XMANAI Reference Architecture: Perspectives

The XMANAI reference architecture aims at providing the basis for the detailed specification, development and integration activities of the overall XMANAI platform along its Data and AI related Services Bundles, its XAI Algoritmms and Models Catalogue, and the different Manufacturing Apps. As any reference architecture, it builds on concrete business requirements, technical requirements and the preliminary MVP features and takes into consideration the best practices provided by different manufacturing initiatives (such as RAMI 4.0 and IIRA). To this direction, the XMANAI architecture design involved the high-level design of the XMANAI architecture across different perspectives, namely:

  • The tiers perspective in order to anticipate where and how the XMANAI Platform is delivered and deployed,
  • The services bundle perpective that abstracts how the data and XAI-related services are developed, including eight (8) services bundles, namely the Data Collection & Governance Services, the Scalable Storage Services, the Data Manipulation Services, the XAI Model Lifecycle Management Services, the XAI Execution Services, the XAI Insights Services, the Secure Asset Sharing Services and the Management Services Bundle.
  • The component perspective that outlines the main functionality of the different XMANAI components and how they are brought together in the different tiers and services bundles, and
  • The application perspective that highlights the scope of the manufacturing apps developed per demonstrator, along with their expected interactions with the XMANAI Platform.
XMANAI Reference Architecture from a Tiers Perspective
XMANAI Reference Architecture from a Tiers Perspective

Different workflows that the XMANAI Platform will support have been designed to specify at high-level the expected interactions among the components, and provide indicative specifications about how the XMANAI Platform can be used to implement the different user journeys. Such workflows include:

  • The XAI Preparation Workflow that involves business users, data scientists and data engineers in the necessary activities for collecting, ingesting and understanding the data.
  • The XAI Experimentation Workflow that involves business users, data scientists and data engineers in the collaborative preparation of AI pipelines for the problem at hand.
  • The XAI Insights Workflow that involves business users, and data scientists and encapsulates the extraction of appropriate insights (from Explainable AI pipelines and models) in an understandable, explicit manner.
  • The XAI Application Workflow that involves data scientists and, indirectly, the XMANAI Manufacturing apps in order to retrieve the results (in terms of predictions/decisions and their associated explanations) of XAI pipelines that have been configured to solve the related manufacturing problems.


In summary, the XMANAI Platform shall act as: (I) a manufacturing data space that can be leveraged by any manufacturing app to acquire manufacturing data in a secure and reliable manner; and (II) an XAI intelligence space that allows for collaborative experimentation and putting to production appropriate XAI pipelines to solve specific manufacturing problems.