Building a Production Hotel Management System: HMS Anyaman
HMS Anyaman is a full-featured Hotel Management System I built and maintained over ~2 years. It covers the complete hotel operations lifecycle — from room booking and front desk management to housekeeping scheduling, dynamic rate optimization, event management, and financial reporting.
In this case study, I walk through the architecture, key technical decisions, the recent backend migration from Laravel to Node.js, and lessons learned from building a system with 186 database tables and 80+ API endpoints.
Tech Stack
Next.js React.js Laravel Node.js Express TypeScript MySQL PostgreSQL Prisma ORM PHP REST API Tailwind CSS GitLab CI/CD Nginx Ubuntu
Key Features
1. Room Booking & Front Desk
The core booking engine handles reservation creation, modification, cancellation, and check-in/check-out workflows. It supports drag-and-drop room assignment on a visual floor plan, making it intuitive for front desk staff to manage room availability in real-time.
2. Housekeeping Management
A dedicated housekeeping module tracks room status (clean, dirty, inspected, out-of-order), assigns cleaning tasks to staff, and logs room status history. The scheduler provides a timeline view for coordinating housekeeping shifts.
3. Dynamic Rate Management
The rate engine supports configurable pricing rules based on occupancy, seasonality, and market segments. I built a dynamic rate simulator that forecasts ADR (Average Daily Rate) and helps revenue managers optimize pricing strategies.
4. Role-Based Access Control (RBAC)
A granular permission system with menu-level CRUD flags and custom transaction actions. Each role (developer, administrator, receptionist, etc.) can be configured with specific view/add/edit/delete permissions per menu item. The permission tree is cached and resolved at login.
// Permission structure example
{
"permissions": [{
"key": ["Front", "Desk"],
"label": ["Front", "Desk"],
"isaccess": true,
"access": [{ "view": true, "add": true, "edit": true, "delete": false }]
}]
}
5. OTA Integration
Integrated with STAAH (an OTA channel manager) to synchronize room inventory and rates across multiple booking platforms. The integration ensures that when a booking is made on any connected OTA, the system updates availability in real-time to prevent overbooking.
6. Financial Reporting & Dashboards
Real-time dashboards display key metrics: revenue, occupancy rates, ADR, RevPAR, and booking trends. The reporting module covers profit & loss statements, trial balance reports, and market segment analysis.
Backend Migration: Laravel → Node.js
When I joined the project, the backend was built with Laravel (PHP) + MySQL. The frontend was Next.js. During my tenure, I planned and initiated a migration of the backend to Node.js (Express + TypeScript) + PostgreSQL, while maintaining 100% API compatibility so the frontend required zero changes.
Why Migrate?
- Unified language: Both frontend and backend in JavaScript/TypeScript reduces context switching
- Better ecosystem: Express + Prisma provides cleaner ORM and migration tooling
- Scalability: Node.js handles concurrent connections more efficiently for API workloads
Migration by the Numbers
- 186 tables migrated from MySQL to PostgreSQL
- 574MB of data migrated with 100% integrity verified
- 5-phase roadmap: Database → Core Architecture → Auth → API Endpoints → Testing
- 80+ API endpoints to replicate with identical response shapes
Key constraint: every API response must match the Laravel format exactly — including the { success, data, message, meta } envelope — so the Next.js frontend works without a single line change.
I built automation scripts for the migration: a connection tester, data migrator (MySQL → PostgreSQL), and data verifier with row-count validation. Phase 1 (Database Migration) was completed successfully.
Frontend Architecture
The frontend uses Next.js with a custom permission-aware component system. Every page and UI element conditionally renders based on the user's role and CRUD permissions. The app also has a Capacitor wrapper for mobile deployment via Android WebView.
State management uses Redux with persisted auth tokens. The sidebar and navigation are dynamically generated from the menu tree returned by the backend permission system.
Lessons Learned
- Permission systems are the hardest part — Getting menu-level CRUD right with polymorphic role assignment required careful database design and caching strategy.
- API contract first — Having a strict API response format made the migration feasible. Without it, swapping backend frameworks would have been much riskier.
- Automation saves weeks — Building the migration scripts (schema extract, data mover, verifier) turned a potentially weeks-long manual migration into a 30-minute automated process.
- OTA integration is complex — Synchronizing inventory across multiple channels in real-time requires careful handling of race conditions and partial failures.