【作者】 段晓飞；

【导师】 吴玉成；

【作者基本信息】 重庆大学 ， 通信与信息系统， 2010， 硕士

【摘要】 随着宽带无线通信技术的发展以及多媒体业务的广泛应用,用户要求宽带无线接入技术能够提供服务质量(QoS)保证。为了很好地满足未来宽带无线接入的应用需求,IEEE组织提出了802.16系列标准。IEEE802.16标准定义了物理层技术和媒体接入层(MAC层)机制,并且在MAC层划分了业务的调度类型,使不同类型的业务连接关联不同的QoS参数,从而提供了对业务QoS的支持。此外,协议中还给出了用于主动授权业务(UGS)的带宽调度方法,但是并没有对其它类型业务的具体调度算法以及系统的调度结构予以说明,这些都是带宽调度的重要部分,而带宽调度又是保证各类型业务QoS的关键。因此,对802.16系统中带宽调度结构及算法的研究具有十分重要的意义。论文对现有的调度架构以及相关算法进行了深入的研究和分析。大部分算法关注的焦点都只是在保证业务的QoS要求方面,没有兼顾对系统调度服务公平性的考虑。为了在提供业务QoS保证的同时解决调度过程中服务不公平的问题,提出了一种基于预分配机制的带宽调度算法。该算法首先为高优先级业务预分配一部分带宽资源以保证其QoS对数据速率的要求,对带宽的剩余部分采用调整优先级的亏空公平队列算法在各业务之间再次进行分配,直到可用的带宽为零或者所有业务流都获得了足额的带宽为止。为实现不同SS的各业务连接能够公平有效地使用上行带宽资源,设计了GPSS(Grant per Subscriber Station)方式下的系统带宽调度模型,其中在BS端加入了加权最大最小公平算法,在SS端采用了基于预分配机制的带宽调度算法。加权最大最小公平算法由于考虑了不同业务对QoS的不同要求,实现了带宽资源在各SS之间的公平分配。基于预分配机制的带宽调度算法在为高优先级业务提供QoS保证的同时,又使得低优先级业务也能获得一定的服务机会,从而在一定程度上解决了高优先级业务过多地占用带宽资源致使低优先级业务长期得不到带宽而“饿死”的问题。基于GPSS方式的系统带宽调度模型由于融合了加权最大最小公平算法和基于预分配机制带宽调度算法,所以提高了SS带宽请求的服务率,虽然该策略使实时轮询业务(rtPS业务)的吞吐量和时延特性有局部性的损失,但是仍然保证了各类型业务的相应QoS要求,并且兼顾了对低优先级业务的带宽分配,很好地实现了从用户节点到业务的调度公平性。更多还原

【Abstract】 With the development of broadband wireless communication technology and more and more applications of the multimedia services, broadband wireless access technology should have the ability to guarantee the quality of services (QoS) required by users. In order to meet the application requirements of broadband wireless access in the future, IEEE organization has proposed a series of 802.16 standards. In IEEE802.16 standards the technology adopted by physical layer and the mechanism of MAC are defined detailedly. To support the QoS, services are divided into different scheduling types according to their QoS characters, and various service connections are associated with different QoS parameters. Furthermore, the scheduling algorithm for unsolicited grant service (UGS) is defined, but the standard don’t give explanation for scheduling structures and concrete scheduling algorithms for the other types of services. Therefore, studying on bandwidth scheduling structures and algorithms of 802.16 system has extremely vital significance.The proposed scheduling architectures and algorithms are studied and analyzed in depth in the thesis. Most algorithms are just focus on the aspect of guaranteeing the QoS of services, but not giving considerations to the scheduling fairness of system. Thus a novel bandwidth scheduling algorithm based on pre-allocation mechanism is proposed to provide guarantee for the QoS of services and to resolve the problem of servicing unfair in the scheduling process. To make sure the request of data rate of high priority services, this algorithm will pre-allocate some bandwidth resource for them firstly. Then for the rest of bandwidth, the real time polling services, non-real time polling services and best effort services are allowed to share through the changed priority deficit queue algorithm. The scheduling will not be stopped until the usable bandwidths are zero or all service flows have got required bandwidths.In order to realize that uplink bandwidths could be used fair and efficiently among various service connections which is belong to different SSs, a model of bandwidth scheduling based on Grant Per Subscriber Station (GPSS) mode is established. In this model, the weight max-min fair algorithm is introduced to base station, and the novel bandwidth scheduling algorithm based on pre-allocation mechanism is used by subscriber stations. Because it has considered that various services have different requirement for QoS, so the weight max-min fair algorithm could allocate bandwidths fairly among subscriber stations. The proposed novel algorithm makes the lower priority service flows obtain some service opportunities while the high priority service flows’ QoS are guaranteed as well. From the above, we can see that it has the ability to resolve the problem that the lower priority service flows are starved because most bandwidths is occupied by the high priority service flows and little bandwidths could be provided to the lower priority service flows for a long time in a certain degree.The weighted max-min fair scheduling algorithm and the bandwidth scheduling algorithm based on pre-allocate mechanism are embedded into the designed system model, so the service rate of the subscriber station’s bandwidth requests could be raised. Although the throughput and delay characteristic of real-time polling services are lost in some cases, but it gives consideration to allocate bandwidths for the lower priority service flows, so it has achieved scheduling fairness well from user nodes to services.更多还原