On July 21st, Polimi partners took part at a workshop, held for the EUCloudEdgeIoT project and led by Atos, with the objective to identify the thematic areas and building blocks composing a reference architecture in the field of IoT/Edge and Cloud computing. We started from an initial idea, summarised as a schema for the building blocks and arisen from the results of the previous workshops.
Initial blocks included the following concepts: Security, Trust & Reputation, Data Management, Resource Management, Orchestration, Network, Monitoring & Observability, AI.
The main goal of the discussion was to identify the missing blocks and functionalities in order to update the proposed model, which aims at becoming a standard, promoting a European reference architecture for the continuum.
In order to collaborate and make a joined effort to propose updates for the model, all the partners were divided into two groups.
Here is a brief summary of the Group 1 discussion:
First, it was asked why the concept of network, being it the layer for all the computing resources, was emphasized, when in the context of a project, network is usually identified as a computing resource itself. It was then pointed out that computer network resources should be part of a federation to orchestrate both resources and services, and that on top of the network, there could be software establishing security and separation over the network.
Such a concept of orchestration of resources and services could be ideally consists of a path crossing separate domains: for example, the sharing of monitoring and observability could be served on top of all federations of resources.
The discussion then focused on the role of data within the architecture: the need to express the concept of data sharing and links among blocks was identified, thus leading to the idea of adding new concepts, such as data fabric and data governance, to the architecture, focusing on the need to represent the relationship between the data and the orchestration services.
The need to establish secure communication among different points in the continuum was then strongly reiterated, emphasizing the need to depict a data driven architecture and a common exchange protocol.
Finally, the discussion moved around the concept of orchestration, focusing on the different ways such a concept is perceived in different contexts: for example, in the swarm computing perspective, orchestration is considered to be distributed.
After the discussion, results from each of the two groups were presented.
Group 2 identified the need to include the concept of Privacy (Roles and Authorization) within the Security block, emphasizing the concept of DevSecOps as a functionality belonging to the block, and to include Accountability within the Trust & Reputation block. Then, the group pointed out the need to create a space that spans across the different blocks (Cross-Layer functionalities).
The final proposal was to allow all the members to list all the terms related to the different projects and topics, specifying the perceived definition of such terms for their thematic areas, with the objective to merge and deliver such terms to a common glossary including the different meanings for each term.
To do so, members were allowed to modify a shared document to include the different points of view, thus building a full taxonomy for the continuum.
Moreover, participants were encouraged to participate to the next workshops and to fill a survey form, which will be sent to all the members in September, to further investigate the best approach to collaborate.