Decentralised technologies for orchestrated cloud-to-edge intelligence
The emerging spring of Artificial Intelligence (AI) will enable innovative applications exploiting the myriad of connected sensors and appliances embedded in every corner of modern life. Currently, AI requires high computational resources only available in high-performance data centers; therefore, realizing an architecture capable of securely processing this unprecedented amount of remotely sensed and potentially sensitive data, as well as conveying timely responses to pervasive configurable actuators is a non-trivial endeavour, requiring the cooperation of multiple parties.To address these challenges, DECENTER aims to realise a robust Fog Computing platform, covering the whole Cloud-to-Things Continuum, that will provide AI application-aware orchestration and provisioning of resources. The project will enrich existing Cloud and IoT solutions with advanced capabilities to abstract features and process data closer to where it is produced. DECENTER will enable a collaborative environment in which multiple stakeholders (Cloud and IoT providers) can securely share and harmoniously manage resources, in dynamically created multi-cloud/edge, federated environments. Cross-border infrastructure federation will be realized via Blockchain-based Smart Contracts defining customized Service Level Agreements, used to commit the execution of verified workloads across multiple, potentially remote, administrative domains. Through such novelties, DECENTER will unlock the potential of innovative decentralised AI algorithms and models, by deploying them across multiple tiers of the infrastructure and federated clouds. The project will follow a lean implementation methodology and validate its concept with real-world pilots executed in urban, industrial and home environments. With its approach, DECENTER will target the emergence of innovative digital businesses, thus providing a competitive advantage to EU and Korean industry and fostering cross-border collaboration.