Alegor Tech
TELEMATICS & REAL-TIME PLATFORMS

450,000 Devices. One System. Real-Time. From physical tracker and cellular networks to real-time processing, time-series data, APIs, and operational services.

Gelios is a telematics platform that, as of August 2026, continuously processes data from approximately 450,000 tracking devices in active production. The platform maintains real-time position and status, reconstructs historical trajectories, and delivers live data to services in both commercial and public sectors.

Project Gelios
Domain Telematics / IoT
Type Real-time Platform
Status Active Production
Scale (Aug 2026) ~450 000 enheter
1. Theory versus Reality

Scalability is not about how many devices a system could theoretically handle. It is about how many it actually handles when everything must work reliably every single day.

Many systems are labeled with generic industry terms like "scalable", "high performance", "enterprise ready", or "mission critical". But real-world software engineering is not about marketing labels or synthetic benchmarks in sterile test environments.

Gelios does not need abstract claims. In August 2026, the platform continuously ingests, parses, validates, and stores telemetry from approximately 450,000 tracking devices in active production. This is not a benchmark, not theoretical capacity, and not a future roadmap. It is live production – and that is the definitive proof.

Live Production
~450 000
Active tracking devices
Verified Timestamp
August 2026
Verified production status
Continuity
24/7/365
Continuous telemetry ingestion
Deployment
Commercial & Public
Mission-critical services
2. Understanding the Scale

What does 450,000 active devices really mean?

A tracker is not a passive row in a database that updates once a month. It communicates constantly. Units move at highway speeds, jump between cellular towers, cross borders, lose signal in tunnels, buffer data packets locally, and reconnect in rapid burst transmissions.

Therefore, the platform cannot operate on the traditional web application lifecycle:

User opens webpage → Server computes → Response returned → Work completed.

Instead, the infrastructure operates in a continuous, high-concurrency ingestion loop:

  • 450,000 physical devices emit continuous telemetry
  • Real-time ingestion, protocol parsing, and payload validation
  • Parallel updates: Current State + Time-Series History + Events
  • Instant delivery to external APIs, live map views, and enterprise services
Continuous Telemetry Convergence
~450k DEVICES
450 000 DEVICES ACTIVE TRACKERS
Position, speed, ignition, battery, sensors, CAN bus
REALTIME INGESTION & PROCESSING Go / MQTT / EMQX
Ingestion → Device Auth → Parsing → Validation → Async Dispatch
CURRENT STATE
Instant location, online status & memory cache (Redis)
TIME-SERIES
Historical routes, telemetry archives & analytics (TimescaleDB)
APIs & OPERATIVA TJÄNSTER Live maps, fleet management, external enterprise ERPs & public agencies
3. The Technical Chain

A single dot on a map requires an entire ecosystem behind it.

What the end user sees in the interface may be a simple icon moving along a road. But before that dot is rendered, the data must traverse, validate, and compute across multiple specialized technological layers.

01
SATELLITE / VEHICLE

GPS & Vehicle Telemetry

Satellite fix (GPS/A-GPS), CAN bus metrics, ignition state, and inertial sensor telemetry.

02
TRACKING DEVICE & NETWORK

Hardware & Cellular Net

Hardware firmware serializes telemetry packets and transmits over LTE-M/GSM via compact protocols.

03
INGESTION SERVICE

Ingestion & Parsing

High-concurrency Go services and MQTT brokers ingest hundred-thousands of connections and parse payloads.

04
REALTIME PROCESSING

Real-time Processing

Jitter filtering, map matching, continuous mileage computation, and geofence boundary checks.

05
CURRENT STATE

Instant State (Redis)

Last known coordinates, online status, velocity, and sensor state saved in sub-millisecond memory cache.

06
TIME-SERIES STORAGE

TimescaleDB Archives

Billions of telemetry points committed asynchronously to TimescaleDB for historical trajectory querying.

07
LARAVEL BUSINESS API

Business & REST API

Tenant authorization, fleet scoping, data normalization, and delivery via REST & webhook endpoints.

08
MAP / EXTERNAL SERVICES

Live Map & Operations

The icon updates on screen – validated, geo-indexed, and integrated into customer workflows.

