엣지 디바이스에서의 병렬 프로그래밍 모델 성능 비교 연구

남덕윤,Dukyun Nam 대한임베디드공학회 2023 대한임베디드공학회논문지 Vol.18 No.4

Heterogeneous computing is a technology that utilizes different types of processors to perform parallel processing. It maximizes task processing and energy efficiency by leveraging various computing resources such as CPUs, GPUs, and FPGAs. On the other hand, edge computing has developed with IoT and 5G technologies. It is a distributed computing that utilizes computing resources close to clients, thereby offloading the central server. It has evolved to intelligent edge computing combined with artificial intelligence. Intelligent edge computing enables total data processing, such as context awareness, prediction, control, and simple processing for the data collected on the edge. If heterogeneous computing can be successfully applied in the edge, it is expected to maximize job processing efficiency while minimizing dependence on the central server. In this paper, experiments were conducted to verify the feasibility of various parallel programming models on high-end and low-end edge devices by using benchmark applications. We analyzed the performance of five parallel programming models on the Raspberry Pi 4 and Jetson Orin Nano as low-end and high-end devices, respectively. In the experiment, OpenACC showed the best performance on the low-end edge device and OpenSYCL on the high-end device due to the stability and optimization of system libraries.

IoT 기반 스마트 공장 구현을 위한 엣지 클라우드 플랫폼 개발

김형선(Hyung-Sun Kim),이홍철(Hong-Chul Lee) 한국컴퓨터정보학회 2019 韓國컴퓨터情報學會論文誌 Vol.24 No.5

In this paper, we propose an edge cloud platform architecture for implementing smart factory. The edge cloud platform is one of edge computing architecture which is mainly focusing on the efficient computing between IoT devices and central cloud. So far, edge computing has put emphasis on reducing latency, bandwidth and computing cost in areas like smart homes and self-driving cars. On the other hand, in this paper, we suggest not only common functional architecture of edge system but also light weight cloud based architecture to apply to the specialized requirements of smart factory. Cloud based edge architecture has many advantages in terms of scalability and reliability of resources and operation of various independent edge functions compare to typical edge system architecture. To make sure the availability of edge cloud platform in smart factory, we also analyze requirements of smart factory edge. We redefine requirements from a 4M1E(man, machine, material, method, element) perspective which are essentially needed to be digitalized and intelligent for physical operation of smart factory. Based on these requirements, we suggest layered(IoT Gateway, Edge Cloud, Central Cloud) application and data architecture. we also propose edge cloud platform architecture using lightweight container virtualization technology. Finally, we validate its implementation effects with case study. we apply proposed edge cloud architecture to the real manufacturing process and compare to existing equipment engineering system. As a result, we prove that the response performance of the proposed approach was improved by 84 to 92% better than existing method.

An Overview of Mobile Edge Computing: Architecture, Technology and Direction

( Arslan Rasheed ),( Peter Han Joo Chong ),( Ivan Wang-hei Ho ),( Xue Jun Li ),( And William Liu ) 한국인터넷정보학회 2019 KSII Transactions on Internet and Information Syst Vol.13 No.10

Modern applications such as augmented reality, connected vehicles, video streaming and gaming have stringent requirements on latency, bandwidth and computation resources. The explosion in data generation by mobile devices has further exacerbated the situation. Mobile Edge Computing (MEC) is a recent addition to the edge computing paradigm that amalgamates the cloud computing capabilities with cellular communications. The concept of MEC is to relocate the cloud capabilities to the edge of the network for yielding ultra-low latency, high computation, high bandwidth, low burden on the core network, enhanced quality of experience (QoE), and efficient resource utilization. In this paper, we provide a comprehensive overview on different traits of MEC including its use cases, architecture, computation offloading, security, economic aspects, research challenges, and potential future directions.

