{"title":"Edge Federation: A Dependency-Aware Multi-Task Dispatching and Co-location in Federated Edge Container-Instances","authors":"Uchechukwu Awada, Jiankang Zhang","doi":"10.1109/EDGE50951.2020.00021","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00021","url":null,"abstract":"Existing research on Edge Computing has proposed several edge deployment types, such as unmanned aerial vehicles (UAV)-enabled edge computing, telecommunication base stations endowed with edge clusters, fog nodes, cloudlets, etc. However, none of them consider the ability of keeping edge resources running across various edge deployments in a single pool, such that these resources can be holistically managed and controlled from a single federated plane, also eliminate vendor lock-in situations. Moreover, these schemes assume each server can only execute one task or job at any time. They do not consider lightweight tools that can wrap, isolate and co-locate multiple tasks in a virtualized server. As such, they can easily result in fragmentation and over-allocation of resources. Furthermore, as modern applications are becoming more complex, relatively few research considers the different characteristics of tasks, like the class, constraints and dependencies. This research aims to address these limitations, extend the state-of-the-art by providing an intelligent resource scheduling and optimization solutions for high performance in a federated edge computing system, considering both task dependencies and heterogeneous resource demands at the same time. Extensive simulations on real-world data-trace from the recent Alibaba cluster trace, with information on task dependencies and resource demands, show the effectiveness, faster executions, and resource efficiency of our approach.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121558751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. Ashouri, F. Lorig, P. Davidsson, Romina Spalazzese, Sergej Svorobej
{"title":"Analyzing Distributed Deep Neural Network Deployment on Edge and Cloud Nodes in IoT Systems","authors":"M. Ashouri, F. Lorig, P. Davidsson, Romina Spalazzese, Sergej Svorobej","doi":"10.1109/EDGE50951.2020.00017","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00017","url":null,"abstract":"For the efficient execution of Deep Neural Networks (DNN) in the Internet of Things, computation tasks can be distributed and deployed on edge nodes. In contrast to deploying all computation to the cloud, the use of Distributed DNN (DDNN) often results in a reduced amount of data that is sent through the network and thus might increase the overall performance of the system. However, finding an appropriate deployment scenario is often a complex task and requires considering several criteria. In this paper, we introduce a multicriteria decision-making method based on the Analytical Hierarchy Process for the comparison and selection of deployment alternatives. We use the RECAP simulation framework to model and simulate DDNN deployments on different scales to provide a comprehensive assessment of deployments to system designers. In a case study, we apply the method to a smart city scenario where different distributions and deployments of a DNN are analyzed and compared.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"504 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132479313","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Service Continuity Management Method for MEC-Assisted C-V2X Applications","authors":"Liping Hao, Wei Liu, Ye Wang, Yuanyuan Bao, Feng Li, Wai Chen","doi":"10.1109/EDGE50951.2020.00013","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00013","url":null,"abstract":"Vehicle's high mobility and the limited coverage of edge servers may result in interruption of ongoing edge services. Therefore, how to ensure the high reliability and availability of services needs to be solved. In this paper, we propose a service continuity management method that considers vehicle behavior, energy consumption, and application redirection in the fifth generation (5G) mobile networks. Firstly, we predict vehicle destination by adopting extreme gradient boosting (XGBoost) to decide where services are migrated following vehicle mobility. Secondly, we propose a MEC handover strategy based on markov decision process to reduce the overall energy consumption. Thirdly, we illustrate application redirection process in application domain based on 5G core network. To evaluate our method, we conducted experiments using open source dataset and simulation data. Our experimental results show that the proposed service continuity management method can achieve a destination prediction accuracy of 87.77%, and can significantly reduce overall energy consumption by 38.7% as compared to frequent mlaration strategy.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"379 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121757375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yanjin Fu, Daxin Tian, Xunting Duan, Jianshan Zhou, P. Lang, Chunmian Lin, Xin You
{"title":"A Camera–Radar Fusion Method Based on Edge Computing","authors":"Yanjin Fu, Daxin Tian, Xunting Duan, Jianshan Zhou, P. Lang, Chunmian Lin, Xin You","doi":"10.1109/EDGE50951.2020.00009","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00009","url":null,"abstract":"Multi-access edge computing provides a low-latency and high-performance network environment for the Internet of Vehicle services by migrating computing and storage resources to the edge of networks and supports the deployment of more regional-specific vehicle-to-everything services. In this paper, we propose a fusion perception method for roadside cameras and millimeter wave radar as an application of edge computing. First, we used YOLOv3 to process camera data and DBSCAN clustering method to process radar data to obtain the position, speed, and category of detections. Next, we used joint calibration to spatially synchronize camera and radar data and a direct update method to temporally synchronize camera and radar data. Furthermore, we used the Munkres algorithm to associate camera and radar detections, and Kalman filter to track the proposed fusion method's perception results. Finally, we conducted simulation experiments to evaluate the proposed method. The simulation results demonstrate the effectiveness of our algorithm.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"119 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131998658","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yongjing Li, Dan Wang, Tao Sun, Xiaodong Duan, Lu Lu
