1.3 - App Development Lifecycle

Today we explore the complete application development lifecycle using a helpful analogy: moving into an apartment. This will help us understand the different layers of cloud application development.

What You’ll Learn

• The four key phases of cloud application development
• How infrastructure, architecture, and application code relate to each other
• The apartment analogy for understanding cloud resources
• Why automation and templates matter at each phase

Resources

Before diving in, review these resources:

1. 📘 Azure Well-Architected Framework - Best practices for cloud applications
2. 📘 Cloud Adoption Framework - Comprehensive guidance for Azure adoption
3. 📘 Infrastructure as Code Overview - Understanding IaC principles

The Apartment Analogy

Building a cloud application is like moving into and furnishing an apartment. Let’s explore each phase:

Phase 1: The Building (Cloud Platform)

Apartment Analogy: The apartment building itself—the structure, utilities, security, and shared amenities.

Cloud Equivalent: The Azure platform—regions, availability zones, networking backbone, and core services.

What You Don’t Do:

• You don’t build the building
• You don’t manage electricity/water infrastructure
• You don’t handle building security

What Azure Provides:

• Physical data centers
• Network infrastructure
• Security perimeter
• Redundancy and disaster recovery
• Compliance certifications

Your Responsibility: Choose the right region and understand available services.

Phase 2: The Floor Plan (Architecture)

Apartment Analogy: The apartment layout—number of rooms, how they connect, where utilities are placed.

Cloud Equivalent: Your application architecture—which services you’ll use, how they communicate, data flow patterns.

What You Design:

• Which Azure services to use (App Service, Functions, Cosmos DB, etc.)
• How services connect and communicate
• Data flow and storage patterns
• Security boundaries
• Scaling strategy

Example for Retail AI:

Frontend (App Service)
    ↓
API Gateway (API Management)
    ↓
Backend Services (Container Apps)
    ↓
AI Services (Azure OpenAI) + Data (Cosmos DB + AI Search)

Your Responsibility: Design an architecture that meets your requirements for performance, cost, security, and scalability.

Phase 3: The Infrastructure (Provisioning)

Apartment Analogy: Actually getting the apartment—signing the lease, getting keys, turning on utilities, setting up internet.

Cloud Equivalent: Provisioning Azure resources—creating resource groups, configuring services, setting up networking.

Three Approaches to Provisioning:

Manual Provisioning (Day 5)

Analogy: Going to each utility company, filling out forms, making phone calls, waiting for activation.

Cloud: Clicking through Azure Portal, filling forms, waiting for resources to deploy.

Pros: Visual, good for learning Cons: Time-consuming, error-prone, not repeatable

ARM/Bicep Templates (Day 6)

Analogy: Having a checklist and calling utility companies yourself, following instructions.

Cloud: Writing Infrastructure as Code that describes resources, running deployment commands.

Pros: Repeatable, version controlled, automated Cons: Still requires expertise, no application code included

AZD Templates (Day 7)

Analogy: Hiring a move-in coordinator who handles everything—utilities, internet, even basic furniture setup.

Cloud: Using azd up to provision infrastructure AND deploy application code in one command.

Pros: Complete solution, best practices built-in, repeatable Cons: Less flexibility (but can be customized)

Phase 4: The Furnishing (Application)

Apartment Analogy: Moving in your furniture, decorating, personalizing the space, making it functional.

Cloud Equivalent: Deploying application code, configuring services, loading data, testing functionality.

What You Customize:

• Application code and logic
• Configuration settings
• Data and content
• User interfaces
• Business rules

Example for Retail AI:

• Deploy chatbot application code
• Upload product catalog
• Configure AI prompts and behaviors
• Set up monitoring dashboards
• Test with sample conversations

The Complete Lifecycle

Let’s see how this applies to our retail AI scenario:

1. Platform Selection (Building)

✅ Choose Azure as cloud platform ✅ Select East US region for deployment ✅ Understand available AI services

2. Architecture Design (Floor Plan)

✅ Design RAG-based chatbot architecture ✅ Select services: Azure OpenAI, AI Search, App Service ✅ Plan data flow and security boundaries

3. Infrastructure Provisioning (Infrastructure)

📍 This is where we are now—deciding HOW to provision

• Manual? (Day 5)
• ARM Template? (Day 6)
• AZD Template? (Day 7)

4. Application Deployment (Furnishing)

⏳ Coming after infrastructure is ready

• Deploy chatbot code
• Configure AI models
• Load product data
• Test and iterate

Why This Matters

Understanding these phases helps you:

1. Communicate Clearly: “Are we discussing architecture or implementation?”
2. Choose Right Tools: Different tools for different phases
3. Plan Effectively: Know what needs to happen when
4. Troubleshoot Better: Identify which layer has issues

The Problem We’re Solving

Traditional approach challenges:

• Phase 2-3 Gap: Architecture designs don’t automatically become infrastructure
• Documentation Drift: Manual steps get forgotten or outdated
• Inconsistency: Dev, test, and prod environments differ
• Slow Onboarding: New team members struggle to set up environments

AZD templates bridge all phases, providing:

• Architecture documentation (in code)
• Automated provisioning (infrastructure)
• Application deployment (code)
• Best practices (built-in)

Ask Copilot

Explore these concepts further:

1. “What are the advantages and disadvantages of using Infrastructure as Code compared to manual resource provisioning in Azure Portal?”
2. “How does the separation of infrastructure and application layers improve maintainability and team collaboration?”
3. “What are the key considerations when choosing between Azure App Service, Container Apps, and Azure Functions for hosting an AI application?”

Related Resources

• Azure Application Architecture Guide
• Infrastructure as Code Best Practices
• Azure Landing Zones
• Cloud Design Patterns

---

Next: Day 4 - AI App Architecture

Tomorrow we’ll dive deep into AI application architecture patterns and learn to identify key components.