Object Storage in Modern Systems

What Is Object Storage?

Object storage treats data as immutable blobs identified by a key (like a filename). Unlike block or file storage, there is no directory tree — just flat namespaces called buckets. Access is via HTTP REST (GET/PUT/DELETE).

Key Features

Feature	Description
Infinite scale	Petabytes stored without pre-provisioning capacity
Durability	11 nines (99.999999999%) — data replicated across ≥3 AZs
Eventual consistency	Strong read-after-write for new objects; eventually consistent for overwrite/delete
Versioning	Keep all versions; never accidentally delete production data
Lifecycle policies	Auto-transition to cheaper tiers (S3 → S3-IA → Glacier) after N days
Presigned URLs	Time-limited signed URLs for secure direct browser upload/download
Multipart upload	Upload large files in parallel chunks; resume on failure

import { S3Client, PutObjectCommand } from '@aws-sdk/client-s3';
import { getSignedUrl } from '@aws-sdk/s3-request-presigner';

const s3 = new S3Client({ region: 'us-east-1' });

// Generate a presigned URL valid for 5 minutes
// Client uploads directly to S3 — your server never touches the bytes
export async function getUploadUrl(key: string, contentType: string) {
  const command = new PutObjectCommand({
    Bucket: process.env.MEDIA_BUCKET!,
    Key: key,
    ContentType: contentType,
    // Enforce max file size via Content-Length condition
  });

  return getSignedUrl(s3, command, { expiresIn: 300 }); // 5 min
}

// After upload, client calls your API to confirm
// You then serve the object via CloudFront (CDN), not S3 directly

TIP: Never serve objects directly from S3. Put CloudFront (or any CDN) in front of S3. This gives you edge caching, HTTPS, custom domains, and shields S3 from direct traffic — cutting egress costs by 60–90%.