{"title":"Solutions for Variant Manufacturing Factory Scenarios Based on 5G Edge Features","authors":"Yongjing Li, Dan Wang, Tao Sun, Xiaodong Duan, Lu Lu","doi":"10.1109/EDGE50951.2020.00016","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00016","url":null,"abstract":"Manufacturing factory is an important scenario of vertical industries served by 5G network and is generally categorized as outside and inside the factory building. In this paper, we first analyzed the network requirements of two scenarios, and introduced key 5G network technologies that can fit in. Then we proposed an Out-Factory 5G edge network architecture based on Edge Computing and 5G LAN technologies, an In-Factory 5G edge network architecture with high reliability, low latency and determinacy based on Edge Computing, URLLC and TSN technologies and an integrated solution for factory manufacturing to meet the requirements of both scenarios. Finally, we introduced the practical case of remote control of gantry crane based on 5G Edge Computing in reality. We found that the end-to-end control delay of the gantry crane was effectively reduced, the production efficiency was improved, and the safety of the manufacturing was also improved with less labor cost.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127054387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Tongzheng Sun, J. Xiong, Yang Wang, Tianhui Meng, Xi Chen, Chengzhong Xu
{"title":"RS-pCloud: A Peer-to-Peer Based Edge-Cloud System for Fast Remote Sensing Image Processing","authors":"Tongzheng Sun, J. Xiong, Yang Wang, Tianhui Meng, Xi Chen, Chengzhong Xu","doi":"10.1109/EDGE50951.2020.00010","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00010","url":null,"abstract":"Modern remote sensing (RS) image application systems often distribute image processing tasks among multiple data centers and then gather the processed images from each center to efficiently synthesize the final product. In this paper, we exploit the edge-cloud architecture to design and implement a novel RS image service system, called RS-pCloud, which leverages the Peer-to-Peer (P2P) model to integrate multiple data centers and their associated edge networks. The data center as cloud platform is responsible for the storage and processing of original RS images, as well as the storage of partial processed images while the edge network is mainly for caching and sharing the processed images. With this design, RS-pCloud not only achieves the load sharing of processing works but also attains the data efficiency among the edges at the same time, which in turn improves the performance of the image processing and reduce the cost of the data transmission as well. RS-pCloud is designed to be used in a transparent way where it receives a query task from the user through a certain cloud platform, split the task into different sub-tasks, according to the location of the data they required, and then distribute the sub-tasks to corresponding clouds for near-data processing, the returned results from each cloud are first cached in specific edge for further sharing and then gathered at the client to synthesize the final product. We implemented and deployed RS-pCloud on three clusters in conjunction with an edge network to show its performance advantages over traditional single-cluster systems.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129663514","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
H. Badri, Tayebeh Bahreini, Daniel Grosu, Kai Yang
{"title":"Risk-Aware Application Placement in Mobile Edge Computing Systems: A Learning-based Optimization Approach","authors":"H. Badri, Tayebeh Bahreini, Daniel Grosu, Kai Yang","doi":"10.1109/EDGE50951.2020.00020","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00020","url":null,"abstract":"In this paper, we address the problem of application placement in MEC systems that takes into account the risk of exceeding the energy budget of the edge servers. We formulate the problem as a chance-constrained program, where the objective is to maximize the total quality of service in the system, while keeping the expected risk of exceeding the edge servers' energy budget within an acceptable threshold. We develop a learning-based method to solve the problem which requires a very small execution time for large size instances. We evaluate the performance of the proposed method by conducting an extensive experimental analysis.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115513513","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Leveraging 5G TSN in V2X Communication for Cloud Vehicle","authors":"Dan Wang, Tao Sun","doi":"10.1109/EDGE50951.2020.00023","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00023","url":null,"abstract":"With the development of 5G network, the research on 5G supporting autonomous driving is being carried out rapidly in the industry. Autonomous driving is a quite complex system, and Cloud Vehicle is high-level of Autonomous driving. The deterministic 5G network can be regarded as the first step towards Autonomous driving. Considering 5G is capable of providing high bandwidth, low latency, and deterministic transmission in the near future (3GPP R16), 5G supporting V2X is gaining industry recognition and the corresponding research is prevailing. In the paper, based on the analysis of typical V2X scenario, the requirements for 5G V2X of deterministic transmission and low latency are identified. Then a solution for supporting cloud vehicle is proposed, which is based on the feasibility analysis of utilizing 5G TSN for transmitting V2X control signals deterministically, and the main cases include communication for V2V, V2N and vehicle mobility.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124030138","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Joint Optimization of Task Offloading and Resource Allocation via Deep Reinforcement Learning for Augmented Reality in Mobile Edge Network","authors":"Xing Chen, Guizhong Liu","doi":"10.1109/EDGE50951.2020.00019","DOIUrl":"https://doi.org/10.1109/EDGE50951.2020.00019","url":null,"abstract":"Mobile edge computing (MEC) has been recognized as emerging techniques in 5G to provide powerful computing capabilities for the Ultra Reliable Low Latency Communication (URLLC) applications. In this paper, a MEC enable multi-user wireless network is considered by offloading the computation task to MEC server, reducing latency and energy consumption of user terminal for Augmented Reality (AR) application. The joint optimization problem of resource allocation and task offloading is studied to minimize the energy consumption of each user subject to the delay requirement and the limited resources. We propose a deep reinforcement learning algorithm based on a multi-agent deep deterministic policy gradient (MADDPG) to solve this problem. Simulation results show that the proposed algorithm can greatly reduce energy consumption of the users.","PeriodicalId":418280,"journal":{"name":"2020 IEEE International Conference on Edge Computing (EDGE)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124444166","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}