DevOpsAWSCloud

AWS Solutions Architect Patterns: Well-Architected Framework & Designs

Master AWS solutions architect patterns — the Well-Architected Framework 6 pillars, three-tier web architecture, serverless patterns, and event-driven design for SAA-C03.

TT
Sarah Mitchell
6 min read
AWS Solutions Architect Patterns: Well-Architected Framework & Designs

Knowing individual AWS services is not the same as knowing how to architect with them. A Solutions Architect combines services into systems that are secure, reliable, cost-efficient, and operationally excellent. AWS has codified the principles behind great cloud architecture in the Well-Architected Framework — and the SAA-C03 exam tests your ability to apply them.


The AWS Well-Architected Framework

The Well-Architected Framework is a set of architectural best practices and design principles organised into six pillars. Use it to evaluate and improve your architectures.

PillarCore Question
Operational ExcellenceCan you run, monitor, and improve your systems effectively?
SecurityAre you protecting data, systems, and assets?
ReliabilityCan your system recover from failures and meet demand?
Performance EfficiencyAre you using resources efficiently as requirements change?
Cost OptimisationAre you paying only for what you use?
SustainabilityAre you minimising environmental impact?

Pillar 1: Operational Excellence

Define everything as code (CloudFormation, CDK, Terraform), make frequent small reversible changes, anticipate failure with game days and runbooks, and learn from all operational events through blameless post-mortems. AWS services: CloudFormation, AWS CDK, Systems Manager, CloudWatch, X-Ray, Config.

Pillar 2: Security

Implement a strong identity foundation with least privilege and MFA, enable traceability (CloudTrail, VPC Flow Logs), apply security at every layer (defence-in-depth), protect data in transit (TLS) and at rest (KMS), and automate security responses with GuardDuty and Security Hub. AWS services: IAM, KMS, CloudTrail, GuardDuty, Security Hub, WAF, Shield, Macie.

Pillar 3: Reliability

Design for automatic failure recovery with health checks and multi-AZ, test recovery procedures by actually restoring backups, scale horizontally, stop guessing capacity with Auto Scaling, and manage change through automation. AWS services: Route 53, ELB, Auto Scaling, RDS Multi-AZ, S3, CloudFormation.

Pillar 4: Performance Efficiency

Use the right tool for the job, go global in minutes with CloudFront and Global Accelerator, use serverless to remove capacity management overhead, and experiment often. AWS services: CloudFront, Lambda, ECS Fargate, DynamoDB, ElastiCache, Global Accelerator.

Pillar 5: Cost Optimisation

Adopt a consumption model (pay only for what you use), measure overall efficiency (cost per transaction), stop spending on undifferentiated heavy lifting by using managed services, and analyse expenditure with tags. AWS services: Cost Explorer, AWS Budgets, Compute Savings Plans, Reserved Instances, Spot Instances, S3 Lifecycle policies, Trusted Advisor.

Pillar 6: Sustainability

Right-size infrastructure, maximise utilisation with Spot and serverless, use lifecycle policies to move data to lower-energy storage classes, and measure with the Customer Carbon Footprint Tool.


Three-Tier Web Architecture

The three-tier architecture is the most commonly tested pattern in the SAA-C03 exam. It separates concerns into three distinct layers, each scalable independently.

text
┌─────────────────────────────────────────────────────────┐
│                       Internet                          │
└───────────────────────────┬─────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│              Presentation Tier (Web)                    │
│  CloudFront CDN → ALB → EC2/ECS (public subnets)        │
│  Serves static assets (S3) + dynamic HTML               │
└───────────────────────────┬─────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│              Application Tier (API)                     │
│  ALB → EC2/ECS/Lambda (private subnets)                 │
│  Business logic, API endpoints                          │
└───────────────────────────┬─────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│              Data Tier (Database)                       │
│  RDS Multi-AZ (private subnets)                         │
│  ElastiCache Redis (session + query cache)              │
└─────────────────────────────────────────────────────────┘

Security group rules for three-tier: Web tier SG allows 80/443 from internet; App tier SG allows 8080 from Web tier SG only; DB tier SG allows 5432 from App tier SG only. No direct internet access to the application or database tier.


Serverless Architecture Patterns

Serverless means no servers to provision, patch, or scale. You write code; AWS manages the infrastructure.

Core Serverless Services

ServicePurpose
LambdaRun code in response to events — no servers
API GatewayHTTP API frontend for Lambda functions
DynamoDBServerless NoSQL database
S3Serverless object storage
SQSServerless message queue
SNSServerless pub/sub notifications
EventBridgeServerless event bus
Step FunctionsServerless workflow orchestration

Serverless REST API Pattern

text
Client → API Gateway → Lambda → DynamoDB
                     ↓
                 S3 (file storage)
                 SES (email)

Lambda key concepts: maximum execution time is 15 minutes, maximum memory is 10 GB, billed per 1ms of execution, cold starts (~100ms–1s) on first invocation after idle, and concurrency scales to thousands of concurrent executions automatically.

Event-Driven Architecture

In event-driven architecture, components communicate by emitting and consuming events — they are decoupled and do not call each other directly.

text
Order Service  →  SQS Queue  →  Lambda (process payment)
               →  SNS Topic  →  Lambda (send confirmation email)
                             →  Lambda (update inventory)
                             →  SQS (notify warehouse)

SQS vs SNS vs EventBridge

SQSSNSEventBridge
PatternPoint-to-point queuePub/sub fan-outEvent routing bus
ConsumersOne consumer per messageMany subscribersMany targets via rules
FilteringNoMessage filteringRich content-based rules
RetentionUp to 14 daysNo retentionNo retention

Disaster Recovery Strategies

AWS defines four DR strategies, ranked from cheapest/slowest to most expensive/fastest:

StrategyDescriptionRTORPOCost
Backup & RestoreBack up data to S3; restore to new infrastructure when disaster occursHoursHoursLowest
Pilot LightMinimal replica of core infrastructure running at all times; scale up when needed10s of minutesMinutesLow
Warm StandbyScaled-down but fully functional replica; scale up to full production during disasterMinutesSecondsMedium
Multi-Site Active/ActiveFull production running in two+ regions simultaneously~0~0Highest

Choose DR strategy based on your RTO/RPO requirements vs. budget. Most organisations use Warm Standby or Pilot Light for non-critical systems and Multi-Site for mission-critical ones.


Previous: Lesson 8 — High Availability & Auto Scaling | Next: Lesson 10 — SAA-C03 Exam Guide


Part of the AWS Cloud Fundamentals course.

External references: