This paper presents the state of the art of the Quality of Service (QoS) and mobility support mechanisms for mobile IP networks, which includes the issues and challenges in QoS support, an overview of the Mobile IP pr...This paper presents the state of the art of the Quality of Service (QoS) and mobility support mechanisms for mobile IP networks, which includes the issues and challenges in QoS support, an overview of the Mobile IP protocol, a general description of the QoS and Mobility framework, and the End-To-End QoS architecture in the next-generation all-IP mobile network.展开更多
Mobility management is a challenging topic in mobile computing environment.Studying the situation of mobiles crossing the boundaries of location areas is significant forevaluating the costs and performances of various...Mobility management is a challenging topic in mobile computing environment.Studying the situation of mobiles crossing the boundaries of location areas is significant forevaluating the costs and performances of various location management strategies. Hitherto, severalformulae were derived to describe the probability of the number of location areas' boundariescrossed by a mobile. Some of them were widely used in analyzing the costs and performances ofmobility management strategies. Utilizing the density evolution method of vector Markov processes,we propose a general probability formula of the number of location areas' boundaries crossed by amobile between two successive calls. Fortunately, several widely-used formulae are special cases ofthe proposed formula.展开更多
在5G边缘网络飞速发展的过程中,边缘用户对高带宽、低时延的网络服务的质量要求也显著提高.从移动边缘网络的角度来看,网络内的整体服务质量与边缘用户的分配息息相关,用户移动的复杂性为边缘用户分配带来困难,边缘用户分配过程中还存...在5G边缘网络飞速发展的过程中,边缘用户对高带宽、低时延的网络服务的质量要求也显著提高.从移动边缘网络的角度来看,网络内的整体服务质量与边缘用户的分配息息相关,用户移动的复杂性为边缘用户分配带来困难,边缘用户分配过程中还存在隐私泄露问题.本文提出一种移动边缘环境下基于联邦学习的动态QoS(Quality of Service)优化方法MECFLD_QoS,基于联邦学习的思想,优化边缘区域的服务缓存,在动态移动场景下根据用户位置分配边缘服务器,有效保护用户隐私,实现区域服务质量优化,对动态用户移动场景有更好的适应性.MECFLD_QoS主要做了以下几个方面的优化工作:(1)优化了传统QoS数据集,将数据集映射到边缘网络环境中,充分考虑边缘计算的移动、分布式、实时性、复杂场景等特点,形成边缘QoS特征数据集;(2)优化了边缘服务器缓存,在用户终端训练用户偏好模型,与区域公有模型交互时只传输参数,将用户的隐私数据封装在用户终端中,避免数据的传输,可以有效地保护用户特征隐私;(3)优化了用户移动场景,在动态移动场景中收集用户移动信息,利用用户接入基站的地理位置拟合用户的移动轨迹进行预测,有效地模糊了用户的真实位置,在轨迹预测的同时有效地保护了用户的位置隐私;(4)优化了用户分配方法,提出改进的基于二维解的人工蜂群算法对边缘网络中的用户分配问题进行优化,事实证明改进的人工蜂群算法针对其多变量多峰值的特点有效地优化了用户分配,达到了较优的分配效果.通过边缘QoS特征数据集实验表明,本方法在多变量多峰值的用户分配问题中能产生全局最优的分配.展开更多
文摘This paper presents the state of the art of the Quality of Service (QoS) and mobility support mechanisms for mobile IP networks, which includes the issues and challenges in QoS support, an overview of the Mobile IP protocol, a general description of the QoS and Mobility framework, and the End-To-End QoS architecture in the next-generation all-IP mobile network.
文摘Mobility management is a challenging topic in mobile computing environment.Studying the situation of mobiles crossing the boundaries of location areas is significant forevaluating the costs and performances of various location management strategies. Hitherto, severalformulae were derived to describe the probability of the number of location areas' boundariescrossed by a mobile. Some of them were widely used in analyzing the costs and performances ofmobility management strategies. Utilizing the density evolution method of vector Markov processes,we propose a general probability formula of the number of location areas' boundaries crossed by amobile between two successive calls. Fortunately, several widely-used formulae are special cases ofthe proposed formula.
文摘在5G边缘网络飞速发展的过程中,边缘用户对高带宽、低时延的网络服务的质量要求也显著提高.从移动边缘网络的角度来看,网络内的整体服务质量与边缘用户的分配息息相关,用户移动的复杂性为边缘用户分配带来困难,边缘用户分配过程中还存在隐私泄露问题.本文提出一种移动边缘环境下基于联邦学习的动态QoS(Quality of Service)优化方法MECFLD_QoS,基于联邦学习的思想,优化边缘区域的服务缓存,在动态移动场景下根据用户位置分配边缘服务器,有效保护用户隐私,实现区域服务质量优化,对动态用户移动场景有更好的适应性.MECFLD_QoS主要做了以下几个方面的优化工作:(1)优化了传统QoS数据集,将数据集映射到边缘网络环境中,充分考虑边缘计算的移动、分布式、实时性、复杂场景等特点,形成边缘QoS特征数据集;(2)优化了边缘服务器缓存,在用户终端训练用户偏好模型,与区域公有模型交互时只传输参数,将用户的隐私数据封装在用户终端中,避免数据的传输,可以有效地保护用户特征隐私;(3)优化了用户移动场景,在动态移动场景中收集用户移动信息,利用用户接入基站的地理位置拟合用户的移动轨迹进行预测,有效地模糊了用户的真实位置,在轨迹预测的同时有效地保护了用户的位置隐私;(4)优化了用户分配方法,提出改进的基于二维解的人工蜂群算法对边缘网络中的用户分配问题进行优化,事实证明改进的人工蜂群算法针对其多变量多峰值的特点有效地优化了用户分配,达到了较优的分配效果.通过边缘QoS特征数据集实验表明,本方法在多变量多峰值的用户分配问题中能产生全局最优的分配.
