22 - Electronics - IoT

MQTT - OPC-UA - LoRaWAN - Node-RED - InfluxDB

From field to cloud. Every reading, every second.

-- Definition

IoT (Internet of Things) is the discipline of connecting your machines and sensors to the internet and aggregating data into a single panel. Industry 4.0 is the use of that data for production decisions, maintenance prediction, and automated control. The Senkronix IoT platform covers everything end to end -- from MQTT/OPC-UA/LoRaWAN protocols to digital twins and predictive maintenance -- delivered by a single team.

N° I -- Why Custom IoT?

Plug into the cloud,
then get locked in.

AWS IoT, Azure IoT Hub, Google Cloud IoT, and similar off-the-shelf platforms provide strong capabilities -- but their per-message pricing multiplies cost as you scale. Receiving one message per second from 10,000 sensors can produce 30,000-80,000 TL per month in cloud fees. Worse: your data lives in that platform's data model and is processed through that platform's services, making migration to another provider a rewrite.

At Senkronix the IoT platform is built vendor-agnostic. On your own server if desired, on a private cloud if desired, or hybrid. The data is yours, the schema is yours, it is portable. Core components are open source (EMQX, InfluxDB, Grafana, Node-RED, TimescaleDB), so while cloud prices may rise in five years, the platform will not shut down. On top of that we add your business rules -- you are tied to your business, not a vendor.

Advantages of custom IoT

No per-message fees -- on-premise or fixed cloud cost
KVKK compliant; data can be hosted in Türkiye with no cross-border transfer
No vendor lock-in -- every protocol and tool used is standard or open source
Edge computing -- process data in the field, send only summaries to the cloud; bandwidth savings
Retrofits to existing machines -- legacy OPC DA, Modbus devices are converted
Long-term ownership -- total cost advantage over 10+ years of scale

N° II -- Modules

Eight layers.
From sensor to decision.

-- Device Network Topology / N° II-A

Five layers · From sensor to dashboard

● Topology node→ Data flow

Device Management

Device registration, identification, configuration, OTA firmware updates, health monitoring.

Data Ingestion

MQTT (EMQX, Mosquitto), OPC-UA, HTTP/REST, CoAP, LoRaWAN Network Server, TCP/UDP.

Edge Computing

Data filtering, summarization, local rules engine, offline buffering, and synchronization on field gateways.

Time-Series DB

InfluxDB, TimescaleDB, QuestDB; high-volume data, automatic downsampling, retention policies.

Real-Time Dashboards

Grafana, custom dashboards, live charts, maps, sensor mapping on 3D models.

Alarms and Automation

Threshold alarms, anomaly detection, SMS/email/WhatsApp notifications, rules engine (Node-RED).

Digital Twin

Virtual copy of the machine/facility; sensor data flows onto the 3D model, virtual testing is possible.

ML and Predictive Maintenance

Anomaly detection, failure prediction, energy optimization; models based on Python, TensorFlow, PyTorch.

N° III -- Protocols and Technology

Field protocols are many.
The right one is singular.

-- Protocol Map / N° III-A

Six protocols · IoT core

● Protocol core□ Communication family

The real complexity of IoT is at the protocol level. Is the sensor battery-powered? What is the range? How many messages per second? These questions determine which protocol to select -- the wrong choice forces starting over. During Discovery Senkronix evaluates all these variables together.

Communication protocols

MQTT -- Lightweight, publish/subscribe, the default choice for most IoT deployments. EMQX, Mosquitto, HiveMQ brokers
OPC-UA -- Industrial standard; direct communication with PLC, DCS, SCADA systems
LoRaWAN -- Long range (10+ km), low power; smart city, agriculture, meter monitoring
NB-IoT / LTE-M -- Carrier-based; balance of network coverage and battery life
Zigbee / Thread / Matter -- Home/building automation; mesh network topology
Modbus TCP/RTU -- For legacy industrial devices; PLC, RTU, inverters
CoAP, HTTP/2, WebSocket -- For web-connected devices

Data layer and tools

Broker and Gateway: EMQX (MQTT), Node-RED (visual programming), Kepware (OT integration)
Database: InfluxDB, TimescaleDB (time-series); PostgreSQL, MongoDB (metadata)
Visualization: Grafana, Apache Superset, custom React dashboards
Stream processing: Apache Kafka, Redis Streams, Apache Flink
Cloud options: On-premise, AWS IoT Core, Azure IoT Hub, Turkcell Cloud -- all supported
ML/AI: Python, Scikit-learn, TensorFlow Lite Edge, ONNX Runtime

N° IV -- Who is it for?

