November 2024 - The European Commission (EC) recently organised a workshop on the topic of ''Enhancing IoT Semantic Interoperability by SAREF for Digital Twins''. The workshop focused on the applicability of the Smart Applications REFerence (SAREF) ontology to the emerging field of digital twins. SAREF is a framework that enables smart device interoperability in IoT ecosystems.
In this interview, Laura Daniele of TNO (NL) describes EC and industry initiatives leading up to this workshop. She also explains why consumer appliance manufacturers, building managers and smart grid stakeholders want to learn about exchanging information collaboratively between connected devices, platforms and systems as well as across operational processes.
Q: Would you begin by telling us about your background and interests in the IoT domain?
LD: I am originally from Italy and now living in the Netherlands. I came here to study for my PhD and stayed on after joining TNO where I work as a Research Scientist. My day-to-day activities involve projects and collaborations with industry bodies, universities, and standardization organizations. I also participate in big European Union (EU) research and innovation projects. I also co-lead a group on semantic interoperability for the Alliance for IoT and Edge Computing Innovation (AIOTI), which brings together semantic experts across the world.
My background is in electronic engineering and the classics, which includes topics such as language and philosophy. This combination plays a significant role in my research interests in two ways. One is to help stakeholders in a system or business environment to talk together and to communicate information in a meaningful way. The second is the creation of models using ontology engineering techniques. These models are the raw materials for cross-stakeholder communications. An ontology, based on the definition given by Gruber in the 90’s, is an explicit specification of a certain domain of interest, a view of the world that we wish to represent for some purpose. In the IoT domain, you can think of ontologies as models of smart or connected devices that provide the structure to communicate about their attributes, their relationships to other relevant entities and the logical constraints among them.
Q: And how would you describe TNO and its activities?
LD: TNO is the largest independent research organization in the Netherlands employing about 4,300 people operating with a budget of about EUR680m. Roughly half comes from state funding and the rest is market revenue from commercial and research contracts.
TNO’s tagline is “Innovation for Life.” We have a unique role in applying our expertise to connect stakeholders in large information ecosystems. I work in TNO’s ICT Strategy & Policy unit, one of six units. My unit differs from others in that we take a horizontal perspective whereas others have vertical domain specializations.
Q: What made you develop an interest in ontologies?
LD: Modeling and design are core areas of expertise in my professional work, and I also had experience of working in a group at the forefront of research into ontologies during my PhD. These elements came together in an early project at TNO in the logistics domain where information sharing was key. Things took off after that when the EC funded a project on ontologies for smart appliances that led to the development of SAREF.
Q: Before we speak about SAREF, would you share an example that illustrates how an ontology enables interoperability in IoT systems?
LD: Let us think about a situation where different parties want to talk to each other. They need a way to exchange messages, which is a form of interoperability. Practical ways to exchange messages can involve the postal service, email, or WhatsApp, for example. We refer to this as ‘connectivity,’ and it represents the lowest level of technical interoperability.
If we want our correspondents to make sense of our messages, we need to agree on a language which consists of symbols and meanings. Let us say our stakeholders agree on a language of choice and its alphabet. That gives them a common set of symbols and words. As an example, we might agree to send data either in text (twenty-two) or in numerical (22) form. This corresponds to the next level of interoperability, what we call ‘syntactic’ interoperability.
Next, we focus on ‘meaning’ so that message senders and receivers have an agreed way of interpreting messages. Based on what our messaging parties agree, the data we send is commonly interpreted as a specific numerical value (22) because of the meaning we associate as humans to the symbol ‘22’ or the word ‘twenty-two’. Moreover, 22 could stand for the age of a person or a temperature reading using the Celsius scale, which as humans we are able to infer from the context. However, machines need an explicit representation of this information, therefore we need to specify in formal language that the data represents a temperature reading using the Celsius scale. This is an example of semantic interoperability.
To summarize, connectivity and networks allow us to exchange messages. On top of that, a data format, or form of encoding, provides syntactic interoperability. In the IoT domain, JSON and XML are examples of encoding formats. On top of everything, we add ‘meaning’ through ontological capabilities using a formal language, for example the Web Ontology Language (OWL) based on description logic and standardized by W3C.
In the IoT context, we can define an ontology for different connected devices in terms of the measurements they make, the functions they support, the services they provide and the commands they can issue. Knowing all of this, a developer is better able to interact with that device. This is the chain of logic that led to the SAREF initiative.
