Enterprise integration
In the enterprise context, managing and optimizing resources is achieved through a service-oriented approach leveraged by the use of the Cyber-Physical Systems paradigm. Cyber-Physical Systems are systems of cooperating computational agents that possess a high degree of connectivity with the surrounding physical environment and ongoing processes [87]. Cyber-Physical business processes orchestrate the actions of IoT (Internet of Things) devices and embedded ICT (Information and Communications Technology) systems (e.g., smartphones, sensors), and in doing so, they strongly determine the coordination of real-world entities (e.g., humans, robots, etc.) [88]. The Openfabric platform provides integration tools for augmenting the enterprise environment with a cognitive layer.
Through this, the enterprise application can access Openfabric AI services that are used for the better management and optimization of Cyber-Physical Systems processes, resources, and activities. Promoting innovative changes in business ecosystems often represents a disruptive, costly, and inefficient process, and is one of the most significant impediments in the adoption of new technologies. The proposed integration model offers a symbiotic, non-invasive approach with a minimal impact on existing processes. The integration is achieved by providing extension points at the Workflows and Microservices [89] level. Fig. 11 presents a simplified enterprise model covering the integration part. From a logical perspective, this can be structured into the following layers:
- Cyber-Physical layer - represents the connection with the surrounding physical environment;
- Service layer - contains computing systems used to orchestrate and manage enterprise resources;
- Cognitive layer - leverages AI services through the enterprise.
Fig.11: Enterprise integration
A. Cognitive Layer
The cognitive layer provides a set of microservices, connectors, adapters, and REST services, granting access to the Openfabric ecosystem. From an infrastructure viewpoint, these services may run on the company infrastructure, or they may be offered by external providers.
B. Service Layer
The service layer contains enterprise domain-specific business rules, data models, and workflows. Preparing the company’s unstructured data for AI ingestion requires passing them through a normalization / cleanup process. This process makes use of ontology-based data abstractions [62]. From a user perspective through the use of workflows (sequences, flowcharts, and transactional business processes), complex business problems could be formally described and automatized with ease [90], [91]. Workflows operating in an unattended (without human supervision) or assisted (with human guidance) context can be enhanced with AI capabilities that are provided by the cognitive layer. The workflow activities, tasks, and decision blocks could be connected to Openfabric AIs through the use of microservice connectors.
C. Cyber-Physical layer
The Cyber-Physical layer represents the connection with the surrounding physical environment. It supervises digital twin instances of IoT devices, robots, industrial machinery, processes, and humans [92], [93].