The Internet of Things (IoT), and in particular the Industrial Internet of Things (IIoT), is a powerful, flexible technology being used and adopted by numerous industries. But effective deployment and integration of sensors and devices, particularly alongside other complex technologies such as edge computing and cloud computing, has proven challenging.
Many companies and organisations have tackled these challenges, with varying degrees of success. As a result, there has been a rapid and widespread proliferation of methodologies. New challenges have emerged from this abundance of methodologies, particularly when it comes to taking edge-data, feeding it to IoT and IIoT devices, and synchronising communication and coordination via the cloud. Interoperability and convergence have proven difficult issues to resolve.
This is the issue the NEPHELE Research Project is tackling with:
- An IoT and edge computing software stack
- A synergetic meta-orchestration framework
The objective is to create a system-of-systems meta-orchestration approach that will allow developers to effectively integrate IoT devices into a distributed computing framework and manage the deployment of distributed applications over heterogeneous infrastructure.
From Research Work in a Lab Towards International Cooperation
The NEPHELE Research Project was inspired by research work done at the Network Management and Optimal Design Laboratory (Netmode) of the National Technical University of Athens, and developed by considering existing works and collaborating with other research partners across Europe. We work a lot on IoT and edge computing technologies in the Netmode laboratory at the university, considering computing resources that span across the computing continuum. So, many of the challenges industries face were already familiar to us.
Based on that knowledge and experience, we built a consortium, which includes universities, small and medium to large enterprises, as well as the open source and standardisation community.
From there we were able to apply for funding, which led to the current three-year project that started September 1, 2022 and wraps up by the end of August 2025. The project has made a lot of progress so far thanks to collaboration with various open source communities and integration with the W3C Web of Things (WoT) community, and Eclipse Thingweb, which provides components for leveraging W3C WoT standards.
Minimising Responsibility With a System-of-Systems Approach
The project takes a two-part approach to simplifying edge-to-cloud IoT interoperability and cloud convergence.
The first component is an IoT and edge computing software stack. This stack allows for virtualised IoT devices to be leveraged at the edge of a network infrastructure. It supports openness and interoperability in a device-independent way.
The second component is the meta-orchestration framework, which enables management of cloud and edge computing orchestration platforms via high-level scheduling supervision and definition.
Together, these systems provide the system-of-systems approach that enables NEPHELE to simplify deployment and management of distributed applications. The top-level entity is responsible for deployment and has overall responsibility, but control is distributed across the various entities and their various layers.
This also includes both virtual objects (VO) and composite virtual objects (cVO) that act as the virtualised instances of the IoT devices, allowing for the software stack to interact with IoT devices via these VOs and cVOs as needed.
Composite virtual objects are particularly interesting. These cVOs combine inputs from multiple VOs, which can help produce much more robust decision-making and integrate with more complex systems.
For example, if you had a forest fire monitoring system that included cameras to watch for smoke, as well as temperature sensors throughout a forest. Combining inputs would be necessary to ensure possible fires aren’t missed while also ensuring the system doesn’t get bogged down by false positives from warm spots or small patches of smoke that don’t necessarily indicate fire. These cVOs enable that combination of inputs.
This capability is quite similar to the role of digital twins, and the NEPHELE software stack supports the integration of digital twins as well.
Also worth noting is that the software stack supports multiple communication protocols, including MQTT, HTTP, and CoAP, and work continues to provide support for semantic interoperability with OMA LwM2M and NGSI-LD specifications. This helps resolve many of the interoperability issues in IoT-to-edge-to-cloud systems.
With all these capabilities, NEPHELE provides a full, end-to-end device and industry-agnostic software stack that can be used for connections between views and applications, as well as tackle semantic interoperability.
Project Proof of Value of Open Source Planning
Undoubtedly, what we’ve been able to accomplish thus far in the NEPHELE project is proof of the value of integrating open source communities into the process. Existing open source implementations of virtual objects in Python (specifically the wot-py implementation of the W3C WoT working group) and in Java (implementation provided by the colleagues in CNIT) was actually the starting point for the project. The current implementation is available at the provided resources to developers from the W3C WoT under the naming VO-WoT.
The open source connection has paid dividends throughout development. We’ve been able to access very rich information related to virtualisation thanks to our open source connections, as well as being able to use some very excellent specifications during development. Plus, the quality of the code ends up being much higher with all the contributions from the open source community.
Nevertheless, it was a different process than what many of us were used to when leading research projects within universities or other closed-source institutions. Maintaining engagement and repositories is a different challenge that isn’t as relevant in those spheres but is crucial for a successful open source project. I think it’s crucial for any research project considering open source integration to plan for it from the outset, particularly with how they’re going to manage the project and keep the development process going.
Functional Code and Use Cases Available But Work Ongoing
Having crossed over the halfway mark for the research project, we’re proud to say that a lot has been accomplished already. The architecture described above, while still undergoing a refinement and improvement process, is available and functional.
We’re also proud to say that throughout the process, we’ve been able to give back to the open source community. Apart from the NEPHELE codebase itself, we’ve tried to ensure we disseminate the ideas and outcomes with the community.
If you’re curious about how and where the NEPHELE project can apply, we currently have four use cases available:
- Disaster recovery
- Logistic operations in ports
- Smart cities
- Telehealth services
We are, however, expanding these use cases based on contributions from open calls to third-parties to develop their own industry-specific solutions based on the available software. Thus far we’ve received input from organisations in agriculture, electric vehicles, climate optimisation, haptic communications, and wearable devices. And we expect to see more with our second open call, which is open until 20 October.
The best place to learn more and get involved with the project, if you’re interested, is the website, which has a plethora of resources. The development of the open source stack is ongoing in the Eclipse Research Labs GitLab.
About the Author
Anastasios Zafeiropoulos
Anastasios Zafeiropoulos is a postdoctoral researcher at the National Technical University of Athens and the technical manager for the NEPHELE Project. Contact him at tzafeir@cn.ntua.gr.