無線網路之媒體接取控制與排程機制

Abstract

在無線網路中保障服務品質是很重要的議題。現今，不同的網路對不同的服務最佳化，在每個網路中提供多樣性的服務變的越來越急迫了。為了要解決這個問題，網路中的媒體控制協定與排程機制需要再重新設計，來滿足多媒體服務的服務品質並提升系統效能。 在無線區域網路中，媒體接取控制機制的目的是解決使用者間相互競爭的問題。它可利用不同的仲裁訊框間隔 (Arbitration Inter Frame Space) 來區分不同的服務以提升服務品質，但是效能並不是很好。為了要在無線區域網路保證多媒體服務的服務品質，提出一個適應性p-persistentr (APP) 基礎的媒體接取控制法。APP媒體接取控制法依照不同的服務種類給予不同的允諾機率來服務多媒體使用者並利用傳送的允諾機率來區分使用者。當使用者有較大的延遲封包他會有較高的允諾機率。從分析與模擬結果中証實，APP可以降低使用者間的延遲變異達15%，並降低高優先權服務的失敗率。 在IEEE 802.16 都會型網路中，基地台要服務大量的使用者與提供不同種類的服務型態，所以需要排程機制提升系統效能，並兼顧使用者的服務品質需求。我們在IEEE 802.16 上鏈路提出了一個基於動態優先次序的資源分配(dynamic priority-based resource allocation, DPRA)機制，在下鏈路提出一個最大化系統產出並降低複雜度的功效式排程機制 (utility-based throughput maximization and complexity reduction (U_TMCR) scheduling)。 DPRA機制對於急迫性較高的服務，我們給予較高的優先次序值(Priority value)，使具有較高優先次序的使用者能優先被分配系統資源做傳輸。我們也會根據每一種服務在不同時間的急迫性，動態調整其優先次序。我們提出的DPRA機制會在子通道(subchannel), 調變方法(modulation order), 以及能量(power)三方面找尋最佳化的資源分配方法，並且對同一個使用者做一致性分配(consistent allocation)。由模擬結果顯示，我們提出的方法可以達到傳輸速率最佳化以及QoS的滿足，並且能減少標頭傳輸(transmission overhead)以及降低運算複雜度。 U_TMCR程機制不只在保證服務品質 (QoS) 的情形下最大化系統效能，同時也降低計算複雜度，並針對多媒體使用者做通道配置、天線選擇與決定調變方法。U_TMCR 機制根據通道品質和使用者的服務品質需求，為每個使用者設計效用函數 (utility function)，將排程問題轉成考慮系統限制對整個系統最佳化效用的問題。U_TMCR 機制也提出一個低複雜度演算法來解決所提出的最佳化問題。由模擬驗證，U_TMCR可以提升8%的系統效能並降低6.25%~29.2%的計算複雜度。

To guarantee the quality of service (QoS) in wireless network is an important issue. Currently, the different networks are optimized for different services, but it becomes urgent to provide varied service in wireless networks. To address this problem, there need to re-design a medium access control (MAC) or scheduling scheme to satisfy the QoS of multimedia service and to enhance the network utilization. In the WLAN, the goal of the medium access control (MAC) protocol is to deal with the contention of stations. It uses the different arbitration inter frame space to differentiate the services to promote the service quality, but the QoS satisfaction is not good enough. In order to support multimedia services in the WLAN, an adaptive p-persistent-based (APP) MAC scheme for IEEE 802.11 WLAN is investigated. The APP MAC scheme can further differentiate priorities of access categories by the initial permission probabilities and adaptively adjust permission probabilities to transmission stations according to its transmission state. Numerical and simulation results show that the APP MAC scheme can reduce the dropping probability of high priority service and effectively reduce the delay variance by 15%. The base station has to serve the massive users with different service type in IEEE 802.16 WiMAX system. Therefore, it needs to elaborately design the scheduling scheme to enhance QoS satisfaction with high system efficiency. In the IEEE 802.16 system, a dynamic priority resource allocation (DPRA) scheme for uplink and a utility-based throughput maximization and complexity reduction (U_TMCR) scheduling scheme for downlink are investigated. The DPRA scheme dynamically gives priority values to difference services based on the urgency degrees and allocates system radio resources according to the priority values. It can maximize the system throughput and satisfy differentiated QoS requirements. Also, the DPRA scheme performs consistent allocation to conform the uplink frame structure of IEEE 802.16, to fulfill QoS requirement, and to reduce the computational complexity. Simulation results show that the proposed DPRA scheme performs very close to the optimal method, which is by exhaustive search, in system throughput; and it outperforms the conventional EFS algorithm [39] in the performance measures such as system throughput, rtPS packet dropping rate, ratio of unsatisfied nrtPS, and average transmission rate of BE. The goals of the U_TMCR scheme are not only to maximize system throughput under QoS guarantee but also to reduce computational complexity. Based on channel quality and QoS requirements of each user, the U_TMCR scheme designs a utility function for every user and formulates the scheduling into an optimization problem of overall system utility function subject to system constraints. It also contains a heuristic TMCR algorithm to efficiently solve the optimization problem. Simulation results show that the U_TMCR scheme can improve the system throughput by 8% and reduce the computational complexity by 6.25%~29.2%.