A new constrained-send mechanism to enhance the performance of IEEE 802.11 DCF

In this paper, a modified access mechanism named Constrained-send DCF (CDCF) is proposed to improve the performance of IEEE 802.11 DCF protocol. It is found that, in DCF, the transmission probability is higher than the reasonable value when the node number is greater than 4 under basic access scheme or than 17 under RTS/CTS scheme, and it results in serious collision. To avoid collision of high access loading, a constrained-send probability is introduced at the end of each back off procedure for the station transmitting. The performance of this mechanism is analyzed based on a 2-Dimension Markov analytical model, after that the optimum constrained-send probability is derived. Numerical results show that the CDCF mechanism has much better performance than DCF with respect to both system throughput and average packet delay. And under RTS/CTS scheme, although CDCF slightly improve the throughput performance (due to the natural good throughput performance of RTS/CTS-scheme DCF), it leads to a much better average packet delay performance compared to DCF. The CDCF keeps all the features of the IEEE 802.11 DCF protocol and is quite easy to implement.     

高效的动态按需时隙分配协议

针对时隙分配中的预留冲突和网络效率较低的问题,提出一种海上无线环境的高效分布式动态按需TDMA时隙分配协议——DRSA(on-demand random slot assignment protocol).该协议可以根据网络节点自身的业务负载情况,在两跳范围内随机地为各节点分配时隙资源,且在分配时隙的过程中无需任何有关具体网络拓扑结构的先验信息.模型分析和仿真结果表明:即使在拥挤网络环境下DRSA协议仍然能够保证信息时隙的有效分配,使得网络以较高效率运行;对于因节点移动、业务分布变化而引起的网络环境及网络规模变化,DRSA协议也具有较强的适应性、健壮性和可扩展性.     

Adaptive mechanism for the frequency hopping MAC

For the frequency hopping media access control protocol (FH_MAC), there is a parabola relationship between throughput and traffic, and then an adaptive FH_MAC (AFH_MAC) is proposed for the maximum throughput, which is fit for timeliness traffic in aeronautical ad hoc networks. The paper works out the expression for the throughput, success transmission probability and average time delay based on the queuing theory, combination theory and discrete laplase transformation. The results show that, the time delay performance of the FH_MAC is better than that of the time division multiple access (TDMA) and IEEE 802.11 distributed coordination function (DCF), that the throughput and success transmission probability of the FH_MAC are higher than that of the ALOHA when the traffic load is low, and that the theoretic analysis of the AFH_MAC is in accordance with simulation, and the adaptive mechanism works well.     

基于多FPGA的NoC多核处理器验证平台设计

为了能够灵活地验证和实现自主设计的基于NoC的多核处理器,缩短NoC多核处理器的设计周期,提出了设计集成4片Virtex-6-550T FPGA的NoC多核处理器原型芯片设计/验证平台.分析和评估了NoC多核处理器的规模以及对FPGA硬件资源的需求,在此基础上给出了集成4片FPGA的开发板详细设计方案,并对各主要模块如互联架构、电源、板级时钟分布、接口技术、存储资源等关键设计要点进行阐述.描述了开发板各个主要模块的测试过程和结果,表明了该设计的可行性.     

LTE系统中提高TCP性能的资源调度算法研究

LTE可以提供真正无处不在基于IP的移动宽带业务,但随着承载网的IP化,网络拥塞、丢包、抖动、延时等质量问题将影响到LTE业务层的QoS质量.作为LTE无线资源管理的核心,研究并设计一个良好的资源调度算法是提高数据业务的性能和终端用户的体验是一个亟待解决的重要任务.本文通过借鉴LTE对VoIP数据分组的半持续调度算法的思想,提出了一种LTE的无线资源调度的改进方案.方案将TCP确认包映射到具有更高优先级的空闲逻辑信道,从而降低了ACK包在无线信道中发生丢弃和拥塞的概率,避免了TCP拥塞控制机制的频繁开启.仿真结果表明,本文提出的TCP确认包映射转换方案在RTT时延、吞吐量等方面均有一定的提升,具有一定的稳定性和性能优势.     

一种新型ICA算法在RFID系统中的应用

针对无线射频识别系统在通信过程中的防碰撞问题,从信号处理的角度,提出一种基于盲源信号并行分离技术的防碰撞处理算法,其中包括信号白化技术,以及一种改进的独立元分析(ICA)信号并行分离方法.功能仿真结果证明,该算法的吞吐率最高达到69％,与吞吐率只有18％的传统Aloha算法相比,该算法能提高防碰撞能力.     

卫星CICQ交换系统调度算法研究

联合输入交叉点排队(CICQ)交换结构由于在交叉点引入少量缓存,可以将输入端口和输出端口进行有效隔离,降低调度算法的复杂度,并适用于大容量交换。为此,研究基于交叉点缓存的各种调度算法和基于CICQ的交换结构,提出一种基于流量控制的FCSA算法,通过OPNET仿真分析表明该算法在均匀分布和突发业务源的情况下具有较好的时延性能,并且复杂度低,吞吐量大。将该算法应用于星载交换机,结果表明,该算法可以满足星载交换机多业务突发传输的特点,易于硬件实施。     

基于PCA的神经网络在港口货物吞吐量中的预测

分析选取了可能影响港口货物吞吐量的因素,采用PCA技术提取关键因子,最后以提取的关键因子作为神经网络的神经输入元,分别建立BP神经网络预测模型和GRNN神经网络预测模型。以上海港口为例,对港口货物吞吐量进行预测并对预测结果给予分析。     

The analysis of an adaptive workload balancing strategy in computing system resources management

An adaptive workload balancing scheme is analyzed. This scheme is intended to automatically dissolve the imbalance in resource utilization of computing systems and to bring about the proper balance between throughput and turnaround time. This is done by means of the invisible hand due to the bidding of priorities of resource use, with no contradiction between the flexible bid for each resource and the stable price for the entire service. The result of this analysis shows that this scheme would regulate those imbalances, and that it includes the load balancing policy of IBM OS/VS2-2 as a special case. Some considerations and methods on an actual implementation are also presented.This paper is a revised version of material which appeared as Research Report RC5072 of IBM Thomas J. Watson Research Center, Yorktown Heights, New York.On leave from the University of Electro-Communications, Chofu, Tokyo, Japan.     

Research on switching throughput of traffic manager in core routers

A general model is made to analyze switching throughput of traffic manager in core routers. By designing a real traffic manager that uses the OC-48c interface, the whole system is analyzed and it is pointed out that at least four HSSLs should be employed per CSIX interface when using Vitesse’s GigaStream switch chipset. Meanwhile, at the CSIX interface, the CFrame should be constructed according to the actual size of the last cell of each IP packet. The above principles can guarantee forwarding of IP packets at line rate. A general relationship between throughput and buffering scheme of IP packets in the external memory is given, which is useful in the design of switch fabric in core routers. __________ Translated from Acta Electronica Sinica, 2005, 33(7): 1243–1246 (in Chinese)