Edge AI-based Brain-Computer Interface for Real-time Applications

Henar Mike Canilang,Chigozie Uzochukwu Udeogu,James Rigor Camacho,Erick Valverde,Angela Caliwag,Wansu Lim 대한인간공학회 2021 대한인간공학회 학술대회논문집 Vol.2021 No.11

Objective: This study aims to integrate brain computer interface (BCI) to edge AI devices for real-time EEG signal processing applications. For the specific implementation in this paper, we applied edge AI device-based EEG signal processing for emotion recognition. Background: The emergence of Electroencephalogram (EEG) based applications for intelligent applications is projected to have rapid advancements in the future. The BCI system enables efficient brain signal acquisition. Current intelligent convergence of EEG based applications includes brain signal processing integrated to deep learning models. It is expected that this convergence in intelligent EEG based applications will push through to on-device local processing such as edge AI devices for portability in state-of-the-art applications. The portability and practical usage of these systems in real-world applications could lead to the development and deployment of many other advanced embedded systems for EEG-based applications. Systems that can run locally on the edge without needing to be connected to a mobile network. Edge AI devices are the leading-edge computing platforms that process data locally to overcome the current constraints of IoT application. This paves way to the integration of edge-based processing as the computing paradigm to process and acquire EEG signals. Owing to the current research advancement for both EEG and edge applications, this paper aims to propose one of the many systematic applications of deploying edge-based EEG using a brain computer interface. Method: The input for this edge-based EEG signal processing is through the BCI interfaced to the edge AI device. The edge AI device deployed with a deep learning model for specific applications locally processes the acquired signal. These acquired signals are valuable for training deep learning models to realize practical applications at the edge. The processed EEG signals enable the system response of the system such as rapid emotion recognition. Results: Varying EEG signals were acquired in each of the BCI channels. These brain signals are segmented to different brain signal clusters such as Gamma waves (30㎐ to 100㎐), Beta waves (12㎐ – 30㎐), Alpha waves (7.5㎐ – 12㎐), Theta waves (4㎐-7.5㎐) and Delta waves (0.1㎐-4㎐) which have specific brain wave description. As for EEG emotion recognition applications, these wave signals are essential for efficient and accurate emotion recognition. The alpha, beta, and gamma waves are identified to be the most discriminative frequency ranges to identify emotion. Each of the EEG signal is classified for emotion recognition and identification such as 1) valence, 2) dominance, 3) arousal and 4) liking. High and low responses from these wave signals have corresponding positive, neutral, and negative emotions based on their neural patterns at parietal and occipital sites. Other applications can use the acquired EEG signals thus maximizing the possible application of edge-based EEG signal processing. Conclusion: The local processing of the EEG signal at the edge enables the edge-based EEG system application thus enabling system response and actuation. Edge EEG also enables local and cloud co-processing whereas this maximizes the benefits of the edge computing paradigm. With this co-processing capability, it enables an adaptive and portable real-time EEG signal processing which is a constraint to conventional EEG based emotion recognition system. Application: EEG is a physiological based emotion recognition which proves to be more accurate than conventional non-physiological emotion recognition. Also, with an edge-based EEG application, it enables portability and flexibility in terms of its deployment. This application aims to be a state-of-the-art innovation to existing physiological and non-physiological emotion recognition. Furthermore, this research paper implementation aims to emphasize the vast possible applications of edge-based EEG signal processing to bridge