When this multi-tier pipeline executes continuously for ~450,000 active devices around the clock, architectural discipline, fault isolation, and efficiency become essential.

4. Core Architectural Principle

Real-time state and historical time-series are two distinct engineering challenges

A high-scale telematics engine must swiftly answer two fundamentally different queries:

Query 1: Real-time
”Var befinner sig enheten just nu?”
Query 2: History
”Hur har den rört sig under en viss tidsperiod?”

Attempting to render live fleet views by querying historical telemetry tables would overwhelm any database architecture under 450,000 active device streams.

Therefore, instantaneous device state (Current State) is strictly decoupled from long-term time-series archives. This decoupling is a cornerstone that makes massive scale stable.

INCOMING CONTINUOUS TELEMETRY
CURRENT STATE Redis / Fast RAM
  • • Senaste kända GPS-koordinat
  • • Online / Offline heartbeat
  • • Aktuell hastighet & riktning
  • • Senaste sensortillstånd
  • • Tidpunkt för senaste kontakt
FAST ACCESS (< 2ms)
TIME-SERIES TimescaleDB
  • • Alla relevanta datapunkter
  • • Historiska rörelsemönster
  • • Rekonstruerade rutter & stopp
  • • Körjournaler & rapporter
  • • Tidsserier för sensorhistorik
LONG-TERM STORAGE
UNIFIED LARAVEL API & REALTIME WEBSOCKET SERVICES
5. Trajectory & Trip Analytics

Real-time for the present. Time-series for historical depth.

Gelios does not simply overwrite an asset’s latest coordinates. The engine continuously aggregates, verifies, and constructs comprehensive trajectory histories for hundreds of thousands of vehicles.

Trip & Stop Reconstruction

Algorithms stitch raw GPS telemetry points into smooth trips, filter stationary sensor drift, and compute precise departure, arrival, and idling durations.

Real-time Geofencing

Polygonal and radial geofences are computed on the fly to detect arrivals, departures, unauthorized night shifts, and automated job site billing.

Time-Series Telemetry

Speed profiles, fuel rates, engine hours, CAN bus diagnostic codes, and external temperature sensors stored in high-density time-series hypercubes.

6. 24/7 Background Workload

A backend infrastructure that never sleeps

Traditional business systems or B2B SaaS web applications follow predictable human work patterns: high traffic during office hours, quiet evenings, and near-zero load during nights and weekends.

Gelios operates in the opposite paradigm. Hundreds of thousands of physical vehicles, heavy equipment, and telemetry units transmit continuously:

  • In the middle of the night
  • On weekends and holidays
  • When no administrator is logged in
  • When API clients are not actively polling

Not a single user needs to be logged in for the system to operate under full active load.

Standard Web App vs Real-Time Infrastructure

Standard Web Application
Load driven by active browser sessions. Idle after business hours.
Gelios Telematics Platform
Load driven by ~450,000 physical IoT trackers transmitting continuous telemetry 24/7/365.
7. Resilience & Reliability

Long-term operational stability beats a synthetic benchmark

Real-time telemetry is only valuable when it keeps functioning flawlessly under relentless real-world load. The Gelios architecture is engineered to isolate faults and eliminate bottlenecks.

01 / ARCHITECTURE

Specialized Microservices

High-concurrency Go and Node.js microservices dedicated solely to protocol decoding and telemetry ingestion.

02 / MEMORY

High-Velocity Caching

High-frequency tracker requests and live states served directly from memory to prevent database thrashing.

03 / ISOLATION

State & History Isolation

Real-time fleet maps query lightweight state without getting stalled by heavy analytical reporting jobs.

04 / TIME-SERIES

Time-Series Hypercubes

Automated hypertable partitioning and data compression for billions of data points with indexed time windows.

05 / BATCHING

Intelligent Batch Writes

Database writes are consolidated into micro-batches to maximize I/O throughput during sudden traffic spikes.

06 / ASYNC

Asynchronous Queuing

Heavy analytical jobs, alert rule evaluation, and push dispatches run asynchronously without blocking ingestion.

