Task Statement 2.2: Design highly available and/or fault-tolerant architectures.

📘AWS Certified Solutions Architect – (SAA-C03)

1. First, Understand the Goal

✅ High Availability (HA)

System stays up and running with minimal downtime.

✅ Fault Tolerance (FT)

System continues working even if a component fails completely.

2. Core Building Blocks in AWS

Before choosing services, understand the infrastructure levels:

1. Region

A geographic area (e.g., us-east-1)
Used for disaster recovery

2. Availability Zone (AZ)

Multiple isolated data centers inside a Region
Used for high availability

👉 Exam Tip:

Multi-AZ = High Availability
Multi-Region = Disaster Recovery + Fault Tolerance

3. Compute Layer (Application Servers)

Services to Use:

🔹 Amazon EC2

Run application servers
Deploy across multiple AZs

🔹 AWS Auto Scaling

Automatically adds/removes EC2 instances
Ensures system stays available during:
- Failures
- Traffic spikes

🔹 AWS Elastic Beanstalk

Automatically deploys apps across AZs
Handles scaling + health checks

Key Design Rule:

Always run at least 2 EC2 instances in different AZs

4. Load Balancing Layer

🔹 Elastic Load Balancing

Types:

Application Load Balancer (ALB)
Network Load Balancer (NLB)

What it does:

Distributes traffic across multiple servers
Detects unhealthy instances
Sends traffic only to healthy ones

Key Design Rule:

Always place Load Balancer across multiple AZs

5. Database Layer

🔹 Amazon RDS

Features:

Multi-AZ Deployment
- Primary DB + standby DB in another AZ
- Automatic failover

🔹 Amazon Aurora

More advanced than RDS
Automatically replicates across multiple AZs
Faster failover

🔹 Amazon DynamoDB

Fully managed NoSQL DB
Built-in multi-AZ replication
Option for Global Tables (multi-region)

Key Design Rule:

Use Multi-AZ databases for HA
Use cross-region replication for DR

6. Storage Layer

🔹 Amazon S3

Automatically stores data across multiple AZs
Very high durability (11 9’s)

🔹 Amazon EBS

Attached to EC2
Stored in one AZ only
Use snapshots for backup

🔹 Amazon EFS

Shared file system
Works across multiple AZs

Key Design Rule:

Use S3 or EFS for multi-AZ resilience
Avoid single-AZ storage for critical systems

7. Networking & Traffic Routing

🔹 Amazon Route 53

Features:

Health checks
DNS-based routing

Routing Types:

Failover routing
Latency-based routing
Weighted routing
Geolocation routing

🔹 Amazon CloudFront

Global content delivery
Improves availability and performance

Key Design Rule:

Use Route 53 for multi-region failover

8. Multi-Region Architecture Services

To design across Regions:

🔹 Amazon S3 Cross-Region Replication

Automatically copies data to another Region

🔹 Amazon RDS Read Replicas

Replicate database to another Region
Can promote to primary during failure

🔹 AWS Backup

Centralized backup across Regions

🔹 AWS Elastic Disaster Recovery

Full disaster recovery solution

9. Common Architectures You Must Know for Exam

1. Multi-AZ Architecture (Most Common)

Components:

EC2 in multiple AZs
Auto Scaling
Load Balancer
Multi-AZ RDS

✅ Provides:

High Availability
Automatic failover

2. Active-Passive (Multi-Region)

One Region = active
Another Region = standby

Uses:

Route 53 failover
Data replication

3. Active-Active (Multi-Region)

Both Regions serve traffic
Load distributed globally

Uses:

Route 53 latency routing
DynamoDB Global Tables

10. Key Exam Decision Patterns

Scenario 1:

“Application must survive AZ failure”

👉 Use:

Multi-AZ EC2 + ALB
Multi-AZ RDS

Scenario 2:

“Application must survive Region failure”

👉 Use:

Route 53 failover
Cross-region replication

Scenario 3:

“Need automatic scaling + availability”

👉 Use:

Auto Scaling + ELB

Scenario 4:

“Database must not lose data”

👉 Use:

RDS Multi-AZ
Aurora

Scenario 5:

“Global users need low latency”

👉 Use:

CloudFront
Route 53 latency routing

11. Important Comparison Table

Layer	Multi-AZ Solution	Multi-Region Solution
Compute	EC2 + Auto Scaling	Deploy in multiple Regions
Database	RDS Multi-AZ	Read Replicas / Global DB
Storage	S3	Cross-Region Replication
Traffic	ALB	Route 53
DR	Backup	Multi-Region failover

12. Common Exam Traps ⚠️

❌ Single EC2 instance → NOT highly available
❌ EBS only → Single AZ risk
❌ RDS without Multi-AZ → No failover
❌ No health checks → No automatic recovery

13. Final Summary (Must Remember)

To design HA & Fault-Tolerant architecture:

✔ Always:

Use multiple AZs
Add load balancing
Enable auto scaling
Use managed services

✔ For critical systems:

Add multi-region setup
Use Route 53 failover
Replicate data across Regions

14. Quick Exam Memory Trick

👉 “Compute + Load Balancer + Database + Replication + DNS”

If all 5 are covered → system is highly available and fault tolerant.