These are my personal notes to better understand AWS compute services and be able to answer related questions in the Cloud Practitioner certification. I’m keeping it simple and clear — not trying to be exhaustive, just making sure I get it.
🖥️ Amazon EC2 (Elastic Compute Cloud)
EC2 lets me launch virtual servers in the cloud. I can choose how much memory, CPU, storage, and which OS I want — and AWS handles the physical hardware. It’s like renting a server, but only when I need it.
⚙️ EC2 Instance Types
There are different types of EC2 instances depending on what the app needs:
- General Purpose → balanced: good default choice
- Compute Optimized → more CPU power (e.g., data processing)
- Memory Optimized → more RAM (e.g., databases, analytics)
- Storage Optimized → fast access to local disk (e.g., big data)
- Accelerated Computing → GPU or FPGA for ML or graphics
💸 EC2 Pricing Options
☁️ On-Demand
I pay by the second or hour. No commitment. Great for testing or short-lived stuff.
📆 Reserved Instances (RI)
I commit to using a specific instance type in a region for 1 or 3 years. In return, I get up to 72% discount. Good if I know I’ll use it consistently.
💵 Savings Plans
I commit to spending a fixed amount per hour (e.g., $5/hour), and AWS gives me discounts on EC2, Lambda, and Fargate — up to that amount. It’s more flexible than Reserved Instances.
🔁 Spot Instances
These use spare capacity, and I can save up to 90%. But AWS can stop the instance at any time (with a 2-minute warning). Great for background jobs or anything I can restart easily.
🧍 Dedicated Hosts
This is for when I need to run my stuff on a dedicated physical server — maybe for licensing or compliance reasons.
📈 Auto Scaling & Load Balancing
🔄 EC2 Auto Scaling
This automatically adds or removes EC2 instances depending on usage (CPU, memory, etc.). It’s useful to handle peaks and save money when there’s less traffic.
⚖️ Elastic Load Balancer (ELB)
This spreads incoming traffic across multiple instances. If one fails, traffic goes to the others. Great for high availability.
📣 Amazon SNS (Simple Notification Service)
SNS lets me send notifications or messages to:
- Email addresses
- SMS numbers
- SQS queues
- Lambda functions
- HTTP endpoints (webhooks)
It’s useful to connect AWS services together or notify someone when something happens.
📬 Amazon SQS (Simple Queue Service)
SQS is a message queue. One service puts a message in the queue, another service reads it later. It’s a way to decouple services so they don’t depend on each other directly.
SNS + SQS = classic combo. 💪
🧠 AWS Lambda
Lambda runs small bits of code without servers. I just upload the function, set a trigger, and AWS runs it.
✅ Pros:
- No server to manage
- Auto scaling
- Pay only for what I use
⚠️ Limitations:
- Max 15 minutes per execution
- Stateless (no memory between runs)
🐳 ECS & ☸️ EKS (Container Orchestration)
🔵 ECS (Elastic Container Service)
This is AWS’s own container service. I define what Docker container to run, and ECS runs it on EC2 or Fargate. Simple, AWS-native.
⚙️ EKS (Elastic Kubernetes Service)
This is managed Kubernetes on AWS. More flexible and powerful, but more complex too. Ideal if I (or my team) already know Kubernetes.
🚀 AWS Fargate
Fargate is awesome: I just define how much memory and CPU I need for my container — and AWS runs it for me.
No EC2 instances, no infrastructure, just my app.
👍 Super simple
💸 More expensive if I run something 24/7
Works with both ECS and EKS.