엣지 컴퓨팅에서 딥러닝 기반의 침입 탐지 시스템 설계 및 구현

김종욱,최미정 한국통신학회 2022 韓國通信學會論文誌 Vol.47 No.8

Edge computing is a new distributed computing technology that adds an edge layer between the cloud and device layer. Edge computing offers more targets than cloud computing, and attackers exploit vulnerabilities, DoS/DDoS, man-in-the-middle attacks, and authentication bypasses to threaten them. Security systems such as intrusion detection systems (IDS), firewalls, and anti-virus software are unsuitable for edge computing due to low accuracy and high false positives. It also revealed limitations such as a lack of security personnel and solutions to respond. This paper proposes a deep learning-based IDS to overcome the limitations of edge computing. We implemented a deep learning-based IDS in KubeEdge that a highly scalable edge computing platform and extracted important features using sparsity constraints to train an intrusion detection model. The model deployed in edge computing achieved 98.96% accuracy, 99.41% F1-Score, 2.270% false-positive, and 0.4990% undetected rate. The system reported to the user that an intrusion had occurred and took appropriate actions to black the attackers' IP. 엣지 컴퓨팅은 클라우드 계층과 단말 계층 사이에 엣지 계층을 추가한 새로운 분산 컴퓨팅 기술이다. 엣지 컴퓨팅은 클라우드 컴퓨팅보다 상대적으로 더 많은 표적을 제공하기 때문에 공격자는 취약점 악용, DoS/DDoS, 중간자 공격, 그리고 인증 우회 등을 사용한다. 다양한 위협을 탐지하기 위해 침입 탐지 시스템, 방화벽, 안티 바이러스 소프트웨어와 같은 보안 시스템들이 사용되지만 낮은 정확도, 높은 오탐으로 인해 엣지 컴퓨팅에는 부적합하다. 또한, 침입 탐지를 위해 필요한 전문 인력과 대응할 수 있는 솔루션의 부족과 같은 한계점도 드러났다. 본 논문에서는 엣지 컴퓨팅에서 한계점을 극복하기 위해 딥러닝 기반의 침입 탐지 시스템을 제안한다. 제안하는 딥러닝기반의 침입 탐지 시스템은 엣지 컴퓨팅 플랫폼인 KubeEdge를 사용하여 엣지 컴퓨팅을 구성하였다. 침입 탐지모델을 생성하기 위해 희소성 제약을 사용하여 중요한 특징들을 추출 및 학습했다. 학습된 침입 탐지 모델을 엣지컴퓨팅에 배포 및 운영하였을 때 평균 98.96%의 정확도, 99.41%의 F1-Score, 2.270%의 오탐률, 0.4990%의 미탐률을 달성했으며 사용자에게 침입이 발생했음을 보고하고 공격자 IP를 차단하는 적절한 대응을 수행했다.

Edge-of-things computing framework for cost-effective provisioning of healthcare data

Alam, Md. Golam Rabiul,Munir, Md. Shirajum,Uddin, Md. Zia,Alam, Mohammed Shamsul,Dang, Tri Nguyen,Hong, Choong Seon Elsevier 2019 Journal of parallel and distributed computing Vol.123 No.-

<P><B>Abstract</B></P> <P>Edge-of-Things (EoT)-based healthcare services are forthcoming patient-care amenities related to autonomic and persuasive healthcare, where an EoT broker usually works as a middleman between the Healthcare Service Consumers (HSC) and Computing Service Providers (CSP). The computing service providers are the edge computing service providers (ECSP) and cloud computing service provider (CCSP). Sensor observations from a patient’s body area networks (BAN) and patients’ medical and genetic historical data are very sensitive and have a high degree of interdependency. It follows that EoT based patient monitoring systems or applications are tightly coupled and require obstinate synchronization. Therefore, this paper proposes a portfolio optimization solution for the selection of virtual machines (VMs) of edge and/or cloud computing service providers. The dynamic pricing for an EoT computation service is considered by the EoT broker for optimal VM provisioning in an EoT environment. The proposed portfolio optimization solution is compared with the traditional certainty equivalent approach. As the portfolio optimization is a centralized solution approach, this paper also proposes an alternating direction method of multipliers (ADMM) based distributed provisioning method for the healthcare data in the EoT computing environment. A comparative study shows the cost-effective provisioning for the healthcare data through portfolio optimization and ADMM methods over the traditional certainty equivalent and greedy approach, respectively.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Edge-of-Things (EoT) computation framework for healthcare service provisioning. </LI> <LI> Portfolio optimization approach for cost-effective healthcare data provisioning. </LI> <LI> Alternating direction method of multipliers (ADMM) for healthcare data offloading. </LI> </UL> </P>

