Augmented Computing at the Edge Using Named Data Networking
Augmented Computing at the Edge Using Named Data Networking
复制标题
使用命名数据网络进行边缘增强计算
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Y. Koucheryavy
中科院分区：
文献类型：
--
作者：
R. Pirmagomedov;S. Srikanteswara;D. Moltchanov;G. Arrobo;Yi Zhang;N. Himayat;Y. Koucheryavy
边缘计算被认为对物联网的发展至关重要，它能够为利用外部资源的受限设备及时执行各种计算任务。传统的基于主机的网络架构成为边缘计算进一步发展的瓶颈，主要是在服务延迟敏感的应用程序时。此外，现有的方法没有利用网络层中的复杂数据相关性进行优化。本文证明了命名数据网络（NDN）有可能为移动的用户提供有效的支持，将他们对时间敏感的计算任务卸载到边缘服务器上。为此，NDN协议通过服务器选择过程进行了增强，能够根据边缘服务器上不同的资源可用性进行调整。实验的结果显示了在这些场景中使用NDN的明确支持，不仅来自兴趣聚合和缓存（NDN功能），而且还来自动态服务器选择。
Edge computing is considered vital to IoT evolution, enabling the timely execution of various computational tasks for constrained devices utilizing external resources. The conventional host-based network architectures become a bottleneck for further development of edge computing, primarily when serving latencysensitive applications. Further, existing approaches do not exploit complex data correlations in the network layer for optimization. This paper demonstrates that Named Data Networking (NDN) has the potential to enable efficient support for mobile users offloading their time-sensitive computing tasks to edge servers. For this purpose, the NDN protocol was enhanced with a server selection procedure, capable of adjusting for the varying resource availability on edge servers. The results of the experiments show clear support for using NDN in these scenarios, with individual gains coming not just from Interest aggregation and caching, which are NDN features, but also from dynamic server selection.