Search

Computational requirements for data processing at different stages of the radiology value chain are increasing. Cone beam computed tomography (CBCT) is a diagnostic imaging technique used in dental and extremity imaging, involving a highly demanding image reconstruction task. In turn, artificial intelligence (AI) assisted diagnostics are becoming increasingly popular, thus increasing the use of computation resources. Furthermore, the need for fully independent imaging units outside radiology departments and with remotely performed diagnostics emphasize the need for wireless connectivity between the imaging unit and hospital infrastructure. In this feasibility study, we propose an approach based on a distributed edge-cloud computing platform, consisting of small-scale local edge nodes, edge servers with traditional cloud resources to perform data processing tasks in radiology. We are interested in the use of local computing resources with Graphics Processing Units (GPUs), in our case Jetson Xavier NX, for hosting the algorithms for two use-cases, namely image reconstruction in cone beam computed tomography and AI-assisted cancer detection from mammographic images. Particularly, we wanted to determine the technical requirements for local edge computing platform for these two tasks and whether CBCT image reconstruction and breast cancer detection tasks are possible in a diagnostically acceptable time frame. We validated the use-cases and the proposed edge computing platform in two stages. First, the algorithms were validated use-case-wise by comparing the computing performance of the edge nodes against a reference setup (regular workstation). Second, we performed qualitative evaluation on the edge computing platform by running the algorithms as nanoservices. Our results, obtained through real-life prototyping, indicate that it is possible and technically feasible to run both reconstruction and AI-assisted image analysis functions in a diagnostically acceptable

According to World Health Organization (WHO), the worldwide prevalence of chronic diseases increases fast and new threats, such as Covid-19 pandemic, continue to emerge, while the aging population continues decaying the dependency ratio. These challenges will cause a huge pressure on the efficacy and cost-efficiency of healthcare systems worldwide. Thanks to the emerging technologies, such as novel medical imaging and monitoring instrumentation, and Internet of Medical Things (IoMT), more accurate and versatile patient data than ever is available for medical use. To transform the technology advancements into better outcome and improved efficiency of healthcare, seamless interoperation of the underlying key technologies needs to be ensured. Novel IoT and communication technologies, edge computing and virtualization have a major role in this transformation. In this article, we explore the combined use of these technologies for managing complex tasks of connecting patients, personnel, hospital systems, electronic health records and medical instrumentation. We summarize our joint effort of four recent scientific articles that together demonstrate the potential of the edge-cloud continuum as the base approach for providing efficient and secure distributed e-health and e-welfare services. Finally, we provide an outlook for future research needs.

Effective, long-lasting Industrial IoT (IIoT) solutions start with short-term gains and progressively mature with added capabilities and value. The heterogeneous nature of IIoT devices and services suggests frequent changes in resource requirements for different services, applications, and use cases. With such unpredictability, resource orchestration can be quite complicated even in basic use cases and almost impossible to handle in some extensively dynamic use cases. In this paper, we propose SDRM; an SDN-enabled Resource Management scheme. This novel orchestration methodology automatically computes the optimal resource allocation for different IIoT network models and dynamically adjust assigned resources based on predefined constraints to ensure Service Level Agreement (SLA). The proposed approach models resource allocation as a Constraint Satisfaction Problem (CSP) where optimality is based on the solution of a predefined Satisfiability (SAT) problem. This model supports centralized management of all resources using a software defined approach. Such resources include memory, power, bandwidth, and edge-cloud resources. SDRM aims at accelerating efficient resource orchestration through dynamic workload balancing and edge-cloud resource utilization, thereby reducing the cost of IIoT system deployment and improving the overall ROI for adopting IIoT solutions. We model our resource allocation approach on SAVILE ROW using ESSENSE PRIME modeling language, we then implement the network model on CloudSimSDN and PureEdgeSim. We present a detailed analysis of the system architecture and the key technologies of the model. We finally demonstrate the efficiency of the model by presenting experimental results from a prototype system. Our test results show an extremely low solver time ranging from 0.47 ms to 0.5 ms for nodes ranging from 100 to 500 nodes. With edge-cloud collaboration, our results show about 4 percent improvement in overall task success rates.

Edge computing is a novel computing paradigm moving server resources closer to end-devices. It helps unleashing the full potential of high-performance access networks with respect to ultra-low latency and transfer rate and improve resilience to problems at core networks and data centers. Multi-access Edge Computing (MEC), is a standard solution by European Telecommunications Standards Institute (ETSI) for access network-level edge computing. MEC, operating at access network level, is an ideal solution for the most cases. However, there are still some challenges to address: first is related to the vulnerability to access network problems and the second is about the high load inflicted to access networks and MEC servers. This is a particular issue in massive-scale Internet of Things (IoT) use cases, where numerous sensors may produce high amounts of data, or where critical system functionalities must be ensured also during access network problems. In this paper, we study the feasibility of bringing some edge functions to the local level as virtualized and dynamically deployable components utilizing local hardware capacity. For the study, we have implemented a local edge networking prototype based on local microservices, called nanoservices, implemented using Docker containers and deployed using Docker Swarm-based orchestration. Since IoT networks typically consist of constrained-capacity devices, our focus is in optimizing the resources of the proposed nanoservices.

In the envisioned ubiquitous world, services will follow users as they move across smart surroundings. Services are instantiated to users through the environment, appearing and disappearing as they move, which reduces the need for personal communication devices such as smartphones or tablets. To facilitate this development, service architectures need to support virtualized, on-demand service composition based on the hardware and software resources available at the current user location. The technical context for this type of user interaction with digital services through smart surroundings is called Internet of Everything (IoE). Today’s service architectures will be too inflexible in this highly decentralized and dynamic environment. Hence, in this article we propose a novel service model called nanoEdge, where nodes collaboratively provide needed functions for virtual services that need to be deployed locally due to performance, efficiency or reliability requirements, for example. The main contributions of this article are the nanoEdge conceptual model and its proof-of-concept (PoC) implementation to show that the model is feasible with regard to performance and resource-efficiency. The successful demonstration of PoC implementation exemplifies future IoE service scenarios with today’s hardware components.