Leveraging DMWay middleware to manage customised data models

Leveraging DMWay middleware to manage customised data models

Illustration of the Sensirion brand air quality sensors support

Today, connected objects use a multitude of communication protocols, proprietary or not, to transmit their data through IoT networks.
Managing these heterogeneous data sources is tedious because the backend applications need suitable connectors to support the multitude of protocols each requiring support for their own data formats. Users often find themselves constrained by software solutions hindering them from interacting flexibly with their sensors in all their diversity, sometimes even for sensors from the same brand.
CETIC proposes its middleware DMWay which allows the exploitation of data sources associated with communicating sensors and supporting their protocols diversity while at the same time neatly uncorrelating data management from hardware specificities (brand, version, physical interface, etc.).

Date: 8 May 2024


Scalability of embedded systems and IoT networks 

Domaine: Software industry 

Innovation theme: Edge Computing 

Asset: Dmway 

About project: Marina 

In this article, we focus on a fairly common family of sensors: those using serial communications such as UART [1], SPI [2], I2C [3] or 1-Wire [4]. To illustrate our approach ofintegrating and exploiting their data, we leverage data modeling capabilities of the DMWay middleware and we chose, as example, to implement air quality sensors from Sensirion and customise a data model for the intended use case.

Data model specific to Sensirion sensor brand

DMWay offers its built-in IoT data models like LWM2M [5] or OCF [6] but also allows users to encode their own data models matching their needs.

In our case, this involves configuring a data model that takes into account the air quality data sent by the Sensirion sensors.
the two sensors SEN55 and SPS30, both measure the concentration of PM1.0, PM2.5, PM4 and PM10 particles. SEN55 also allows to measure the level of VOC (Volatile Organic Compounds), NOX (nitrogen oxides) as well as temperature and humidity which is also used by the sensor’s internal algorithms to automatically adjust the measurements.

The first step consists of configuring the data model that supports all our measurements. The figures below show the DMWay configurator that enables user to define the required parameters.

DMWay data model configurator, pm1 particles example

The following figure shows the mapper which is the middleware module that connects the data sources to the information model or infomodel. The latter represents a central concept of DMWay middleware. The infomodel consists of a hierarchical structure of objects. These objects can be values, actions, events or composite objects (i.e. objects containing other objects).
In DMWay, the concepts of Datamodel (defining the meaning of manipulated data) and Infomodel (defining the organisation of data) are strongly linked. The detailed explanation is outside the scope of this article. The important thing to note is that DMWay inherently offers the means to specify and customise the representation of data using these two concepts.

Mapper configurator For SEN55 sensor

Simultaneous handling of sensors data from different protocols

As DMWay is compatible with a multitude of IoT protocols, it offers the user a wide freedom to choose their sensors without worrying about explicitly managing from end-to-end, the type or communication technology used by the sensor.
In our case, the Sensirion SPS30 is connected to DMWay through a serial connection and the data from the SEN55 is transmitted to DMWay via a programmable gateway (an ESP32 module configured to encapsulate serial measurements through a Wifi link). The firmware developed for the ESP32 interrogates the SEN55 sensor through the I2C link and sends the data to DMWay through MQTT protocol as shown in the following figure.

PoC architecture

The figure below shows the dashboards created with Thnigsboard, based largely on our Sensirion air quality sensors.

Thingsboard air quality dashboard


This PoC illustrates the flexibility of DMWay in allowing specification and management of personalised data models. This helps satisfying specific user requirements in their data management processes and allows easier and straightforward exploitation by IoT data platforms such as Thingsboard or other backend software.

Interested in learning more about the capabilities of DMWay middleware or seeing it implemented on a live demonstrator?
Do not hesitate:
to contact us on info@cetic.be
to visit the tool page: https://asset.cetic.be/en/dmway/
or to visit our demonstration space at our address [Avenue Jean Mermoz 28, 6041 Charleroi (Belgium)]


[1UART: Universal Asynchronous Receiver / Transmitter

[2SPI: Serial Peripheral Interface

[3I2C: Inter-Integrated Circuit

[41-Wireor OneWire

[5LWM2M: Light Weight Machine to Machine

[6OCF: Open Connectivity Foundation