07 / RESILIENCE

Fault Isolation

An anomaly in a single device protocol or external webhook will never degrade core ingestion throughput.

08 / OBSERVABILITY

Continuous Telemetry APM

Real-time telemetry monitoring packet latencies, queue depths, buffer limits, and node health.

8. Architectural Evolution

Same product. Different technical challenges. Different tools.

Not all components of an enterprise platform have the same workload characteristics. Forcing high-frequency IoT traffic through standard web MVC frameworks inevitably creates severe bottlenecks.

Business logic, tenant permissions, user administration, and reporting thrive in robust application frameworks. In contrast, hundreds of thousands of high-frequency tracker requests require specialized services for:

  • • Device configuration distribution
  • • Telemetry ingestion & protocol decoding
  • • Realtime tracker state tracking
  • • Statistics & geospatial processing
  • • High-frequency MQTT message collection
  • • Ultra-low-latency cache-based request handling
BUSINESS APPLICATION (LARAVEL) ENTERPRISE LAYER
Users & Orgs
Permissions
Fleet Config
Reports & PDF
REST APIs
Integrations
PARALLELL SAMVERKAN I REALTID
REALTIME SERVICES (GO / NODE / MQTT) ~450k DEVICES
MQTT Ingestion
Binary Protocols
Tracker State
Config Cache
TimescaleDB Sync
Geofence Triggers
9. Performance Engineering Case

Scaling is not just adding more servers. Sometimes the path through the system must be completely re-engineered.

A prime example is device configuration distribution. When hundreds of thousands of trackers regularly poll for updated parameters, traditional request cycles cause massive unnecessary CPU load.

FÖRE (TRADITIONELL KEDJA) HÖG CPU-BELASTNING
1. Tracker skickar config request Inbound
2. Tung applikationsstart & routing Overhead
3. Databasfråga mot relationstabeller I/O Bottleneck
4. Ramverksserialisering & respons Outbound
Repeated hundreds of thousands of times daily, this created massive compute overhead.
EFTER (SPECIALISERAD TJÄNST + CACHE) MINIMAL CPU-LAST
1. Tracker skickar config request Inbound
2. Dedikerad Go/Node-service fångar anropet Sub-ms routing
3. Direktläsning från minnescache (Redis) Zero DB Query
4. Blixtsnabbt förkompilerat svar < 1ms latency
Dramatic reduction in compute resources and instant sub-millisecond response times.
10. Production Technology Stack

The Technology Behind Real-Time Scale

Each component in the architecture is selected for specific concurrency, throughput, and reliability requirements.

Application & APIs

Laravel / PHP

Robust enterprise application layer managing multi-tenancy, authentication, role policies, and REST APIs.

Realtime Services

Go & Node.js

Ultra-high concurrency compiled ingestion microservices handling continuous packet deserialization.

Telemetry Ingestion

MQTT / EMQX

Distributed high-throughput MQTT messaging broker purpose-built for low-overhead IoT telemetry.

Time-Series Telemetry

TimescaleDB

PostgreSQL-native time-series database managing billions of data points with automated chunk compression.

Operational State

PostgreSQL

ACID-compliant relational database for fleet registers, organization hierarchies, and business logic.

Fast State & Caching

Redis / Memcached

In-memory key-value stores for instant tracker position state, geofence caching, and config lookups.

Architectural Data Flow

DEVICES (~450k) MQTT / HTTP GO REALTIME SERVICES REDIS & TIMESCALEDB LARAVEL BUSINESS API EXTERNAL SERVICES
11. Operational Ecosystem

Infrastructure that external systems build upon

Gelios is not just a standalone user interface. The platform’s telemetry streams and APIs serve as the operational backbone for external commercial applications and public sector operations.

When external systems depend on continuous location data for day-to-day operations, system uptime, backward compatibility, and reliable API interfaces become paramount.

OPERATIONAL USAGE

Wide Range of Applications

  • Commercial fleet management and transport logistics
  • Public sector services and municipal vehicle coordination
  • Construction machinery, field service vans, and asset security
  • Third-party dispatch APIs and custom enterprise dashboards
12. Physical Hardware Layer

The system begins in the physical world.

