HoyMismoGPS V2: Enterprise Fleet Management
A complete architectural overhaul — from a Python Asyncio prototype to a full Google Cloud Platform enterprise system capable of monitoring 500+ vehicles with sub-second GPS updates, executive-grade analytics, and 99.9% uptime SLA.
Live demo — real-time tracking, route replay, and executive dashboard
Platform Interface
Executive Dashboard
KPIs, alerts, fleet overview
Real-Time Map
Animated vehicle markers
Fleet Management
Status, driver scores, alerts
The Architectural Transformation: V1 → V2
V1 was built as a proof-of-concept: a Python Asyncio TCP server that decoded Teltonika binary packets, pushed coordinates to Redis, and rendered them on a Mapbox GL map. It worked — but it couldn't scale, had no analytics, no alerting pipeline, and required manual server management. V2 is a complete reimagination on Google Cloud Platform.
V1 — Prototype Limitations
- ✕ Python Asyncio + single VPS (no auto-scaling)
- ✕ Redis for pub/sub (self-managed, no SLA)
- ✕ PostgreSQL for history (no BigQuery analytics)
- ✕ Mapbox GL only (no traffic, no Street View)
- ✕ No driver scoring, no automated reports
- ✕ Single-tenant, no white-label support
V2 — Enterprise GCP Architecture
- ✓ Cloud Run (auto-scaling, zero-downtime deploys)
- ✓ Cloud Memorystore (managed Redis with 99.9% SLA)
- ✓ Firestore (real-time) + BigQuery (analytics/history)
- ✓ Google Maps Platform (live traffic, Street View)
- ✓ Driver scoring, smart alerts, automated PDF reports
- ✓ Multi-tenant architecture with role-based access
Data Pipeline: Device to Dashboard
Every GPS ping from a Teltonika FMB920 or FMB140 device travels through a cloud-native event pipeline before appearing on the map in under one second. The architecture uses Pub/Sub as the backbone so that multiple consumers (real-time map, alert engine, analytics writer) can process the same event independently and at scale.
Full Data Flow
Teltonika FMB920/FMB140
TCP Port 5000 (Codec 8 / Codec 8 Extended)
|
v
GPS Server — Cloud Run (auto-scaling)
Decodes binary packet → validates IMEI → extracts:
{ lat, lng, speed, ignition, odometer, fuel, ... }
|
v
Cloud Pub/Sub — Topic: "gps-events"
|
+──────────────────────────────────────────+
| | |
v v v
Cloud Function: Cloud Function: Cloud Function:
save-location check-alerts update-stats
| | |
+──> Firestore +──> FCM Push +──> BigQuery
(real-time pos) (speeding alert) (daily analytics)
+──> Memorystore +──> Firestore +──> Firestore
(cache last pos) (alert record) (KPI counters)Enterprise Platform Features
Full Technology Stack
Enterprise-Grade at SMB Pricing
One of the key design goals for V2 was cost efficiency without sacrificing reliability. By leveraging serverless Cloud Run (pay-per-request), managed Firestore (pay-per-read/write), and BigQuery's free tier for analytics queries under 1TB/month, the platform achieves enterprise-grade SLA at a fraction of traditional fleet management software costs.
Comparison: Traditional fleet management platforms (Samsara, Verizon Connect) charge $25–50/vehicle/month. The V2 GCP architecture delivers comparable features at roughly $3.50–5/vehicle/month for a 20-vehicle fleet, with the cost-per-vehicle dropping significantly as the fleet grows.
Need Enterprise GPS Fleet Management?
I can build a similar system for your fleet — from hardware integration (Teltonika, Concox, Queclink) to a fully branded dashboard with your company's branding and custom alert rules.