Q: Let us now turn to SAREF. What is the history of its development?
LD: I mentioned some earlier work in the logistics sector. That did not progress far because it was difficult to persuade industry about the need for interoperability and the technology, in the 2010’s, was mature in the research community but not for the market. Nevertheless, the need for interoperability did not go away.
SAREF came from the home appliances industry and was facilitated by the EC. Vendors saw the need to improve communications between home appliances because households rarely buy everything from one supplier, even if you could find a supplier that sells every conceivable home appliance. Following an industry consultation, the EC launched a tender to create an ontology with two goals. One was to enable semantic interoperability and the second was to chart a path to standardization.
TNO submitted a response based on our expertise in connectivity, protocols, and ontology. The standardization part of the tender brought us into contact with ETSI and its SmartM2M technical committee. For the project itself, we consulted widely with associations such as EEBUS in Germany and Energy@Home in Italy. From these associations, the BSH Home appliance Group and Whirlpool were the most interested and active manufacturers that became industry pioneers for SAREF. We also spoke to businesses such as NEC, Deutsche Telekom, France Telecom, Telecom Italia, and technology representatives for the OSGI Alliance, Zigbee and Z-Wave, among others.
Our market analysis covered fifty technical protocols and ontologies from which we identified many commonalities and a rationale for standardization. That work led to the publication of SAREF in 2015. After that, the EC launched another tender to study opportunities beyond the home. This focused on the energy industry and the challenge of managing flexibility in the demand side of the supply chain that connects households to the energy grid and eventually to an array of energy stakeholders. For this project, TNO worked with energy sector bodies such as ESMIG, the European Smart Metering Industry Group, and DNV-GL. One of the outcomes was a first working demonstrator showing the use of SAREF with actual devices.
In the meantime, ETSI and the EC funded several Specialist Task Forces (STFs) on SAREF. Now, SAREF has become a modular framework that comprises a generic core ontology for IoT and twelve domain-specific extensions, including SAREF for Energy. The framework is maintained and evolved by experts from several organizations under an ETSI umbrella. Users and researchers can access a portal that is open to non-ETSI members where stakeholders can contribute directly to the SAREF evolution.
With growing interest and a desire to scale, TNO later responded to a project call under the Horizon-2020 program for large-scale pilots. We participated in the InterConnect project with about fifty other partners from across Europe. This EUR36m project united many distributed projects and efforts from prior years into a large-scale project that had to work across different countries and different regulatory regimes. We developed and demonstrated advanced solutions for connecting and converging digital homes and buildings with the electricity sector, using SAREF as main pillar. Companies such as Bosch, Siemens, Miele, Whirlpool, Daikin, Vaillant built SAREF into their products. We also validated SAREF-based interoperability in seven connected large-scale test-sites in Portugal, Belgium, Germany, Netherlands, Italy, Greece, and France. TNO currently continues to support large-scale deployment of SAREF-based interoperability in the EUR22m Horizon Europe Hedge-IoT project, which promotes a Holistic Approach towards Empowerment of the DiGitalization of the Energy Ecosystem through adoption of IoT solutions.
Q: What market adoption challenges do you foresee for ontology and semantic interoperability techniques?
LD: Before I come onto challenges, let me comment on the opportunity for interoperability in a world where IoT technologies are going to be common. From my experience, the leading sectors or industry verticals are the home environment, the built environment (for smart building and interdependencies between buildings), and the energy and smart grid sectors. In the case of the latter, we need to think about how renewable sources and distributed generation are going to affect the grid. We also need to consider how building managers manage their buildings and how homeowners manage home automation and energy use. Each sector benefits from ontologies and each sector interacts with others, adding to the value of ontological approaches.
As for the challenges, the most immediate ones are to explain semantic interoperability and why organizations should embrace it. Developers play a pivotal role in the process. However, they seem more interested in working with APIs at the syntactic level (using JSON, XML, Rest API) than in in learning and applying semantic techniques. This means that their applications send data in correctly formatted messages but there is no commonly shared meaning which is problematic when working across boundaries. There might be an element of fear that change will disrupt their roles and that their expertise, techniques, and platforms could be eliminated. That is not how I see it. If a developer needs a device to communicate with another one in a meaningful and interoperable way, then it is necessary to add something to what already exists. Semantic interoperability allows them to add this extra value on top of what they currently do.
