Requirements

Functional Requirements:

• Create a short URL for a given long URL.
• Return the long URL associated with a given short URL.

Non-Functional Requirements:

Availability

• Should have 99.99% uptime

Reliability

• No data loss

Performance

• Should be under 100 ms to be redirected from short URL to original URL
• Should take under 200 ms to create a short url from an original URL

Throughput

• Read requests - 100,000 to 1,000,000 reads/second
• Write requests - 10,000 to 100,000 writes/second
• Should implement caching underneath the hood

Scalability

• Horizontal scaling
• User Base:
  • 100 million daily active users
• Storage
  • 100 million URLs
  • If each mapping takes 100 bytes, need 10 GB for 100 million mappings
  • DB: Dynamodb/Redis/S3
• Load Calculation
  • 100 million users make 10 requests/day -> 1 billion requests

Backup/Recovery

• Regularly backup data to prevent data loss

Security

• Implement https to prevent misuse + secure transmission

API Design

API 1 -> Create short URL:

Action: Users provides a short URL from a long URL

• POST /shorten
• Content-Type: Application/json

{

long-url: string

}

Response: HttpCreated 201

Content-Type: application/json

{

"short_url": string,

"expires_at": datetime

}

API 2 -> Retrieve/Redirect to local URL:

Action:

1. Lookup short_code in DynamoDB
2. If found, return HTTP 302 redirect to long_url
3. Else return 404 Not Found

Request:

• GET /{short_code}
• Content-Type: Application-json

High-Level Design

• CloudFront caches popular URL mappings globally,which is the edge server. It will do rate-limiting & security checks.
• We will then send it to API gateway to handle incoming requests
• We will then have a load balancer (Lambda@Edge), a serverless compute service that is beneficial for caching responses & performing lightweight processes such as URL redirection without hitting your backend service for every request. This reduces latency + improve performance.
• URL Shortening service will create the url, put entry into dynamodb and store off in the cache
• Redirect service reads from cache. On a cache hit, it retrieves the original URL. On a cache miss, it retrieves data from dynamodb and then updates the cache.
• We send the data over to

Detailed Component Design

URL Shortening Service

• Functionality
  • Accepts long URL
  • Generate unique alias for the given URL
  • Stores that mapping to short URLs in dynamo db
• Tech stack
  • Node.js
  • AWS Elasticache, Dynamodb

Redirecting Service

• Functionality
  • Generate unique alias for the given URL
  • Stores that mapping to short URLs in dynamo db
• Tech stack
  • Node.js
  • AWS Elasticache, Dynamodb

Analytics Service

• Functionality
  • Tracks logs/usage data from each service
• Tech stack
  • Nodejs
  • Cloudwatch