A REVIEW ON FOG COMPUTING: ARCHITECTURE, FOG WITH IoT, ALGORITHMS AND RESEARCH CHALLENGES

Sabireen H.,Neelanarayanan V. 한국통신학회 2021 ICT Express Vol.7 No.2

With the increasing advancement in the applications of the Internet of Things (IoT), the integrated Cloud Computing (CC) faces numerous threats such as performance, security, latency, and network breakdown. With the discovery of Fog Computing these issues are addressed by taking CC nearer to the Internet of Things (IoT). The key functionality of the fog is to provide the data generated by the IoT devices near the edge. Processing of the data and data storage is done locally at the fog node rather than moving the information to the cloud server. In comparison with the cloud, Fog Computing delivers services with high quality and quick response time. Hence, Fog Computing might be the optimal option to allow the Internet of Things to deliver an efficient and highly secured service to numerous IoT clients. It allows the administration of the services and resource provisioning outside CC, nearer to devices, at the network edge, or ultimately at places specified by Service Level Agreements (SLA’s). Fog Computing is not a replacement to CC, but a prevailing component. It allows the processing of the information at the edge though still delivering the option to connect with the data center of the cloud. In this paper, we put forward various computing paradigms, features of fog computing, an in-depth reference architecture of fog with its various levels, a detailed analysis of fog with IoT, various fog system algorithms and also systematically examine the challenges in Fog Computing which acts as a middle layer between IoT sensors or devices and data centers of the cloud.

Cluster Balancing Scheme for Home Edge Computing (HEC)

Md. Sajjad Hossain,Fabliha Bushra Islam,Jae Min Lee,Dong-Seong Kim 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.2

To overcome the limitations of cloud computing edge computing facilitates with the low response time and resource management at the edge of the network. But The edge computing scheme appeals a lot more attention as well as projected to fulfill ultra-low reaction time demanded by developing IoT applications. A new dimension of edge comput-ing named Home Ege Computing (HEC) is proposed to fulfill the requirements of some delay sensitive systems like traffic system. This HEC consists of three layers as MEC (Multi-access Edge Computing), HEC servers and cloud server. This article suggests a solution to explain and solve the challenges of latency on HEC servers affected through their inadequate resources. with the increasing number of traffic it crates a very long queue while serving the job. By taking the technique of load balancing and clustering this paper propose a Home Edge Computing Clustering balance technique. This technique permits us to disseminate the applications hierarchically on the HEC clusters as well as an additional emphasis of the planning is to avert bottleneck on a HEC server to lessen the overall latency. We proposed to compare this HEC-Clustering Balance result with some basic clustering approach and load balancing approach.

엣지 컴퓨팅에서 트래픽 분산을 위한 흐름 예측 기반 동적 클러스터링 기법

이창우,Lee, Chang Woo 한국멀티미디어학회 2022 멀티미디어학회논문지 Vol.25 No.8

This paper is a method for efficient traffic prediction in mobile edge computing, where many studies have recently been conducted. For distributed processing in mobile edge computing, tasks offloading from each mobile edge must be processed within the limited computing power of the edge. As a result, in the mobile nodes, it is necessary to efficiently select the surrounding edge server in consideration of performance dynamically. This paper aims to suggest the efficient clustering method by selecting edges in a cloud environment and predicting mobile traffic. Then, our dynamic clustering method is to reduce offloading overload to the edge server when offloading required by mobile terminals affects the performance of the edge server compared with the existing offloading schemes.

