Proposal of an edge serverless computing platform with an asynchronous communication model

• Item type: Symposium(1)
• 公開日: 2023-12-20
• タイトル: Proposal of an edge serverless computing platform with an asynchronous communication model
• タイトル / 言語: en
• タイトル: Proposal of an edge serverless computing platform with an asynchronous communication model
• 言語: eng
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_5794
• 資源タイプ: conference paper
• 著者所属: Shibaura Institute of Technology
• 著者所属: Shibaura Institute of Technology
• 著者所属(英): en
• 著者所属(英): Shibaura Institute of Technology
• 著者所属(英): en
• 著者所属(英): Shibaura Institute of Technology
• 著者名: Yanzhi, Li; Midori, Sugaya
• 著者名(英): Yanzhi, Li; Midori, Sugaya
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: In recent years, the number of IoT devices that access to the networks is growing rapidly. The outrageous number of network connections and large amount of data transmission generated by these devices bring new challenges to cloud computing. Edge computing has received much attention as one of the solutions to the data transmission bottleneck in cloud computing networks. Currently, edge computing, represented by content delivery networks, has achieved great success in static data distribution and has started to provide computing services in edge computing nodes in a serverless computing paradigm. However, the synchronous computing model used in existing Edge serverless computing platforms has the problem of difficulty in running time-consuming computational procedures such as AI inference. To address this problem, we propose an edge serverless computing platform based on an asynchronous computing model. This study decouples the computational process from the communication process through asynchronous computing, which increases the flexibility of computational resource scheduling, and in turn permits more efficient use of edge computing resources. Also, the serverless computing model and the global computing library introduced in this study can be used to hide hardware differences, which allows different hardware to be deployed in the edge site to improve the computing performance. Finally, we discuss the approach to deploying computational programs on this platform, which validates the possibility of using asynchronous computing techniques to deploy computational programs on this platform.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: In recent years, the number of IoT devices that access to the networks is growing rapidly. The outrageous number of network connections and large amount of data transmission generated by these devices bring new challenges to cloud computing. Edge computing has received much attention as one of the solutions to the data transmission bottleneck in cloud computing networks. Currently, edge computing, represented by content delivery networks, has achieved great success in static data distribution and has started to provide computing services in edge computing nodes in a serverless computing paradigm. However, the synchronous computing model used in existing Edge serverless computing platforms has the problem of difficulty in running time-consuming computational procedures such as AI inference. To address this problem, we propose an edge serverless computing platform based on an asynchronous computing model. This study decouples the computational process from the communication process through asynchronous computing, which increases the flexibility of computational resource scheduling, and in turn permits more efficient use of edge computing resources. Also, the serverless computing model and the global computing library introduced in this study can be used to hide hardware differences, which allows different hardware to be deployed in the edge site to improve the computing performance. Finally, we discuss the approach to deploying computational programs on this platform, which validates the possibility of using asynchronous computing techniques to deploy computational programs on this platform.
• 書誌情報: Proceedings of Asia Pacific Conference on Robot IoT System Development and Platform; 巻 2023, p. 75-76, 発行日 2023-12-20
• 出版者 / 言語: ja
• 出版者: 情報処理学会