BitLink

BitLink is a production-grade URL shortening and analytics platform built with a system-design-first approach.

The system is designed to handle high-throughput redirect traffic while collecting rich analytics without impacting redirect latency. Critical workloads such as analytics processing are decoupled from the redirect path using an asynchronous queue architecture.

The goal of the project is to simulate the architecture of a real SaaS backend system, emphasizing scalability, failure isolation, and performance.

---

Core Features

Authentication & Identity

• Email/password authentication
• OAuth 2.0 login (Google, GitHub)
• Stateless JWT-based authentication
• Unified user identity across providers
• Guest link creation with automatic ownership migration after login

---

Link Management

• Short URL generation
• Custom aliases
• Collision detection
• QR code generation
• Link activation and expiration controls
• Ownership migration for guest-created links

---

Analytics System

BitLink records click events instead of simple counters, enabling richer traffic insights.

Tracked metadata includes:

• device type
• browser
• operating system
• geographic location
• referrer source
• timestamp

Analytics capabilities:

• per-link analytics
• account-wide analytics
• time-range filtering
• peak traffic analysis

Analytics processing is asynchronous to ensure redirects remain fast and reliable.

---

System Architecture

                Client
                   │
                   ▼
              CDN / Edge
                   │
                   ▼
                 Nginx
                   │
                   ▼
            Stateless API
                   │
        ┌──────────┴──────────┐
        ▼                     ▼
    Redis Cache           BullMQ Queue
        │                     │
        ▼                     ▼
     MongoDB            Worker Process
                            │
                            ▼
                         MongoDB

The architecture separates latency-critical redirect operations from analytics processing workloads.

---

Redirect Pipeline

User Request
      │
      ▼
CDN / Edge
      │
      ▼
Nginx Reverse Proxy
      │
      ▼
Backend API
      │
      ├── Redis Cache Lookup
      │
      ├── MongoDB Fallback (cache miss)
      │
      ├── Enqueue analytics event
      │
      ▼
Immediate Redirect Response

Key design principle:

Redirect responses never wait for analytics processing.

---

Asynchronous Analytics Processing

Analytics events are processed using BullMQ workers backed by Redis.

Workflow:

Redirect request
      │
      ▼
Analytics job added to queue
      │
      ▼
BullMQ Worker consumes job
      │
      ▼
Event processed and stored

Benefits:

• redirect latency remains minimal
• analytics processing scales independently
• worker failures do not affect redirect availability

---

Tech Stack

Frontend

• React
• Vite
• TailwindCSS

Responsibilities:

• link management UI
• analytics dashboards
• authentication flows

Deployment:

• Vercel (global CDN and automatic deployments)

---

Backend

• Node.js runtime
• Express framework
• Stateless API architecture
• Dockerized services

Responsibilities:

• redirect handling
• link management APIs
• authentication
• analytics event generation

Testing:

• Jest

Code quality:

• ESLint

---

Caching, Queues, and Rate Limiting

Redis

Used for:

• hot redirect caching
• API rate limiting
• queue backend

BullMQ

Used for:

• asynchronous analytics processing
• background job handling

---

Database

MongoDB Atlas

Responsibilities:

• link metadata storage
• analytics event storage
• aggregation queries for dashboards

Indexes are optimized for:

• short code lookups
• analytics queries

---

Reverse Proxy & Traffic Management

Nginx

Used as a reverse proxy to:

• route traffic to backend services
• buffer incoming requests
• optimize redirect handling

CDN functionality is provided via Vercel edge infrastructure.

---

Deployment Architecture

Component	Platform
Frontend	Vercel
Backend API	Render
Database	MongoDB Atlas
Cache	Redis
Queue Workers	BullMQ

The system previously ran on AWS infrastructure with Cloudflare CDN, but was migrated to Vercel and Render to optimize operational costs while maintaining global availability.

---

CI/CD Pipeline

Deployment automation is handled using:

• GitHub Actions
• Docker

Typical pipeline:

Code Push
   │
   ▼
Run Tests (Jest)
   │
   ▼
Build Docker Image
   │
   ▼
Deploy to Render / Vercel

---

Observability

Monitoring and visualization are implemented using:

• Prometheus
• Grafana

Metrics tracked include:

• redirect latency
• queue backlog
• cache hit rate
• API error rates

---

Infrastructure as Code

Infrastructure provisioning is managed using Terraform, allowing consistent environment configuration across:

• development
• staging
• production

This ensures reproducible deployments and simplified infrastructure management.

---

Design Tradeoffs

Several deliberate tradeoffs were made:

• Redirect availability is prioritized over analytics completeness
• Analytics are eventually consistent
• Asynchronous processing introduces slight analytics delays
• Managed platforms were chosen to reduce infrastructure overhead

These decisions prioritize reliability, scalability, and operational simplicity.

---

Future Improvements

Potential improvements include:

• Kafka-based event streaming for analytics
• ClickHouse or time-series database for analytics workloads
• Kubernetes-based orchestration
• Global edge redirect optimization
• Advanced abuse detection and bot filtering