딥러닝을 사용하는 IoT빅데이터 인프라에 필요한 DNA 기술을 위한 분산 엣지 컴퓨팅기술 리뷰

( Temesgen Seyoum Alemayehu ),조위덕 ( We-duke Cho ) 한국정보처리학회 2020 정보처리학회논문지. 컴퓨터 및 통신시스템 Vol.9 No.12

오늘날 데이터 네트워크 AI (DNA) 기반 지능형 서비스 및 애플리케이션은 비즈니스의 삶의 질과 생산성을 향상시키는 새로운 차원의 서비스를 제공하는 것이 현실이 되었다. 인공지능(AI)은 IoT 데이터(IoT 장치에서 수집한 데이터)의 가치를 높이며, 사물 인터넷(IoT)은 AI의 학습 및 지능기능을 촉진한다. 딥러닝을 사용하여 대량의 IoT 데이터에서 실시간으로 인사이트를 추출하려면 데이터가 생성되는 IoT 단말 장치에서의 처리능력이 필요하다. 그러나 딥러닝에는 IoT 최종 장치에서 사용할 수 없는 상당 수의 컴퓨팅 리소스가 필요하다. 이러한 문제는 처리를 위해 IoT 최종 장치에서 클라우드 데이터 센터로 대량의 데이터를 전송함으로써 해결되었다. 그러나 IoT 빅 데이터를 클라우드로 전송하면 엄청나게 높은 전송 지연과 주요 관심사인 개인 정보 보호 문제가 발생한다. 분산 컴퓨팅 노드가 IoT 최종 장치 가까이에 배치되는 엣지 컴퓨팅은 높은 계산 및 짧은 지연 시간 요구 사항을 충족하고 사용자의 개인 정보를 보호하는 실행 가능한 솔루션이다. 본 논문에서는 엣지 컴퓨팅 내에서 딥러닝을 활용하여 IoT 최종 장치에서 생성된 IoT 빅 데이터의 잠재력을 발휘하는 현재 상태에 대한 포괄적인 검토를 제공한다. 우리는 이것이 DNA 기반 지능형 서비스 및 애플리케이션 개발에 기여할 것이라고 본다. 엣지 컴퓨팅 플랫폼의 여러 노드에서 딥러닝 모델의 다양한 분산 교육 및 추론 아키텍처를 설명하고 엣지 컴퓨팅 환경과 네트워크 엣지에서 딥러닝이 유용할 수 있는 다양한 애플리케이션 도메인에서 딥러닝의 다양한 개인정보 보호 접근 방식을 제공한다. 마지막으로 엣지 컴퓨팅 내에서 딥러닝을 활용하는 열린 문제와 과제에 대해 설명한다. Nowadays, Data-Network-AI (DNA)-based intelligent services and applications have become a reality to provide a new dimension of services that improve the quality of life and productivity of businesses. Artificial intelligence (AI) can enhance the value of IoT data (data collected by IoT devices). The internet of things (IoT) promotes the learning and intelligence capability of AI. To extract insights from massive volume IoT data in real-time using deep learning, processing capability needs to happen in the IoT end devices where data is generated. However, deep learning requires a significant number of computational resources that may not be available at the IoT end devices. Such problems have been addressed by transporting bulks of data from the IoT end devices to the cloud datacenters for processing. But transferring IoT big data to the cloud incurs prohibitively high transmission delay and privacy issues which are a major concern. Edge computing, where distributed computing nodes are placed close to the IoT end devices, is a viable solution to meet the high computation and low-latency requirements and to preserve the privacy of users. This paper provides a comprehensive review of the current state of leveraging deep learning within edge computing to unleash the potential of IoT big data generated from IoT end devices. We believe that the revision will have a contribution to the development of DNA-based intelligent services and applications. It describes the different distributed training and inference architectures of deep learning models across multiple nodes of the edge computing platform. It also provides the different privacy-preserving approaches of deep learning on the edge computing environment and the various application domains where deep learning on the network edge can be useful. Finally, it discusses open issues and challenges leveraging deep learning within edge computing.