【作者】 王佳佳；

【导师】 徐大雄； 宋俊德；

【作者基本信息】 北京邮电大学 ， 电磁场与微波技术， 2010， 博士

【摘要】 满足所承载业务的服务质量(QoS),使任何用户随时随地都能获得具有QoS保障的服务,是网络的设计目标之一。伴随着移动通信网络的宽带化和宽带无线网络的移动化进程,基于全IP的异构网络共存和融合已成为下一代无线网络的重要特征。全IP融合网络环境下,多种网络技术并存,各种网络有不同的传输速率、信道特性、系统容量、资源状况、QoS分类、资源预留方式等,不同网络对业务的承载能力和方式不同,如何在业务接入,业务端到端建立和维护以及移动环境达到业务高QoS要求,满足下一代异构网络多样性和服务高质量要求,成为无线异构网络无缝集成的关键。本文首先对无线网络的发展和演进进行了总结,并分析了未来网络的全IP化融合趋势,之后对QoS的定义和分类进行介绍,并详细阐述了当前移动通信网络和IP网络的服务质量保障相关技术,在此基础上,从端到端、接入和移动等方面对全IP融合网络中所面临的QoS问题进行分析、整理和归纳。在端到端QoS保障方面,本文提出了基于策略和下一代信令的端到端QoS保障框架,详细设计了框架内的域内纵向管理控制策略和域间横向信令交互流程；基于所提出的框架,提出了基于最小均方误差的端到端SLA协商算法,根据实际的端到端性能,自适应调整业务在端到端路径上不同网络域能够承诺的服务等级。仿真结果验证了所提出的方案能够自适应调整业务等级,实现不同网络域之间的SLA协商,满足业务通过端到端路径的QoS需求。在业务接入方面,本文提出了异构融合网络环境下的分级联合接纳控制机制(HJCAC, Hierarchical Joint Call Admission Control),在网络侧分别构建本地接纳控制实体和分级联合接纳控制实体,并协同终端共同为业务作出接纳判决和最优接入网络选择。仿真结果证明所提出方案能够在不改变各个网络已有的接纳控制策略的基础上,统一各个网络资源,轻负载时能有效提高业务接纳率10.96%,重负载时仍能提高4.71%；之后,本文提出了自适应接入模式选择算法。终端根据实时监测的网络特性,自适应选择以多跳模式或蜂窝集中模式接入网络,仿真结果表明该算法能够为业务选择合适的模式接入网络,在提高吞吐量同时,维持不同终端之间的公平性。在移动场景下,为解决传统QoS信令机制在移动时的双重预留问题,本文提出了具有移动性感知的QoS信令机制(QNFH,QoSNSLP with FHMIPv6)。所提出方案结合FHMIPv6和NSIS,利用L2检测提前触发三层FHMIPv6,进而触发上层QoS NSLP的提前预留,切换动作完成后,只需对提前预留路径上的无状态预留进行激活。仿真结果表明,相比于FHMIPv6和NSIS协议独立的方案,对于C-Mode模式,新旧路径上的公共部分跳数每增加1跳,所提出方案能够减少切换前后的预留重建时延约1.06ms,并减少信令重建开销约1.87%；对于D-Mode模式,新旧路径上的公共部分跳数每增加1跳,所提出方案能够减少切换前后的预留重建时延约0.98ms,并减少信令重建开销约1.09%。针对路由算法收敛慢和网络环境变化快的矛盾,本文还提出了基于位置信息的智能QoS路由机制,该机制采用蚁群算法,在禁忌表确定阶段添加位置判决函数用于减小算法搜索范围,同时在启发函数内添加位置相关信息和QoS需求信息,使得算法尽可能向最优解逼近。仿真结果验证了所提出算法能够保证在不影响最优寻路效果的同时,提高算法的收敛速度平均达36.96%。本文最后对全文进行了总结,指出了论文工作的不足,并为下一步的研究提出了若干建议。更多还原

