Phase Verbs / Action Phrases
| Phrase | Meaning |
|---|---|
| assign (an IP) | allocate an IP address to an instance |
| mask (a failure) | hide service disruption by quickly moving an Elastic IP to another instance |
| attach (Elastic IP to instance) | associate a fixed public IP with an EC2 instance |
| identify (a machine) | distinguish one host from another using an IP address |
| connect (via NAT) | allow a private IP to reach the internet through a proxy device |
Technical Vocabulary
| Term | Definition |
|---|---|
| IPv4 | IP address format: 4 numbers 0-255 separated by dots (e.g., 192.168.1.1); ~3.7 billion addresses |
| IPv6 | Longer address format (letters + numbers); used mainly in IoT; AWS supports it |
| Public IP | Globally unique IP; accessible from the internet |
| Private IP | Accessible only within a private network; two different networks can share the same private IP range |
| Elastic IP | A static public IPv4 you own and can reassign between instances |
| NAT (Network Address Translation) | Device that lets private IPs communicate with the internet |
| Internet Gateway | AWS network component that connects a VPC to the public internet |
| DNS | Domain Name System — maps human-readable names (like example.com) to IP addresses |
Key Concepts
IPv4 vs IPv6
- IPv4: Most common; 3.7 billion addresses (running low); used everywhere in this course
- IPv6: Much larger address space; mostly for IoT; AWS supports it but not the focus of SAA
Public IP
- Accessible from anywhere on the internet
- Must be globally unique — no two machines can share the same public IP
- AWS assigns one automatically when you launch an EC2 instance (unless disabled)
- Changes when you stop and restart the instance
Private IP
- Accessible only within the same private network (VPC)
- Not routable on the public internet
- Two separate networks can use the same private IP range (no conflict)
- Private IP is stable — does not change when an instance is stopped/started
Elastic IP
- A static public IPv4 address you own (doesn’t change when you stop/start instance)
- You pay for it even when not attached to a running instance
- Can be remapped instantly to another instance (useful for failover)
- Max 5 Elastic IPs per account by default (can request increase)
- Best practice: avoid Elastic IPs — they’re often a sign of poor architecture
Better Alternatives to Elastic IP
| Better Option | Why |
|---|---|
| Use a random public IP + DNS name (Route 53) | More scalable, easier to manage |
| Use a Load Balancer (no public IP on EC2 at all) | Best practice on AWS |
EC2 Default IP Behavior
| When | IP |
|---|---|
| Launch | Gets private IP (always) + public IP (if in public subnet) |
| SSH from outside | Must use public IP (or Elastic IP) — private IP only reachable via VPN |
| Stop → Start | Public IP changes; private IP stays the same |
Exam Tips
- EC2 gets a new public IP each time it stops and starts — use Elastic IP or DNS to avoid this
- Elastic IP = static public IP; max 5 per account; considered poor architecture in most cases
- Private IP = stable; doesn’t change; not internet-routable
- Prefer Load Balancer + DNS over Elastic IP for production workloads