Research in edge and fog computing aims to offer a variety of mechanisms for the acquisition and use of computing resources closer to the data source and data consumption. The availability of edge systems provides applications and developers with new means to change the way these applications interact with the computational infrastructure, reducing delay and total network traffic and improving response times and user QoE [Zhao et al., 2019]. The distributed edge computing infrastructure can be used in conjunction with the central cloud to support a variety of services for applications in future smart cities [Bittencourt et al., 2017], for example to improve traffic as well as vehicles security and passengers comfort (e.g. driving assistance, traffic management, surveillance cameras for traffic violations and vehicle tracking, gaming, augmented reality). The large amount of data collected in real-time from cars can be processed by the intelligent edge to produce timely responses for automated traffic management. Another application that can benefit from the intelligent edge is telemedicine and remote healthcare. Data demands at the edge increase with telemedicine requirements, with for example potentially large image data sets in digital pathology are needed for proper diagnosis and treatment prescription, and can also have strict networking requirements, where reliable, low-latency communications can be necessary for future personalized medicine with automatic dosage adjustments according to real-time patient monitoring. The Tactile Internet will allow more precise remote diagnosis and treatment, where data transfer and processing in real time are of paramount importance.
A fundamental question in edge computing is: where these data should be sent to be processed? The answer to this question is very much related to the answer to: where and when these data will be needed? This is application-dependent and triggers many complex combinations of data distribution and processing throughout an edge-cloud hierarchy. Therefore, management solutions to data transfer and processing at the edge are needed to improve application response time and improve network efficiency by avoiding unnecessary data movements. An intelligent edge can exist in both healthcare facilities and patients’ homes, composing a distributed edge intelligence system that, with continuous learning, can give support to better diagnosis, treatment, as well as understanding medicine collateral effects within a variety of scenarios. Medical e-education and e-training can take advantage of such advanced infrastructures where intelligent edge cloud networks are capable of managing orchestration of services, data transfer and storage, often with privacy requirements.
This research strand will investigate mechanisms to manage the edge computing infrastructure, namely edge / fog computing to deliver an intelligent computing layer acting as a swarm between data producers/consumers and the cloud. This involves adaptive decision-making mechanisms to process data at edge computing resources, but also auto-configure the networking and computing infrastructures to cooperate and achieve the QoS/QoE for widely heterogeneous applications. Ultimately, an autonomic, intelligent swarm should be able to tune data collection needs, e.g. increasing or decreasing sensing frequency, and determine how edge nodes data should be combined and distributed at the edge to be able to: (i) fulfill applications’ SLA; and (ii) run intelligence-based algorithms with the same quality and efficiency as in data farms.