Service-oriented computing is one of the key enabling technologies to enable the digital transformation of production systems, to communicate with each other and rapidly configure themselves to meet dynamic production needs. Service-oriented architectures (SOA) are also crucial to promote the horizontal integration of digital factories across multiple interleaved supply chains, forming the so-called Internet of Production (IoP). The increasing availability of services from multiple supply chains to be aggregated and composed into composite services is leading to a new service ecosystem, named Big Services. In this context, the traditional vision of service registries, with a flat organisation of services to match the mutual requirements of supply chain actors, is no more feasible. A more structured model is required, taking into account the distinction between domain-oriented atomic services, at the single actor level, and demand-oriented composite services, at the supply chain and IoP levels. In this paper, we propose the model of a distributed registry of data-oriented services in an industrial production network. The organisation of services in the registry is guided by multiple perspectives of the production network, namely: (i) the business goal of a real production network; (ii) the perspective on production data that is managed through the services; (iii) the data flow stages implemented through the services (that is, data collection, monitor, dispatch and display). The resulting portfolio of services is distributed over the production network, allowing each actor to preserve control over the owned data and enabling the dynamic composition of services at higher levels. A preliminary validation in a real case study has been performed to demonstrate the feasibility of the approach.
Recently, the digital transformation of smart factories is evolving towards a multiple supply
chains level, where actors of the production environment (e.g., production leaders, suppliers and
customers) are experiencing a horizontal integration across multiple interleaved supply chains,
forming the so-called Internet of Production or Cyber Physical Production Networks [
Quality control order Forger Supplier Nevertheless, this vision is still far from being realised in concrete production supply chains, which are still isolated from each other and do not allow data to cross the borders of actors cooperating in the production environment. Service-oriented architectures come to the rescue, enabling modular and reusable software infrastructures, platform-independent interactions between software components and information hiding for ensuring data sovereignty in a distributed environment.
For example, let’s consider the production network of deep and ultra-deep valves (Figure 1).
The production of the valves as the final product, their installation on-field and maintenance are
time-consuming and costly tasks. Moreover, valves are delivered on-demand in low volumes,
very often designed to serve the specific needs of customers. As for many similar production
networks, diferent categories of actors are involved: the production leader (e.g., the valves
producer); the raw materials suppliers (e.g., the forger); the suppliers of mechanical processing
tasks (who is in charge of machining raw materials provided by the forger to be assembled in
the valves); the suppliers of specific tools used in the production stages (e.g., to perform quality
tests on the valves). Actors may perform diferent tasks, requiring data owned by other actors
and services delivered across actors’ boundaries. Collaboration among them is crucial to deliver
on time high quality products. For instance, the production leader of a supply chain is interested
in the optimisation of the production scheduling, which involves all actors. Therefore, the latter
task requires the leader to (partially) access data about all the production phases and this can be
propagated across multiple supply chains. Knowledge about delays and their causes might be
useful to all actors, but each actor should carefully dispatch only the information about his/her
own machines that is strictly necessary, properly controlling data sovereignty across diferent
chains. Similarly, energy eficiency can be ensured by acting at the multiple supply chains level,
according to recent carbon footprint strategies [
In the considered example, diferent types of services can be used, with diferent purposes:
(a) atomic services, which refer to local services used to access, monitor, and process data
within a single actor; (b) composite services which involve the composition of atomic services
across actors within and between supply chains, to implement domain-specific processes or
to serve immediate, demand-oriented needs emerging from the supply chain customers. The
increasing availability of services from multiple supply chains to be aggregated and composed
into composite services is leading to a new service ecosystem, named Big Services [
The paper is organised as follows: cutting edge features of the approach compared to the state of the art are discussed in Section 2; Section 3 describes the model of the service registry; Section 4 presents the implementation architecture; the experimental evaluation of the approach in a real case study is sketched in Section 5; finally, Section 6 closes the paper.
