MEC增强FiWi网络协同卸载感知动态波长带宽分配算法  

Cooperative offloading-aware dynamic wavelength and bandwidth allocation algorithm for MEC empowered FiWi networks

作　　者：于阳阳 杜俐洁 何荣希[1] YU Yangyang;DU Lijie;HE Rongxi(College of Information Science and Technology,Dalian Maritime University,Dalian Liaoning 116026,China)

机构地区：[1]大连海事大学信息科学技术学院,辽宁大连116026

出　　处：《光通信技术》2022年第5期1-10,共10页Optical Communication Technology

基　　金：国家自然科学基金项目(61371091、61801074)资助;大连市科技创新基金项目(2019J11CY015)资助。

摘　　要：为满足多接入边缘计算(MEC)增强FiWi网络中计算卸载业务和传统宽带业务的差异化需求,基于混合调度和协同卸载机制提出一种协同卸载感知(COA)的动态波长带宽分配(DWBA)算法。该算法利用在线调度回传计算结果,将离线子周期分为计算卸载子周期和宽带业务子周期。光线路终端协同调度波长带宽分配和卸载决策,优先采用水平计算协同卸载方式为超载MEC服务器选择卸载目标,将剩余计算需求卸载到远程集中云。仿真结果表明:该算法能显著提高计算业务完成率,减少平均完成时间,降低对传统宽带业务和核心网负载的不利影响。To meet the differentiated requirements of computing offloading traffic and traditional broadband traffic in multi-access edge computing(MEC) empowered fiber-wireless access networks, this paper proposes a cooperative offloading-aware(COA)dynamic wavelength and bandwidth allocation(DWBA) algorithm based on hybrid scheduling and cooperative offloading mechanism. The algorithm returns computing results by online scheduling, and divides the offline sub-cycle into the computing offloading sub-cycle and broadband traffic sub-cycle. The optical line terminal collaboratively schedules wavelength and bandwidth allocations and offloading decisions, and selects the offloading target for the overloaded MEC server by horizontal computing cooperative offloaing scheme preferentially, and then it offloads the remaining computing requirements to the remote centralized cloud.The simulation results show that the algorithm can improve the completion rate and reduce the completion time of computing tasks, and reduce the adverse impact on traditional broadband traffic and core network load.

关 键 词：光纤无线接入网 多接入边缘计算 动态波长带宽分配 协同卸载 

分 类 号：TN915.6[电子电信—通信与信息系统]