Research On Quality Of Service In Cdma2000 1xEV-DO System

Cdma2000 1xEV-DO has been successfully put into commercial use in a large scale as a Beyond 3G technology, and its core network is evolving to all IP infrastructure gradually as a means of wireless access to the Internet. With the advantage of wide deployment, it not only provides high speed data transmission service, but also the number of the multimedia services run in its network is increasing. Consequently, the incidental challenges in Quality of Service (QoS) become the hotspots of research. This paper deeply analyzes and discusses the five issues in providing the QoS guarantees in the cdma2000 1xEV-DO system, which are the model and infrastructure of providing QoS, the packet scheduling, queue management, call admission control in the forward shared link and the end-to-end performance evaluation of a certain service in the system.Many data services are not sensitive to transmission delay, and their throughputs of the uplink and the downlink are unequal. So the traffic channel in the newly designed forward link of the cdma2000 1xEV-DO is time-multiplexed by all users in the system, and the traffic channel is transmitted to a targeted user with the maximum power of base station at any timeslot. Though it achieves great improvement of data throughput via opportunistic scheduling algorithms, being distinct from the traditional CDMA mobile networks, the shared using of traffic channel brings several problems, which include forward packet scheduling, queue management and call admissions control.This paper is addressed itself to the above issues, and proposes two schemes to simultaneously schedule data packets of the multi-class services. The two schemes are the Multi-Class Simultaneous Scheduling (MCSS) and the Layered Hybrid Scheduling (LHS), and they are all built on the base of the analysis of QoS requirements of the four service classes defined by the international mobile communications standard groups, which are the conversional, the streaming, the interactive and the background. MCSS firstly figures out the relative weight vectors for each of the four service classes via the integration evaluation of fuzzy decision mathematics, and it adopts different approaches to quantify the users' degrees of meeting their individual requirements on QoS attributes in the process of packet transmission. It subsequently gets the weighted sums of the quantified results for every user in the system, and finally makes scheduling decision according to the ranking of the weighted sums.In the LHS's first layer, it allots time slots in the prescribed ratios corresponding to the different traffic classes, because every traffic class has approximate transmission demand, LHS simplifies the packet scheduling belonging to the same traffic class. Then LHS chooses an appropriate scheduling algorithm for every class according to their individual QoS demands in its second layer, the selected algorithms are used to fulfill the allocations of timeslots among users belonging to the same traffic class.As of queue management, this paper analyzes the mode and size of queue in the forward link, along with packet discarding, then mainly explains the manner and the distinctions of packet discarding in order to mitigate the congestion in the forward link. Since discarding different users'packet achieves unequal gains of alleviating congestion, this paper proposes an Adaptive Congestion sensitive Wireless Packet Discarding Scheme (ACWPDS). ACWPDS divides the wireless congestion into three different grades, which are the slight, the middle and the severe, then adopts the corresponding mechanisms to decide how to discard users' packet.Data throughput is limited in the forward link of the 1xEV-DO system, and the degree of the congestion in its forward link exerts great influence on the call admission. Using the above property this paper proposes a Throughput allocation based Multi-class Services Admission Control scheme (TMSAC). TMSAC rejects all new arriving call when severe congestion occurs; when middle congestion happens, TMSAC rejects the new call if the data transmission of the corresponding service class deteriorates; if no congestion exits, or the congestion is slight, TMSAC decides the call admission based on the limit of the overall throughput and throughput allocation ratios.To evaluate the performance of a service in the networks, this paper lastly proposes an end-to-end QoS analytic method for the streaming media using a bottom-up approach, which is applicable to analyze the cumulative performance degradation along the service architecture from the physical layer to the application layer. By means of both mathematic analysis and computer simulations, three QoS attributes affecting the performance of the service are evaluated, which are the delay, throughput and loss rate, and the diversities in providing QoS among different protocol layers can be revealed.