【Abstract】 Satifying with QoS requirements of beared services and making any users obtain services with QoS guarantee anytime and anywhere, is one target of designing networks. With the broadband development of mobile telecommunications network and the mobility development of broadband wireless networks, all-IP heterogeneous networks coexistence and integration is becoming an important feature of next generation wireless networks. In the all-IP integrated networks, multiple networks coexist. Various networks have different transmission rates, channel characteristics, system capacity, resource conditions, QoS classifications, resource reservation methods and so on. The capabilities and methods to bear services are different from each other. To meet the high QoS requirements for next generation heterogeneous networks in service access, end-to-end establishment and maintenance, and even in mobile scenarios, are the key problems to realize the seamless integration of wireless heterogeneous network.The thesis firstly summarized the development and evolution of wireless networks and analyzed the all-IP trend of future networks. Definition and classification of QoS were also explained. Moreover, QoS related technologies were investigated in mobile telecommunication and IP networks. The difficult problems in end-to-end, access and mobile scenarios for QoS guarantee for all-IP heterogeneous networks were also addressed on.In the aspect of end-to-end QoS, the thesis defined an end-to-end QoS provisioning framework, which based on policy and NSIS (Next Steps in Signaling). Vertical management in intra-domain and horizontal signaling negotiation between different network domains were also defined in detail. Based on the proposed framework, the thesis proposed an end-to-end SLA negotiation mechanism taking use of least mean square algorithm. According to the error between measured and negotiated ideal performance along end-to-end path, the proposed mechanism can adjust the level of service in different QoS domains adaptively. Simulation results proved that the proposed mechanism can adjust level of services adaptively, negotiate between different network domains, and satisfy with the QoS requirements along end-to-end path.In the aspect of service access, the thesis proposed hierarchical joint admission control mechanism (HJCAC, Hierarchical Joint Call Admission Control) for heterogeneous integrated networks. Two entities, which included LCAC (Local call admission control) and HJCAC (hierarchical joint call admission control), were added in the side of networks. LCACs and HJCACs cooperated with terminals to make the admission decision and selected the best network for service to access. Traditional admission control policy in each network did not need to make any change. Simulation results proved that proposed mechanism can unify resources in every network, improve the acceptance ratio by 10.96% in light load and 4.71% in heavy load. Moreover, the thesis proposed an adaptive access mode selection algorithm. Mobile terminals can choose access mode between multi-hop and cellular mode by monitoring the interface performance periodly. Simulation results indicated the proposed algorithm can select the most appropriate mode to access the network, enhance the network throughput and keep the fairness among different multimode terminals.In the aspect of mobile scenario, a mobility-aware QoS signaling mechanismQNFH (QoS NSLP with FHMIPv6) was proposed. The aim of the proposed mechanism was to solve the double reservation problem in mobile environment for traditional QoS signaling. The proposed mechanism associated the FHMIPv6 with NSIS. L2 location information can trigger the L3 handover and the resource reservation in advance. Once the L3 handover finished, only the activation of stateless reservation along the previous reservation path was needed. With the increase of hops in the common part of new path and old path, the proposed mechanism can reduce the reservation re-establishment delay by 1.06ms and signaling re-establishment cost by 1.87% for each hop in C-mode. In D-mode, the proposed mechanism can reduce the reservation re-establishement delay by 0.98ms and signaling cost by 1.09% for each hop.In order to solve the conflict between slow routing algorithm convergence rate and rapid change of network topology, an intelligent QoS routing algorithm based on location information was proposed. The proposed algorithm took use of ant colony algorithm. Location judgement function was added in the tabu list in order to reduce the search area. Also the location related information and QoS requirement information were added in the heuristic function, which can force the algorithm to converge to the best solution. Comparing with traditional algorithm, simulation results verified that the proposed algorithm can not only find the best path for the algorithm, but also improve the convergence rate of the algorithm by 36.96%.Finally, the summary of the thesis was given. The shortcomings of the research were pointed out and suggestions for future research were also listed.更多还原