服务部署中基于潮汐模型的负载调节策略

Load Regulation Strategy Based on Tide Model in Service Deployment

随着网络功能虚拟化和软件定义网络被提出和应用,许多虚拟服务功能被部署在基于云的数据中心。然而,如何有效地部署服务功能并动态地部署和分配资源仍然是尚待解决的问题。论文在服务部署过程中,提出潮汐机制来解决不同服务功能的资源再分配和动态调整的问题,避免了自动伸缩方法的冷却时间,减少服务延迟。通过对网络资源、虚拟资源和用户请求建立模型,利用短板理论找到最繁忙的虚拟服务,并启用潮汐虚拟机缓解流量拥塞。与其他方案相比,仿真结果验证了所提出方案的有效性。面对较大量的用户请求时,在系统延迟、丢包率降低以及平衡流量负载方面表现优秀。

As network function virtualization and software-defined networking are proposed and applied,many virtual service functions are deployed in cloud-based data centers.However,how to effectively deploy service functions and dynamically deploy and allocate resources is still an open question.In this article,in the service deployment process,a tidal mechanism is proposed to solve the problem of resource redistribution and dynamic adjustment of different service functions,avoiding the cooling time of the automatic scaling method and reducing service delay.The models of network resources,virtual resources,and user requests,use short-board theory are built to find the busiest virtual services,and enable tidal virtual machines are enabled to alleviate traffic con-gestion.Compared with other schemes,the simulation results verify the effectiveness of the proposed scheme.It performs better in terms of system delay,reduced packet loss rate,and balanced traffic load when facing a larger number of user requests.