LeverX Blog | Career Stories, SAP Insights & IT Tips

AWS for Developers: Building Mental Model of Modern Cloud Architecture

Written by Liza | Jun 24, 2026 7:00:00 PM

How EC2, S3, VPC, IAM, and Serverless Services Fit Together in Real-World Systems

Cloud technologies have become a standard part of modern software development. Yet many experienced developers still interact with cloud platforms through infrastructure teams, CI/CD pipelines, or managed services without fully understanding what happens behind the scenes.

In this article, we will share a practical overview of AWS architecture, focusing on the services developers encounter most often and, more importantly, how they work together.

With over six years at LeverX, I have helped design and operate cloud environments across numerous enterprise projects. This article is a practical guide for software engineers who want to build a stronger cloud foundation without getting lost in infrastructure minutiae.

Why Developers Need a Cloud Mental Model 

One of the biggest misconceptions about cloud computing is that it's fundamentally different from traditional infrastructure. Essentially, it’s just a computer.

Of course, AWS is much more than a single computer. The real innovation is that AWS abstracts away hardware procurement, maintenance, networking, power, cooling, and capacity planning, allowing teams to consume infrastructure on demand.

For developers, this means:

  • Faster product delivery
  • Easier experimentation
  • Reduced operational overhead
  • Better scalability
  • Pay-as-you-go economics

Instead of asking, "Which server should we buy?" teams ask, "Which service best fits this workload?"

Understanding the AWS Building Blocks 

AWS offers more than 200 services, but most applications rely on a relatively small set of core components. According to the masterclass, these components fall into five major categories:

Category Core Services
Compute EC2, ECS, Lambda, AWS Batch
Storage S3, EBS, Glacier
Networking VPC, Route 53, Internet Gateway
Databases RDS, DynamoDB
Security IAM, KMS, Security Groups

 

Understanding these categories is enough to comprehend the architecture of most modern cloud-native applications.

Compute: Choosing Between Servers, Containers, and Serverless

Amazon EC2: Traditional Virtual Servers 

Amazon EC2 (Elastic Compute Cloud) is the most familiar model for many developers.

You provision a virtual machine with a specific amount of CPU and memory, install your software stack, and manage the operating system yourself.

EC2 works particularly well when:

  • Workloads are predictable
  • Applications run continuously
  • Infrastructure requirements are stable

However, EC2 has a downside: you pay for allocated capacity even when your application isn't fully utilizing it.

Good Fit

  • Legacy applications
  • Long-running APIs
Enterprise software with stable traffic patterns

Amazon ECS: Container-Oriented Scaling

For most modern applications, LeverX teams increasingly favor container-based architectures.

Amazon ECS (Elastic Container Service) allows developers to deploy containers without managing complex orchestration platforms directly.

Benefits include:

  • Easier scaling
  • Better resource utilization
  • Smaller deployment units
  • Faster releases

Instead of running one large server, teams can run many smaller containers and scale them dynamically based on demand.

Good Fit

  • Microservices
  • APIs
  • Enterprise SaaS platforms
  • Event-driven systems

AWS Lambda and AWS Batch: Serverless Computing

Serverless services remove infrastructure management entirely.

With Lambda, developers upload code and AWS executes it when triggered.

Key advantages:

  • Scale-to-zero capability
  • Pay only for execution time
  • No server maintenance
  • Automatic scaling

According to Evgeniy, serverless architectures are particularly attractive for workloads that experience irregular traffic patterns.

Good Fit

  • API endpoints
  • Background processing
  • Event-driven workflows
  • AI inference triggers
  • Document processing

Storage: Not All Data Is Equal

A common mistake among developers is treating storage as a single problem. AWS provides different storage technologies optimized for different access patterns.

Amazon S3: The Foundation of Modern Cloud Applications

Amazon S3 (Simple Storage Service) is arguably the most important AWS service. S3 stores objects rather than files on disks.

Typical use cases:

  • Frontend assets
  • User uploads
  • Backups
  • Data lakes
  • Static websites

Many modern frontend applications can be deployed entirely from S3 when combined with CloudFront.

Why Developers Love S3

  • Nearly unlimited scalability
  • High durability
  • Simple API
  • Global availability

Amazon EBS: Cloud Hard Drives

Elastic Block Store behaves more like traditional disk storage. It is attached to EC2 instances and used for:

  • Operating systems
  • Application storage
  • Database storage

Think of EBS as the cloud equivalent of an SSD attached to a server.

Amazon Glacier: Long-Term Archives

Not every piece of data needs instant access. Glacier provides extremely inexpensive storage for:

  • Compliance archives
  • Historical backups
  • Long-term retention

The tradeoff is retrieval speed. Accessing archived data may take hours or even days.

Networking: The Layer Every Developer Should Understand

Many software engineers avoid networking until production issues force them to learn it. Yet AWS networking concepts are surprisingly intuitive.

VPC: Your Private Cloud Network

Amazon VPC (Virtual Private Cloud) is essentially your isolated network inside AWS. Within a VPC, you can create:

  • Public subnets
  • Private subnets
  • Routing rules
  • Security boundaries

A useful mental model is to think of a VPC as your company's office network, except it's entirely virtual.

