標題: Toward UAV-Based Airborne Computing
作者: Lin, Kate Ching-Ju
Wang, Hao-Chen
Lai, Yuan-Cheng
Lin, Ying-Dar
交大名義發表
資訊工程學系
National Chiao Tung University
Department of Computer Science
關鍵字: Delays;Monitoring;Signal to noise ratio;Interference;5G mobile communication;Throughput;Base stations;Edge computing<bold>;</bold>
公開日期: 1-Dec-2019
摘要: The next generation of mobile networks, 5G, aims at supporting lower end-to-end latency, higher reliability and higher throughput, which can be improved by MEC and multi-RAT offloading, respectively. With MEC, a base station in 5G can be equipped with computing power, which can be called as an edge in MEC. With the assistance of the edges, traffic with computational tasks can be directly executed in local servers, without forwarding the tasks to the cloud or core network. Edge computing hence reduces the latency and the traffic load significantly. Conventional multi- RAT offloading decides based only on either communication resources or computing resources. In this work, we argue that, to better utilize the communication and computing resources, neighboring edges should share their resources and cooperatively offload the requests from their clients. To this end, we introduced a double offloading mechanism, called LCCOP, to offload incoming traffic to the best pair of radio and edge subject to the endto- end latency of the requesting connections. We conduct simulations to compare LCCOP with the conventional offloading schemes. The results show that LCCOP's double offloading can significantly improve the satisfaction ratio by up to 83 percent and 143 percent, respectively, as compared to pure computation offloading and pure communication offloading.
URI: http://dx.doi.org/10.1109/MWC.001.1800603
http://hdl.handle.net/11536/153478
ISSN: 1536-1284
DOI: 10.1109/MWC.001.1800603
期刊: IEEE WIRELESS COMMUNICATIONS
Volume: 26
Issue: 6
起始頁: 180
結束頁: 186
Appears in Collections:Articles