Ultimately, the market is reaching a point where stakeholders see value in exchanging information. We see that happening through more use of the word ‘ontology’ beyond academic and research circles. There are many references to SAREF in academic publications and several of the EU’s research and innovation projects are using SAREF out of choice, after seeing its use in the InterConnect project. The challenge now is for the applied research community to make ontologies easier to use and more accessible.
Q: Are these among the issues discussed at the September EC workshop?
LD: Yes, to some extent. The focus of the workshop was to report on the conclusion of an ETSI project (Specialist Task Force #641) and to discuss the future of SAREF. STF 641 was a project to harmonize and consolidate earlier work on semantic interoperability and analyze the potential to extend SAREF as an ontology for digital twins. The project results are promising and confirm that there is a path forward to combining SAREF and digital twins. The ETSI TR 103 827 provides a survey of existing Digital Twin definitions, standards and ontologies. It reveals a highly fragmented landscape. The technical report contains an analysis of six use cases in which SAREF and (some of) its extensions can be used. The technical specification document, ETSI TS 103 828, contains a discussion of gaps in the current SAREF suite of ontologies to represent Digital Twins and shows how to apply the SAREF suite to two of the selected use cases.
Q: What are the next steps for SAREF?
LD: We have made a lot of progress with SAREF through a project-based approach. Now, we need to evolve to something that is continuous. It would be good to set up a user group to share information and experiences. Having taken over the maintenance of SAREF, ETSI obviously plays a key role and sends a positive signal to industry. Within industry, we need to see more progress beyond research so that the base of adopters grows and becomes self-sustaining.
We also need to continue to publicize the core message around semantic interoperability. Specifically, we are adding a layer of value on top of connectivity and syntactic messaging. Keep in mind that appliance vendors do not change anything in their products, so there is no additive cost. However, there needs to be a commitment to adding value on top of the connectivity and messaging layers that already exist. The use of (software) adapters brings these devices into the world of semantic interoperability. This is currently not an industry best-practice, and the availability of supporting tools is lacking. In my opinion, the existence of a user group, with committed founding members who can channel their findings into standardization will also be key.
Q: We have mostly discussed developments in Europe. What are the application opportunities outside of the European community?
LD: The framework I used earlier in our discussion put connectivity at the first layer of interoperability. I also mentioned the smart home as one of the early markets that motivated the development of SAREF. We can observe a similar pattern developing in the case of the Matter protocol that is emerging from the USA via the Connectivity Standards Alliance (CSA). Their members have begun mapping the Matter protocol to SAREF and SAREF4Energy ontologies.
The motivation for this arises from the EU’s Code of Conduct for Energy Smart Appliance manufacturers. As of April 2024, manufacturers can add a product compliance marque in the European Product Registry for Energy Labelling. Ten manufacturers have committed to the initiative and their products will carry a marque in this registry to signify that that they are “interoperable products.” This is an important development that will hopefully stimulate innovation in semantically enabled applications.
Q: Do you have any closing advice for readers or recommendations for how they can learn more about SAREF?
LD: Do not feel scared or threatened to add semantics to your IoT applications. It requires some effort, but it always pays off when the need for (cross-domain) interoperability is strong. We have seen that organizations that are skeptical at the beginning became the best ambassadors afterwards.
Guidance and tooling about semantics are improving all the time. To understand the process of getting there using SAREF, a new European Norm (EN 303 760) “SAREF Guidelines for IoT Semantic Interoperability to develop, apply and evolve Smart Applications ontologies” is underway.
To become familiar with the SAREF framework, visit the official ETSI portal at https://saref.etsi.org/ and read the Technical Specifications of the various ontologies. Please also consider that SAREF experts at ETSI SmartM2M, including TNO, are available for support.
Interested parties can also join the AIOTI’s semantic interoperability group, which meets bi-weekly. The group is organizing a workshop in February 2025 on semantic interoperability and digital twins, and in the past has published useful resources such as a catalogue of ontologies for IoT.
For appliances manufacturers and energy management system providers it is certainly of interest to join the second phase of the EC Code of Conduct. Finally, in the context of the Hedge-IoT project, TNO and partners are developing new training material and tools for external stakeholders according to the latest SAREF and SAREF4ENER specifications, so stay tuned via LinkedIn!
