Self-managed vs Fully Managed

Self-managed (MySQL on EC2)

You handle everything yourself:

You are responsible for:
├── Launching EC2 and installing MySQL
├── Configuring CPU / Memory
├── Expanding disk when nearly full
├── Regular backups
├── Version upgrades / security patches
├── Primary-replica replication (high availability)
├── Monitoring and alerting
└── Fixing issues when things break

Python analogy: Like manually pip install-ing every package, managing virtualenvs yourself, writing your own backup scripts — all infrastructure is your responsibility.

import subprocess

subprocess.run(["apt", "install", "mysql-server"])   # install
subprocess.run(["mysqldump", "-u", "root", "mydb"])  # backup (you write this)
subprocess.run(["systemctl", "restart", "mysql"])    # crashed? restart yourself

---

Fully Managed (DynamoDB)

AWS handles it — you just use it:

AWS is responsible for:
├── Hardware, OS, database engine
├── Auto scaling
├── Automatic backups
├── Automatic cross-AZ replication (high availability)
├── Security patches
└── Monitoring

You only need to:
├── Create tables
├── Read and write data
└── Pay the bill (pay-per-use)

Python analogy: Like using boto3 directly — you only write business logic, infrastructure is completely AWS's concern.

import boto3

# Fully managed: you only write this, everything underneath is AWS's problem
dynamodb = boto3.resource("dynamodb", region_name="us-east-1")
table = dynamodb.Table("my-table")

table.put_item(Item={"id": "123", "value": "hello"})
response = table.get_item(Key={"id": "123"})
# Backups? Scaling? HA? → AWS handles it automatically

---

One-Line Summary

	Self-managed (EC2 + MySQL)	Fully Managed (DynamoDB)
Analogy	Buy ingredients, cook yourself, wash the dishes	Order at a restaurant, kitchen is not your concern
Python comparison	Build server + write all tooling scripts	Just import boto3 and use it
Flexibility	High (full control)	Low (AWS owns the internals)
Ops burden	High	Near zero