Since the first stages of the Industry 4.0 revolution, service-oriented architectures (SOA) have
been proposed to foster the development of platform-independent, interoperable and
componentbased integration of Industry 4.0-compliant devices, machines or parts of production plants [
Business goals Energy consumption optimization Product quality
Monitoring Production Scheduling & Monitoring
Collect
Collect Collect
CollCecotllect
Dispatch
DisDpiastpcahtch
SDeDisrivpsipactaecthcthypes
Dispatch
Display
MoDniistoprlay Display
DMispolnaiytor displayMBoM Display
Monitor displayMBoMDisplay of IIoT devices in manufacturing systems has been discussed. Existing approaches already investigated the relationship between Digital Twins and the Product Lifecycle Management [19], also introducing the novel concept of Digital Thread, intended as the cyber side representation of a product, to enable the holistic view and traceability along its entire lifecycle [20]. Nevertheless, the idea of proposing a service registry, where data services are organised according to diferent perspectives (namely, the three perspectives of product, process and industrial assets) and categorised with respect to the data flow within the Cyber Physical Production Network, is a winning strategy to avoid the proliferation of ad-hoc service-oriented solutions. As a first attempt in this direction, we proposed to adopt the business goal, the perspective on production data and the high level action performed by the service as criteria to organise the service portfolio within the registry.
A methodology to guide the design of a multi-perspective data service registry in a Cyber
Physical Production Network has been described in [
Within the proposed data-oriented service registry model, the business goals correspond to composite services that can be implemented either within or across diferent supply chains. In the real case study, four business goals have been identified: production scheduling, process monitoring, sustainable energy consumption and product quality check. These composite services are composed by combining atomic services, managed by a single actor, specifically designed to access, share and visualize the data among all actors involved in the production network: (i) collect services, used by each actor to acquire data from the physical side of the production network, such as the scheduled production plan or sensor data about machine energy consumption; (ii) monitor services, used to detect anomalies that may lead to higher consumption or breakdown/damage of the work centers, production process failures delays or product quality issues; (iii) dispatch services, used to share data across the actors of the production network, such as the service to collect delivery dates from suppliers in production scheduling; (iv) display services, used to visualise data on dashboards or ad-hoc GUIs. Depending on the type of data that is managed by atomic services, they are further classified as product-, process- and asset-oriented.
Example. Figure 2 reports the set of atomic services used to compose a Production Order Scheduling service (POS, in brief) and a Production Status Monitoring service (PSM, in brief) in the production scheduling and monitoring business goal. The first stage of the POS service concerns the registration of the Engineering Bill of Materials (EBoM) and the Manufacturing Bill of Materials (MBoM) by the production leader, through the corresponding Collect under the product perspective. After the MBoM registration, process phases are scheduled for the product to be created (receiveProductionOrder Collect service) and resources are scheduled (registerResources Collect service). Therefore, a handshaking procedure starts (in the process perspective), where the production leader notifies the suppliers of the product parts to be assembled, according to the MBoM (notifyProductionOrder Dispatch service) and received from the suppliers the delivery date of each required part (collectProductionTimes Collect service). Finally, display services are exposed to visualise data on the process, such as the MBoM (displayMBoM service) and the production scheduling (displayProductionScheduling service). Monitoring of production advancement (PSM composite service) presents a diferent pattern of stages: possible downtimes are first collected from machines (collectMachinesDowntimes Collect service) and delays are identified or predicted (detectProductionDelays Monitor service). The machines downtimes and delays are sent to the production leader, who is interested in this kind of information (displayDelays and displayMachinesDowntimes services), since it has an impact on the production schedule of the final product.
Data and service access policies. In the registry, atomic services and composite ones
(POS/PSM) are organised into two levels. Furthermore, as detailed in [
Data Model
Collected
Data Collect Services
Monitor Services
Multi-perspective dashboards Data Repository Monitoring
Data
Production Scheduling Services
Dispatch services Energy Efficiency … Services
Product Quality Services
Display
Services Services repository the allowed operation on the object and its properties/fields (in terms of CRUD actions). Each actor must have access only to the data and services for which he/she is authorised, i.e., related to his/her own internal company performance and his/her role in the production network. For what concerns the invocation of collect and dispatch services, according to the sequence diagram that models the interactions between actors, token-based mechanisms (e.g., OAuth) can be adopted. Monitor services process collected measures, under the three perspectives, and raise alarms/warnings in case of values overtaking predefined thresholds or in case of concept drifts in the data streams. These warnings/errors are properly notified to target actors identified at design time (e.g., a problem on a work center may be notified to the owner of the machine, while a delay on a process phase execution may be notified to the actors responsible for the phase and possibly to the main producer). Regarding display services, they are provided through a web-based dashboard which incorporates authentication and authorisation mechanisms for users.
