization
• HTTP/2 for efficiency
• Strong typing and contract

Data Modeling

Normalization

• 1NF: Atomic values
• 2NF: Remove partial dependencies
• 3NF: Remove transitive dependencies
• BCNF: Boyce-Codd Normal Form

Denormalization Tradeoffs

• Faster reads, slower writes
• Redundant data
• Consistency challenges

System Design Interview Patterns

Approach

1. Clarify Requirements: Functional and non-functional
2. Estimate Scale: Users, QPS, storage
3. Design High-Level: Component diagram
4. Dive Deep: Pick critical components
5. Discuss Tradeoffs: Pros and cons

Common Systems to Design

• Twitter Feed: Caching, fan-out, databases
• Video Streaming: CDN, encoding, playback
• Messenger: Real-time, consistency, scalability
• Ride-sharing: Geo-location, matching, payments

Architecture Decision Records (ADR)

# ADR 001: Use PostgreSQL for main database

## Context
We need a relational database for our data model.

## Decision
We will use PostgreSQL instead of MySQL.

## Consequences
- Excellent JSON support
- Better concurrency handling
- Steeper learning curve than MySQL

Security Architecture

Defense in Depth

┌──────────────────────────┐
│ Firewall                 │
├──────────────────────────┤
│ API Gateway / WAF        │
├──────────────────────────┤
│ Application Security     │
├──────────────────────────┤
│ Database Encryption      │
├──────────────────────────┤
│ Physical Security        │
└──────────────────────────┘

Performance Architecture

Optimization Levels

1. Database: Indexing, query optimization
2. Application: Caching, algorithms
3. Infrastructure: CDN, regional servers
4. Protocol: HTTP/2, compression

Learning Resources

Books

• "Designing Data-Intensive Applications" by Martin Kleppmann
• "Building Microservices" by Sam Newman
• "System Design Interview" by Alex Xu

Online Resources

• System Design Primer GitHub
• Educative System Design courses
• Interview.io system design prep

Practice

• Design familiar systems (Uber, Instagram)
• Whiteboard with peers
• Read production architecture docs

Tools

Diagramming

• Draw.io: Free diagramming
• Lucidchart: Professional diagrams
• Excalidraw: Quick sketches
• Miro: Collaborative whiteboarding

Architecture Analysis

• PlantUML: Diagram as code
• Structurizr: Architecture as code
• ArchiMate: Standard notation

Roadmap.sh Reference: https://roadmap.sh/system-design

---

Status: ✅ Production Ready | SASMP: v1.3.0 | Bonded Agent: 06-architecture-specialist