摘要
随着大数据、虚拟现实以及人工智能等创新业务的发展,海量数据上传云端逐渐显现出带宽资源紧张、数据处理效率低下等问题。面对这些问题,边缘计算因其灵活性和高能效特性得到广泛发展。边缘数据中心因为更加靠近数据源,在定制化服务能力方面有更高的要求。针对该需求,文章提出一种边缘数据中心物理资源RO构建方法。通过RSD架构服务器可以实现物理资源重构,将计算、存储、网络等各类资源从紧耦合关系变成可以被软件调度的松耦合关系,实现根据业务负载匹配的硬件基础架构。实验结果表明,基于RSD的边缘数据中心构建方法,可以实现硬件资源池化、提高资源利用率,适合应用于边缘数据中心环境中。
With the development of innovative services such as big data, virtual reality and artifcial intelligence, massive data are being uploaded to clouds gradually, which shows problems such as bandwidth-limited network resources and ineffcient data processing. Faced with these problems, edge computing has been widely developed for its fexibility and energy effciency. Edge data centers have higher requirements for customized service capabilities because they are closer to data sources. Aiming at this demand, a method of constructing an edge data center through physical resources based on RSD and OpenStack is proposed. Through the RSD architecture server, physical resource reconstruction can be realized, and various resources such as computing, storage, and network can be changed from tightly coupled relationship to loosely coupled relationship that can be scheduled by software, and the hardware infrastructure based on service load matching is realized. Experimental results show that the RSD-based edge data center construction method can realize hardware resource pooling and improve resource utilization, and is suitable for application in an edge data center environment.
作者
郭玉华
贾宝军
侯乐
程筱彪
徐雷
Guo Yuhua;Jia Baojun;Hou Le;Cheng Xiaobiao;Xu Lei(China Unicom Research Institute,Beijing 100176,China)
出处
《信息通信技术》
2018年第5期16-21,共6页
Information and communications Technologies
