S3 Overview

Amazon S3 is an object-based cloud storage utility to house an unlimited volume of files. Data is physically isolated inside Buckets assigned to specific AWS Regions. Instead of using traditional hierarchical file system, S3 stores data flatly as Objects mapped to unique string Keys. For big data processing, ML training pipelines, or standard web app backups, S3 is the foundational storage bed.

Key Takeaways​

🚨 The Big Namespace Upgrade (Global vs. Regional)​

For over a decade, every developer shared the exact same global headache: S3 bucket names had to be globally unique across the entire planet. If another AWS customer claimed my-logs, you were completely out of luck and had to append long, ugly strings.

AWS officially fixed this by introducing Account Regional Namespaces:

• The Old Global Way: Every bucket name exists in a single global internet pool. Once a name is taken anywhere in the world, it's locked forever.
• The Upgraded Way: You can now create buckets inside your own reserved account-regional namespace. This means you can use clean, simple names like logs, backups, or data in your production account, and use the exact same names in your staging or dev accounts without naming collisions!.
• How it Works: Behind the scenes, AWS seamlessly appends your unique account ID and region as a suffix (e.g., logs-123456789012-us-east-1-an), ensuring the underlying endpoint is perfectly unique without messing up your clean automation scripts.

The Object & Key Matrix (The Directory Illusion)​

S3 is not a file system like the hard drive on your laptop. It is a massive, flat key-value database.
When you see a path in the AWS console like s3://my-bucket/my-folder/another-folder/my-file.txt,

• The Illusion: S3 has zero concept of physical subdirectories or folders.
• The Reality: The entire path is one single flat text string called the Object Key.
• The Anatomy: S3 splits a key into a Prefix (the folder-like paths containing slashes) and the Object Name (the final file descriptor):

$$\text{Object Key} = \underbrace{\text{my\_folder/another\_folder/}}_{\text{Prefix}} + \underbrace{\text{my\_file.txt}}_{\text{Object Name}}$$

Object Constraints & Multi-Part Rule​

When pushing data payloads into your buckets, you must adhere to strict structural size limits:

• The 50 Terabyte Cap: The maximum size limit for any single object stored inside Amazon S3 is 50TB.
• The 5GB Multi-Part Rule: While 50TB is the ceiling, **you cannot upload a file larger than 5GB using a single, basic standard HTTP PUT request. Any file greater than 5GB must be uploaded using **Multi-Part Upload**. This splits the file into independent chunks, upload them in parallel to maximize network throughput, and stitches them back together automatically at the S3 edge.

Exam Tips​

The exam will tests your awareness of S3 object configurations and modern performance models:

The Metadata Lifecycle Shift: S3 allows you to append custom User Metadata (key-value text pairs describing the object) and Object Tags (up to 10 tags per file). The exam loves to test how these impact operations. Object Tags are distinct because they can be dynamically read by S3 Lifecycle Rules and IAM Policies. For example, you can write a rule stating: "If a file contains the tag Confidential=True, block all public access. If it contains Status=Deprecated, automatically move it toS3 Glacier storage after 30 days to save cash."