Data Overlay Mesh in Distributed Clouds Allowing Collaborative Applications

In modern architecture, multi-tenancy is the design mode of operation used to reduce the resource costs of system implementation. This architectural decision does not come without additional costs when managing isolation and data privacy between the tenants. That is a known challenge in cloud computing that translates to systems closer to the users, namely edge computing and micro clouds. Existing data management solutions, while supporting multi-tenancy, need help to handle the dynamic nature of such environments, limiting their ability to scale and support collaborative applications across different clusters or regions. This paper addresses the limitations of current data management solutions in supporting dynamic organization, configuration, and collaboration in multi-tenant, geo-distributed micro clouds at the edge. By extending the existing virtual clouds model, we introduce a new dataspaces layer and protocols for data discovery using the concept of hierarchical namespaces and transforming non-shareable resources, such as files and folders, into collaborative resources. We propose two types of collaborative applications, leveraging stored procedures and event triggers for data consumption. Finally, we ensure that the model provides accurate access and redistribution of shareable (CPU, RAM, disk) and non-shareable resources through graph transformation theory, with correctness validated using multiparty session types. That enables more flexible programming for collaborative applications, improving connectivity across distributed storage systems and allowing resource borrowing between applications within the same organization. This research seeks to bridge the gap between current tools and the demands of modern distributed systems, enabling more robust, adaptable, and collaborative data management solutions.