SEMIoTICS will improve on the state of the art in several areas. Major contributions are listed below:
1) Security, Privacy, Dependability and Interoperability by design: SEMIoTICS adopts a pattern-based approach to guarantee SPDI properties across horizontal and vertical compositional structures of IoT applications.
2) Software defined networking (SDN) and network function virtualization (NFV): SEMIoTICS will develop algorithms for discovering, aggregating and dynamically (re-) assigning physical resources to overlay virtual entities considering the context of the data and services that request access to those entities. SEMIoTICS will develop an SDN architecture for efficient interconnectivity of smart objects. SDN Orchestration will interconnect networks with different QoS levels between different IoT domains to provide End-to-End service connectivity between them and remote computational infrastructures (e.g. cloud, fog), meeting different IoT application requirements related to bandwidth, latency and energy efficiency, using semantic information. SEMIoTICS will also first introduce the deployment of Virtual Network Functions (VNFs) to remove complexity from Local and Access sensor networks. SEMIoTICS will provide the NBI API for adaptable and dynamic networking services to client IoT applications.
3) Monitoring & Adaptation mechanisms: SEMIoTICS will focus on addressing these gaps by developing a dynamically configurable and adaptable infrastructure that will be able to accommodate the different monitors and event captors that may be associated with the different deployment environments, smart objects and components of IoT applications and will allow a scalable cross-layer IoT application monitoring by fusion of the different types of monitoring results generated by monitors of different level.
4) IIoT/IoT platforms: SEMIoTICS will develop a solution integrating SEMIoTICS technology, namely filed-level middleware and networking toolbox and backend API and existing IoT platforms. This solution will be made up of building blocks using an efficient and effective subsystem federation mechanism.
5) Machine-learning algorithms: SEMIoTICS will investigate the engineering and adaptation of existing methods for causal modelling, causal discovery, and causal inference in IoT large-scale applications for predictive modelling, anomaly detection, and diagnosis.
6) End-to-end security and Privacy in IoT: SEMIoTICS will combine novel and standardized technologies to provide lightweight and usable mechanisms for the end-to-end authentication of its entities (devices, applications, users etc.). SEMIoTICS by design will include interfaces to analyse the code that IoT apps are executing, and dedicated analysis nodes for monitoring on a secure runtime environment. This will complement the integrated end-to-end security mechanisms. Finally, via the combination of the above advancements with the guarantees and visibility of the SPDI properties’ status, and consequent enhanced user control and accountability, SEMIoTICS will provide a significantly higher level of security and privacy in IIoT/IoT deployments than what is currently available.
7) Semantic Interoperability: SEMIoTICS will extend the W3C approach so that it considers the needs of the IIoT domain, in particular so that it enables the semi-autonomic bootstrapping & interfacing of things. Another key innovation will be the use of semantic information, such as metadata, sensor/actuator information, sender details, network information, as well as the application requirements, for describing the network resources and the introduction of semantic interoperability information and mechanisms, accessible through the SPDI patterns, to achieve end-to-end semantic interoperability in an IoT system.
The potential impacts of the above expected progress will be evaluated during the 2nd half of the project.
