PROGRAMMING ABSTRACTIONS AND EXECUTION MODELS FOR THE EDGE-TO-CLOUD CONTINUUM
The evolving edge-to-cloud computing continuum combines devices ranging from low-power sensors and IoT endpoints to powerful edge servers and cloud data centers. Dealing with widely distributed infrastructures requires acknowledging and embracing the heterogeneity of the landscape. Current approaches often rely on centralized, provider-specific frameworks that can hinder scalability and adaptability. Rather than treating the continuum as a unified entity, existing solutions tend to create containerized silos, with each tier relying on distinct and non-portable strategies. We investigate methods to enable flexible and efficient software development across the full edge-to-cloud continuum. Our goal is to define programming abstractions and execution models that span a diverse set of devices without being tied to a single vendor’s platform while fully leveraging the specific characteristics of each tier. We focus on high-level programming models and execution abstractions that expose fundamental system properties, such as data locality, resource availability, and execution constraints while decoupling application logic from underlying infrastructure details. These models and abstractions will be implemented and empirically validated in realistic scenarios in terms of usability, scalability, and performance.
Back to Current Students