Nowadays, a plethora of IoT platforms exist for management of IoT devices and groups of nodes. Each platform tackles part of the challenges related to interoperability, efficient data management, support of self-* functionalities, provision of generic IoT enablers and management of IoT applications deployment. In many cases, middleware solutions are required for supporting unified management of swarms of IoT nodes. In NEPHELE, through the release of a lightweight software stack, we aim to make available a set of software libraries and tools, able to manage the integration of IoT and Edge Computing mechanisms and coordinate the execution of IoT functions at both the physical (device) and virtual (edge computing infrastructure) level. The development and release of an open-source software stack (VOStack) that supports virtualization of IoT devices and functions with a twofold perspective (convergence of IoT technologies and unified management of IoT functions by edge/cloud computing orchestration platforms) is considered one of the main novelties introduced in NEPHELE.
In parallel, a plethora of orchestration platforms is available for cloud and edge computing applications, each one targeting a set of application needs. However, the highly distributed nature of such applications creates a need for adoption of synergetic orchestration schemes with dynamic and modular characteristics, where responsibilities for orchestration of parts of the application can be assigned on demand to different platforms. NEPHELE aims to provide an integrated environment for the next-generation hyper-distributed applications management, where IoT and edge computing platforms and orchestration mechanisms will interoperate in a secure and trusted way. A meta-orchestrator undertakes the role of efficiently coordinating the management of distributed compute and network resources, and the enforcement of AI-assisted orchestration mechanisms in the various parts of the compute continuum. Intelligence is continuously injected within the orchestration actions, exploiting advances provided by AI technologies in features detection and inference and leading to the optimal management of the interplay among edge and cloud resources.