DEV Community

Cover image for AWS S3 for Ecommerce Image Storage
AYON
AYON

Posted on

AWS S3 for Ecommerce Image Storage

#architecture #aws #backend #systemdesign

Why AWS S3

We are selecting AWS S3 because it provides scalable, secure, and highly durable object storage, while aligning with our AWS ecosystem strategy and long-term scalability goals.

Key Reasons

  • Reduces backend and database load
  • Lightweight REST API payloads
  • Direct client upload using Pre-Signed URLs
  • Secure authentication and authorization flow
  • CDN support through CloudFront
  • Future image optimization support (thumbnail / medium / large variants)
  • Easier long-term maintenance by staying within the AWS ecosystem

Why Not Store Images as Blob in Database

❌ Traditional approach:

Database (Blob)
      │
Image Binary
      ▼
Spring Boot API
      │
Heavy REST Payload
      ▼
Browser / Mobile
Enter fullscreen mode Exit fullscreen mode

Problems

  • Larger database size
  • Heavy REST API payloads
  • Backend transfers image bytes on every request
  • Higher memory and bandwidth usage
  • Database backup/restore becomes expensive
  • Scaling becomes difficult

Recommended Architecture

Keep binary files in S3, metadata in database.

product_images
--------------
id
product_id
image_url
s3_key
filename
filesize
filetype
created_at
Enter fullscreen mode Exit fullscreen mode

Image bytes never enter PostgreSQL.

Secure Upload Flow 

                 IMAGE UPLOAD FLOW

┌─────────────────┐
│ Client           │
│ Web / Mobile App │
└────────┬────────┘
         │
         │ 1️⃣ Upload Request
         │ JWT / Session Token
         ▼

┌────────────────────────────────┐
│ Spring Boot + Spring Security  │
│                                │
│ ✔ Authentication               │
│ ✔ Authorization                │
│ ✔ Validate Content Type        │
│ ✔ Generate Pre-Signed URL      │
└────────────┬───────────────────┘
             │
             │ 2️⃣ Temporary Upload URL
             ▼

       ┌──────────────┐
       │   Amazon S3  │
       └──────▲───────┘
              │
              │ 3️⃣ Direct Upload
              │ Secure temporary access
              │ Image never touches backend
              │
┌─────────────┴─────────────┐
│ Client                    │
└─────────────┬─────────────┘
              │
              │ 4️⃣ Save Metadata
              │ image_url
              │ s3_key
              │ filename
              ▼

      ┌────────────────┐
      │ PostgreSQL      │
      └────────────────┘
Enter fullscreen mode Exit fullscreen mode

Security Benefits

Backend responsibilities:

  • Authentication
  • Authorization
  • File validation
  • Generate temporary secure upload URL
  • Store metadata only

S3 responsibilities:

  • Secure object storage
  • Temporary Pre-Signed URL access
  • IAM permissions
  • Encryption
  • Private object support

Backend never exposes AWS credentials.

Image Fetch Flow (Optimized) 

Client (Web / Mobile)
        │
        │ GET /products
        ▼

Spring Boot API
        │
        │ Product Data + Image URL only
        │ Lightweight REST payload
        ▼

PostgreSQL

Response:

{
  "name": "Nike Shoes",
  "price": 2999,
  "imageUrl":
  "https://cdn.app.com/products/123.jpg"
}

        │
        │ Browser loads image URL
        ▼

Nearest CloudFront CDN Edge
(Cache Hit → Instant response)

        │ Cache Miss
        ▼

Amazon S3
Enter fullscreen mode Exit fullscreen mode

Performance Benefits

REST API transfers:

✅ Product metadata only

{
  "name": "Nike Shoes",
  "price": 2999,
  "imageUrl": "https://cdn.app.com/products/123/main.jpg"
}
Enter fullscreen mode Exit fullscreen mode

NOT:

❌ Image binary data

Benefits

  • Smaller REST payloads
  • Lower backend CPU usage
  • Lower bandwidth consumption
  • Faster API responses
  • Better scalability during traffic spikes

CDN Optimization (Future Scale)

Future architecture:

Client
   │
   ▼

Nearest CloudFront Edge Location
(Cache Frequently Requested Images)

   │ Cache Miss
   ▼

Amazon S3
Enter fullscreen mode Exit fullscreen mode

Benefits

  • Users receive images from nearest CDN edge location
  • Frequently accessed resources remain cached
  • Lower latency globally
  • Faster image loading
  • Reduced origin load
  • Better UX
  • Better SEO
  • Lower infrastructure cost

Future Image Optimization

Current:

product.jpg (1200px)
Enter fullscreen mode Exit fullscreen mode

Future pipeline:

Upload Original
       │
       ▼

Amazon S3

       │
Lambda / Image Processor
       ▼

thumbnail.jpg  (300px)
medium.jpg     (600px)
large.jpg      (1200px)
Enter fullscreen mode Exit fullscreen mode

Examples:

Product Listing:

cdn.app.com/products/123?size=thumbnail
Enter fullscreen mode Exit fullscreen mode

Product Details:

cdn.app.com/products/123?size=large
Enter fullscreen mode Exit fullscreen mode

In practice, CloudFront can forward these query parameters, and a Lambda image service can return the correct variant while keeping the same clean URL structure.

Why

  • Product list page does not need high-resolution images
  • Lower bandwidth usage
  • Faster mobile experience
  • Better rendering performance

Why AWS S3 over Other Storage Solutions

Criteria AWS S3 EdgeStore Google Cloud Storage
Highly scalable object storage ✅ Depends provider ✅
Secure upload flow ✅ Native Depends implementation ✅
CDN ecosystem CloudFront External setup Cloud CDN
AWS ecosystem integration Excellent Neutral Limited
Future scale readiness Excellent Moderate Excellent
IAM & security controls Strong Depends provider Strong
Durability 99.999999999% Depends provider High

Final Decision

AWS S3 is selected because it provides scalable, secure, and highly durable object storage while reducing REST payload size, eliminating backend image transfer bottlenecks, enabling secure direct uploads, supporting global CDN delivery, and providing a future-proof architecture for ecommerce scale.

Top comments (0)