The hardware tracker collects ground truth from the vehicle and its environment, initiating an unbroken chain that ends in an API, a live map, or an enterprise workflow.

VEHICLE
GPS • CAN • BLE
Sensorer & Tändning
TRACKER
Hårdvara & Firmware
Protokoll & Buffring
GELIOS CLOUD
Ingestion & State
TimescaleDB & Redis
13. Engineering Maturity

Engineering insights that cannot be learned from a tutorial

The definitive value we bring to our clients is real-world experience operating platforms of this magnitude. It yields actionable insights that only emerge under high-concurrency production:

01. BOTTLENECKS

Where real bottlenecks appear

Identifying hidden locking and I/O saturation before it causes production latency.

02. PARTITIONING

How data must be partitioned

Partitioning time-series chunks so analytics never block live state ingestion.

03. CACHING

When caching actually works

Designing cache invalidation without causing cache stampedes under high concurrency.

04. DECOUPLING

When to split the monolith

Pragmatically extracting only high-frequency bottlenecks into dedicated microservices.

05. ASYNC

Asynchronous execution

How background queues and micro-batches smooth out resource spikes effortlessly.

06. ZERO DOWNTIME

Architecture evolution in live ops

Migrating systems and upgrading protocols without a single second of production downtime.

14. The Alegor Philosophy

Reuse the standard. Custom-engineer what is truly hard.

The specialized Gelios real-time ingestion stack is distinct from the Alegor platform core. This is a critical architectural distinction.

Alegor is leveraged for universal application requirements: user management, tenant hierarchy, role permissions, audit logging, and core API structures. The specialized Gelios stack solves domain-specific engineering: high-throughput telemetry ingestion, binary tracker protocols, real-time state, and geospatial time-series data.

Alegor does not replace advanced engineering—it frees our team to focus engineering effort where it matters most.

ALEGOR ARCHITECTURE

Clear Separation of Concerns

ALEGOR CORE (Standardiserat) Auth, Permissions, Tenant Scoping, Admin UI, Notifications, Webhooks
GELIOS SPECIALBYGGT (Realtid) Telemetry Ingestion, MQTT, Binary Protocols, TimescaleDB, Live State
PROVEN IN PRODUCTION · AUGUST 2026
~450 000

active tracking devices in continuous live production

REALTIME
Live Ingestion

Continuous high-throughput telemetry ingestion and state processing 24/7/365.

HISTORY
Time-Series Archives

Trajectories, trips, and diagnostics archived into compressed time-series hypercubes.

DEPENDABILITY
Mission-Critical

Telemetry and APIs power daily operations for enterprise and public sector organizations.

“The definitive test of a real-time platform is not how it behaves with a hundred units in a sandbox. It is whether it continues to run flawlessly when hundreds of thousands of physical devices transmit data every single day.”
16. Key Results & Metrics

Concrete Facts from Live Production

No buzzwords – pure engineering outcomes and verified production stability.

~450 000

Devices in Production

The platform continuously ingests and processes data from ~450,000 physical trackers in live operation (August 2026).

Realtid + Historik

Decoupled Pipelines

Delivers sub-millisecond current state queries while managing deep historical trajectory data.

24/7/365

Uninterrupted Uptime

Engineered for continuous telemetry intake, even when zero human users are logged in.

Öppna API:er

External Integrations

Gelios serves validated telemetry feeds to mission-critical commercial and public sector platforms.

Evolutionär Arkitektur

Architecture Scaled with Load

Specialized Go services, Redis in-memory caching, and TimescaleDB hypercubes were introduced incrementally where live production demanded optimization.

Architecture Consultation

Does your platform need to perform when scale gets serious?

Building the prototype is straightforward. The real engineering begins when data volume, throughput, and dependencies surge. We design and scale business-critical systems built for real-world production.

Cookie Usage

We use cookies to enhance your experience, analyze website traffic, and measure marketing performance. Read more in our cookie policy.

Cookie Usage

Strictly Necessary Cookies

Essential for core website operation and security. Cannot be disabled.