Every sector with machines.

Scenario - 01

Manufacturing Facilities

Machine monitoring, OEE calculation, energy metering, production-line anomaly detection. The first step of the Industry 4.0 transformation.

Scenario - 02

Smart Buildings and Facilities

Single-panel monitoring and automated control of HVAC, lighting, access control, energy, and security systems.

Scenario - 03

Agriculture and Greenhouses

Soil moisture, air temperature, automated irrigation control, precision farming integrated with drone and weather data.

Scenario - 04

Energy and Infrastructure

Smart meters, electricity/water/gas network monitoring, substation control, solar panel performance tracking.

Scenario - 05

Smart City

Parking, air pollution, traffic, waste bin, and lighting sensors. Optimization of municipal services.

Scenario - 06

Fleet and Asset Tracking

GPS, OBD-II, driver behavior, maintenance reminders, fuel tracking; for logistics firms and field teams.

N° V -- Frequently Asked Questions

Clear questions,
clear answers.

Which protocol should we use between field and cloud?+

General rule: OPC-UA for in-factory PLC/industrial devices, HTTPS/MQTT for web-enabled devices, LoRaWAN or NB-IoT for wide-area low-power sensors. In complex projects several protocols coexist and are merged at a gateway. During Discovery we decide together based on range, message frequency, battery life, coverage, and cost.

Our machines are old. How do we add them to IoT?+

We retrofit. An IoT gateway is added (Raspberry Pi, Siemens IOT2050, Moxa gateway, etc.) that reads legacy Modbus RTU, RS-232, or meter outputs and converts them to modern MQTT/OPC-UA. Non-invasive retrofits are also possible -- the machine is monitored via added current sensors, vibration sensors, and temperature sensors.

Will data volume become huge? What about storage cost?+

Yes, it multiplies if unmanaged. 1,000 sensors at 1 message per second equals 86 million records per day. With InfluxDB/TimescaleDB we apply downsampling (per second → per minute → per hour), compress cold data, or move it to S3/MinIO object storage. Smart retention policies typically deliver 80%+ storage savings.

Does predictive maintenance really work?+

Yes, but it requires enough historical data. ML models typically produce meaningful predictions after 6-12 months of data collection. Success rates depend on the hardware: pump/motor vibration analysis can achieve 85%+ early-failure detection. More complex systems start around 60% and improve over time. ROI typically materializes within 12-18 months.

On-premise or cloud?+

Both are possible. On-premise: your server, fixed annual cost, full data control. Cloud: flexibility, low entry cost, but monthly subscription. Hybrid: critical data on-premise, reporting and ML in the cloud. For data sovereignty and KVKK reasons, most of our industrial customers prefer on-premise or hybrid models.

How is security handled?+

Multi-layered: TLS/DTLS encrypted communication at device level, authentication with X.509 certificates, secure boot, anti-tampering. At network level VPN tunnels, firewalls, DMZ. At server level access controls, audit logs, penetration testing. Segregation of OT (operational technology) and IT networks (air gap) is standard in critical projects.

N° VI -- Process

Four stages.
Each one documented.

Discovery

We walk the site, inventory existing equipment, and document measurement needs and KPIs.

Output: Inventory - KPI - Protocol Selection

Design

Architecture: device-gateway-cloud-user layers, data schema, security plan, scaling projection.

Output: HLD - Data Schema - Security

Development

Two-week sprints, pilot line deployment, live data validation, dashboard iteration.

Output: Pilot - Demo - Git

Launch and Support

Rollout, operator training, ML model calibration, continuous monitoring. 24/7 support under SLA.

Output: Live Platform - SLA

-- IoT Proposal Desk

Tell us about your field data.
The dashboard we will build together.

Request an IoT Proposal 444 27 14

bilgi@senkronix.com - Karatay / Konya