Executive Summary
In the high-demand ride-hailing sector, system latency, data integrity, and real-time reliability dictate user trust and platform success. This case study highlights the engineering behind Anvo Autos—a resilient, high-concurrency backend engine built using Node.js, Fastify, PostgreSQL (PostGIS), Redis, and Socket.IO.
The platform guarantees 100% financial accuracy via double-entry ledger wallets, thwarts fare manipulation using automated PostGIS route deviation alerts, and secures passenger safety through zero-PII (Personally Identifiable Information) live trip tracking links.
Project Challenges
Challenge
1
High-Concurrency Driver Availability & Lifecycle: Managing thousands of drivers transitioning between Online, Offline, On Trip, and Paused states simultaneously creates extreme database lock contention. Stale availability states result in matching inefficiencies and drop-offs.
Challenge
2
Sub-Second GPS Ingestion & Network Jitter: Processing GPS coordinates emitted by driver devices every few seconds strains database write capacities, while packet loss causes lagging vehicle markers on client maps.
Challenge
3
Financial Race Conditions & Transaction Integrity: During peak surge hours, concurrent wallet debits, promo applications, and trip cancellations trigger race conditions, causing balance mismatches and double-spending.
Challenge
4
Fare Fraud & Route Deviation Detection: Detecting when a driver intentionally takes a longer, inefficient route to inflate travel fares requires live, continuous spatial tracking without overloading server memory.
Challenge
5
Data Privacy & Safe Trip Sharing: Riders need the ability to share live route tracking links with family. However, exposing internal database IDs or raw APIs risks leaking PII (Personally Identifiable Information).
Challenge
6
Fault-Tolerant Scheduled Rides: Delayed ride bookings must dispatch at the exact minute requested. System restarts or temporary Redis outages cannot be allowed to cause missed bookings.
Our Solutions
1. Robust Driver Lifecycle State Machine
- Enforced strict, backend-controlled transitions (OFFLINE ↔ AVAILABLE ↔ ON_TRIP).
- A background cron job (DriverHealthJob) offlines inactive drivers after 30 seconds of heartbeat loss.
2. Redis GEO & GPS Smoothing
- Cached driver locations in memory using Redis GEO for sub-second updates.
- Implemented a GPS filter that ignores insignificant movement (<10m) to minimize database write-overhead.
3. Immutable Double-Entry Ledger Wallet
- Configured a transaction ledger (wallet_ledger) utilizing Row-Level Locking (SELECT FOR UPDATE) within atomic database transactions.
- Guarantees negative balance protection and prevents race conditions.
4. PostGIS LineString Deviation Tracking
- Decoded planned Google Routes polylines into PostGIS LineString geometry in PostgreSQL.
- Used ST_Distance to compare live driver GPS points against the route, triggering Socket.IO deviation warnings if they stray >500m.
5. Argon2 OTP-Secured Trip Validation
- Secured the trip pickup phase by requiring drivers to input a 4-digit OTP.
- Verified against an Argon2 hash in the database, locking verification after 3 failed attempts to block brute-force attacks.
6. High-Entropy, Zero-PII Trip Sharing
- Developed a tokenized sharing link (64-character hex token).
- The public endpoint filters out sensitive details, streaming progress to read-only Socket.IO rooms.
7. Resilient Scheduled Ride Queuing
- Automated booking dispatches with BullMQ workers on Redis.
- Built a secondary local fallback database scheduler to guarantee execution during temporary Redis outages.
Technologies Used
For a highly concurrent, real-time platform like Anvo Autos, choosing the right technologies was critical to ensure low latency and high availability. We built a robust backend powered by Node.js and Fastify, using PostgreSQL with PostGIS extension for accurate spatial tracking and Redis for sub-second location caching and job queues. This combination enables the system to comfortably handle extreme traffic and high-frequency writes.
Runtime & Framework
01Node.js Fastify TypeScript
Geospatial & ORM
02PostgreSQL + PostGIS Prisma ORM
Caching & Real-Time
03Redis GEO Socket.IO BullMQ
Security & Observability
04Argon2 MSG91 Prometheus
Impact On Business
0% Ledger Discrepancies: The double-entry ledger architecture completely eliminated race conditions, duplicate promo redemptions, and wallet balance errors.
45% Map API Cost Reduction: Caching route geometry in Redis and applying throttling logic successfully decreased recurring Google Maps API costs.
99.9% Booking Dispatch Rate: The combination of BullMQ workers and local database fallbacks ensured scheduled trips were processed without failure.
Zero Passenger Data Leaks: Secure, high-entropy tokenized links enabled anonymous web tracking without exposing internal IDs, usernames, or phone numbers.
Proactive Fraud Mitigation: Real-time route deviation alerts successfully flagged unauthorized route expansions, reducing fare disputes by 30%.
Conclusion
The architecture of Anvo Autos demonstrates that high-concurrency web applications do not need to sacrifice financial accuracy or user privacy for real-time responsiveness. By offloading hot-path GPS data to Redis GEO, utilizing PostGIS for localized spatial logic, and securing transactions with immutable ledgering and row-level locking, the platform stands as a benchmark for modern, scalable, and secure ride-hailing engines.
Case Study
SIP Calculator & Wealth Planner
Advanced Interactive Financial Simulation & Wealth Management Ecosystem
Read More