Public vs Private Subnets

A key architectural principle to remember:

Databases should not be publicly accessible.

Typical architecture:

 

Only the API layer should communicate with databases directly.

Private subnets help enforce this separation.

Route 53: DNS for AWS

Route 53 handles domain management and DNS routing. Examples:

  • api.company.com
  • app.company.com
  • internal service discovery

It becomes the central naming system for cloud infrastructure.

Databases: Relational vs NoSQL

There are two major AWS database approaches.

Amazon RDS

RDS provides managed relational databases such as:

  • PostgreSQL
  • MySQL
  • MariaDB

Developers get:

  • Automated backups
  • Failover support
  • Monitoring
  • Managed updates

without managing database servers manually.

Use RDS When

  • Relationships matter
  • Transactions are important
  • Data models are structured

DynamoDB

DynamoDB is AWS's managed NoSQL database. Advantages:

  • Extremely fast performance
  • Massive scalability
  • Low operational overhead

Use DynamoDB When

  • Scale is unpredictable
  • Workloads are key-value based
  • Millisecond latency is critical

As Evgeniy explained, DynamoDB offers flexibility that traditional relational databases cannot easily provide.

Security: The Services Developers Can't Ignore

A key thing to remember is that security is everyone's responsibility. AWS follows a shared responsibility model:

  • AWS secures the infrastructure.
  • Customers secure applications and data.

IAM: The Heart of AWS Security

Identity and Access Management (IAM) controls:

  • Human access
  • Service permissions
  • Cross-account access
  • Automation credentials

For developers, IAM becomes critical when services need to interact securely. Examples:

  • CloudFront accessing S3
  • Lambda accessing DynamoDB
  • CI/CD pipelines deploying applications

The principle of least privilege should guide all IAM policies.

Security Groups

Security Groups function as virtual firewalls. Common rules include:

  • Allow frontend → backend
  • Allow backend → database
  • Deny everything else

This simple pattern dramatically reduces attack surfaces.

KMS

AWS Key Management Service handles encryption keys. Use cases:

  • HTTPS certificates
  • Data encryption
  • Secrets protection
  • Compliance requirements

For organizations working in regulated industries, KMS often becomes a foundational security component.

What a Modern AWS Architecture Actually Looks Like

One of the most valuable parts of the session was the architecture walkthrough.

A typical production web application might look like this:

Frontend

  • S3
  • CloudFront CDN

API Layer

  • AWS Batch or Lambda
  • ECS containers

Data Layer

  • Amazon RDS

Security Layer

  • IAM
  • Security Groups

Networking Layer

  • VPC
  • Private subnets
  • NAT Gateway
  • Route 53

The result is an architecture that is scalable, fault tolerant, and secure by design.

AWS Best Practices Every Developer Should Know

According to the masterclass, several practices consistently separate successful cloud projects from problematic ones.

Design for Scaling

Use:

  • Auto Scaling
  • Load Balancers
  • Containers
  • Serverless services

instead of relying on a single large server.

Deploy Across Multiple Availability Zones

Availability Zones are independent data centers within an AWS Region.

Distributing workloads across zones improves resilience against outages and infrastructure failures.

Encrypt Everything Important

Especially for:

  • Financial systems
  • Healthcare platforms
  • Enterprise applications

Use HTTPS and managed encryption services whenever possible.

Monitor Costs Early

One of the most practical recommendations from the session was to actively use:

  • AWS Cost Explorer
  • Budget alerts
  • Spot Instances when appropriate

Cloud costs rarely become a problem overnight—but they often become expensive surprises when left unmonitored.

How AI Is Changing AWS Architectures

During the Q&A, participants asked about AI adoption in enterprise projects.

According to Evgeniy, AI services are increasingly becoming part of production environments, not just experiments.

Within AWS, Bedrock provides access to multiple foundation models while preserving AWS-native governance and security controls. For organizations already operating within AWS ecosystems, this can simplify compliance, networking, and permission management.

For developers, this means future architectures will increasingly include:

  • LLM-powered APIs
  • Document intelligence services
  • AI-assisted search
  • Automated content processing

all integrated alongside traditional application components.

Learning AWS Beyond the Basics

While we cannot teach you every AWS service within this article, what we attempted to do was to help you understand the fundamental building blocks and how they connect.

Once that mental model is in place, learning advanced topics such as Kubernetes, Infrastructure as Code, CI/CD pipelines, AI services, or serverless architectures becomes significantly easier.

At LeverX, knowledge sharing is an important part of engineering culture. Internal communities, technical masterclasses, and expert-led sessions help developers continuously expand their skills across cloud, DevOps, AI, SAP, and enterprise technologies.

Final Thoughts

AWS can seem overwhelming when viewed as a catalog of hundreds of services.

But when you reduce it to a few core building blocks—compute, storage, networking, databases, and security—the platform becomes much easier to understand.

For experienced software developers, mastering this mental model is often the fastest path toward designing better systems, collaborating more effectively with DevOps teams, and building cloud-native applications with confidence.

And remember: cloud architecture isn't magic. It's simply a set of well-designed building blocks that, when combined correctly, can power almost any modern application.