Figure 3 presents the architecture that implements the service-oriented approach described in this paper, including the four business goals introduced in the previous section. In the architecture each actor is equipped with (i) his/her own vision of the data that the actor can explore (data repository), organised according to the multi-perspective data model; (ii) a list of services the actor has at his/her disposal to interact with internal industrial assets or machines in the form of Cyber Physical Systems and a list of services exposed by the other actors (service registry); (iii) a web-based multi-perspective dashboard for data exploration and visualisation purposes. The data repository includes both structured data, stored within a relational database according to the multi-perspective model and semi-structured data, stored within a MongoDB NoSQL installation (Collected Data).
MongoDB database stores fine-grained measures collected as a continuous flow of data (data streams) from the production network (e.g., sensors data acquisition). For each measurement, a JSON document is registered, reporting the value of the measure, the timestamp, and the ID of the target entities. The data stream collected from a vibration sensor on a specific work center or component is an example of this kind of parameter. JSON documents can be organised in diferent collections with respect to the physical parameter that is being measured (vibration, electrical current, temperature). Fine-grained processing data can be collected internally, from resources and industrial assets owned by the actor, or externally, provided by the other actors.
Display services populate the web-based multi-perspective dashboard that each actor uses
for data exploration. From the home page of the dashboard, it is possible to start the data
exploration by following one of the three perspectives, namely, product, process and industrial
assets. Each perspective brings to a UI component (tile) implemented using ReactJS libraries: i)
the product synoptic tile allows exploration from the product perspective of each single actor; ii)
the process phases tile allows exploration from the process perspective of each single actor; iii)
the working centers tile allows exploration from the industrial asset perspective of each single
actor. More details on the web-based multi-perspective dashboard can be found in [
The evaluation of scalability has been carried out in the context of an industrial research project, where business goals have been identified. In particular, tests have been focused on POS/PSM services, aimed at registering a production order in the database of the production leader and monitoring the production order progress over the supply chain, respectively. Such services have been chosen due to their complex business logics in terms of executed operations and interactions between actors in a supply chain. In the experiments, response times have been measured for the aforementioned services by varying (in terms of the number of units) a pair of factors that may hamper service scalability: (a) the complexity of the product, represented as the number of components within the product Bill of Materials (BoM); (b) the complexity of the supply chain, intended as the number of actors involved in the production process. Results are shown in Figure 4.
The experiments have been run on a PC equipped with Windows 11 Home, AMD Ryzen PRO
4650U, RAM 8GB. Services are based on the multi-dimensional data-model presented in [
Figure 4(a) reports response times of the POS and PSM services, by varying the number 2.S5calability of POS/PSM service wrt the complexity of the product
POS service
POS service 2.4 PPSSMM sseerrvviiccee Scalability of POS/PSM services wrt the complexity of the supply chain 2.6 PSM service
PSM service 2.5 PPOOSS sseerrvviiccee of components of the product BoM. The performance of the service-oriented architecture decreases as the number of components increases, but scalability is preserved. In particular, in the experiments, it has been observed that the response times are upper bounded by the complexity of the supply chain, more than the number of components in the BoM. On the other hand, as demonstrated in Figure 4(b), as the number of actors increases, the POS and PSM services still remain scalable.
In this paper we propose the model of a registry to organise data-oriented services in a production network, according to diferent perspectives, namely: (i) the business goal of a real production network (e.g., production scheduling, sustainable energy consumption, process monitoring and product quality control); (ii) the perspective on production data that is managed through the service (e.g., the industrial assets owned by actors in the network, the product over its lifecycle, the production process); (iii) the data flow stages implemented through the services (that is, data collection, monitor, dispatch and display). The resulting portfolio of services allows actors of the production network to compose complex services that implement high-level goals, while also maintaining control over their owned data.
The proposed approach will be further developed to increase the standardisation level through
the introduction of Semantic Web technologies to represent data on which services are built,
as well as standard service taxonomies and a more in-depth investigation on domain-oriented
and demand-oriented service composition, according to the newly emerging Big Services
paradigm [
Transactions on Industrial Informatics 15 (2018) 2405–2415. [20] T. Margaria, A. Schieweck, The Digital Thread in Industry 4.0, in: Proceedings of Int.
Conference on Integrated Formal Methods (IFM), 2019, pp. 3–24.