AWS EC2

Amazon Elastic Compute Cloud (EC2) provides scalable computing capacity in the AWS cloud.

Instance Selection

Choose the right instance type for your workload. You can change instance types later, but some changes require stopping the instance.

Key Concepts

Instance Types

EC2 offers various instance types optimized for different use cases:

Family	Use Case	Examples
t	Burstable workloads	t3.micro, t3.medium
m	Balanced compute/memory	m5.large, m6i.xlarge

Best for: Web servers, development environments, small databases

Family	Use Case	Examples
c	Compute optimized	c5.large, c6i.xlarge
r	Memory optimized	r5.large, r6i.xlarge

Best for: Batch processing, gaming servers, in-memory caching

Family	Use Case	Examples
p/g	GPU instances	p3.2xlarge, g4dn.xlarge
i/d	Storage optimized	i3.large, d2.xlarge

Best for: Machine learning, video encoding, data warehousing

AMIs (Amazon Machine Images)

Pre-configured templates containing the OS and applications:

Amazon Linux 2023: AWS-optimized Linux distribution
Ubuntu: Popular open-source Linux
Windows Server: Microsoft Windows OS
Custom AMIs: Your own preconfigured images

Amazon Linux 2023 is recommended for most workloads as it's optimized for AWS and includes long-term support.

Security Groups

Virtual firewalls controlling inbound/outbound traffic:

security-group-rules.json

{
  "IpPermissions": [
    {
      "IpProtocol": "tcp",
      "FromPort": 22,
      "ToPort": 22,
      "IpRanges": [{"CidrIp": "10.0.0.0/8"}]
    },
    {
      "IpProtocol": "tcp",
      "FromPort": 443,
      "ToPort": 443,
      "IpRanges": [{"CidrIp": "0.0.0.0/0"}]
    }
  ]
}

Security Best Practice

Never allow SSH (port 22) from 0.0.0.0/0. Use Systems Manager Session Manager or restrict to specific IPs.

Key Pairs

SSH key pairs for secure access:

Create Key Pair

aws ec2 create-key-pair --key-name my-key --query 'KeyMaterial' --output text > my-key.pem
chmod 400 my-key.pem

Instance Lifecycle

State	Description
pending	Instance is launching
running	Instance is running and ready
stopping	Instance is stopping
stopped	Instance is stopped (no compute charges)
terminated	Instance is permanently deleted

Stopped instances don't incur compute charges but you still pay for attached EBS volumes.

Pricing Models

Pay by the second with no commitments.

Best for:

Unpredictable workloads
Short-term projects
Development and testing

Up to 72% savings with 1 or 3-year commitment.

Options:

Standard Reserved: Fixed instance type
Convertible Reserved: Can change instance types
Scheduled Reserved: Specific time windows

Up to 90% savings for fault-tolerant workloads.

Spot instances can be interrupted with 2 minutes notice. Use for fault-tolerant, stateless workloads.

Flexible pricing with commitment to usage ($/hour).

Types:

Compute Savings Plans: Any region, family, OS
EC2 Instance Savings Plans: Specific region/family

Storage Options

EBS (Elastic Block Store)

Persistent block storage volumes:

Type	Use Case	IOPS	Throughput
gp3	General purpose SSD	3,000-16,000	125-1,000 MB/s
io2	High performance	Up to 64,000	1,000 MB/s
st1	Throughput HDD	500	500 MB/s
sc1	Cold HDD	250	250 MB/s

gp3 is recommended for most workloads. You can provision IOPS and throughput independently.

Instance Store

Temporary storage attached to the host:

High I/O performance (NVMe SSD)
Data lost when instance stops
Good for caches, buffers, and temporary data

Best Practices

Security

Use security groups as the first line of defense
Keep instances in private subnets when possible
Use Systems Manager Session Manager instead of SSH
Enable detailed monitoring for production workloads
Always use IAM roles instead of storing credentials

Cost Optimization

Right-size instances based on actual usage
Use Reserved Instances for steady-state workloads
Use Spot Instances for fault-tolerant workloads
Stop instances when not in use
Use AWS Compute Optimizer for recommendations

High Availability

Distribute instances across multiple AZs
Use Auto Scaling for automatic scaling
Use Elastic Load Balancing for traffic distribution
Implement health checks and auto-recovery

User Data

Run scripts on instance launch:

user-data.sh

#!/bin/bash
yum update -y
yum install -y httpd
systemctl start httpd
systemctl enable httpd
echo "Hello World" > /var/www/html/index.html

User data scripts run as root. Logs are available at /var/log/cloud-init-output.log.