基于突发数据流的丢包智能区分仿真

针对2种不同TCP算法控制下的数据流,后进入的流量容易引起丢包率增加的问题,提出将网络模型中传输层和网络层的拥塞控制算法相结合,对突发数据流引起的丢包率起到一个缓解作用.在该网络环境下,对随机早期检测算法作了改进,通过仿真,新算法在一定程度上能够区分突发流量,从而降低丢包率,为高可靠性网络的发展提供参考.     

异构网络环境中的拥塞丢包率仿真分析

针对TCPVegas算法在异构网络环境中流量公平性导致的拥塞丢包问题,将TCPVegas拥塞控制算法与主动队列控制策略来行结合分析,提出了将网络模型中不同层次的拥塞控制机制进行结合的算法。网络仿真表明,结合算法能有效避免拥塞丢包。该算法对随机早期检测算法进行了改进,使其能够区分突发流量,从而降低拥塞丢包率,这也将为高可靠性网络的发展提供一个优良的参考价值。     

TCP Vegas拥塞控制机制的改进

针对Vegas拥塞控制技术的不足，提出了一种改进方法。首先对发送分组根据长度进行分类，分别计算各长度类别的“往返时间”（RTT）和RTT最小值（base_rtt）；在更新发送窗口以及设置重传定时器时，根据分组长度选择属于同一长度范围的RTT和base_rtt值作为输入参数。这在一定程度上消除分组长度差异对发送窗口更新和重传定时器设定值所产生的影响，进一步提高协议吞吐率。仿真结果表明，改进的Vegas能够将原协议的吞吐率提高多达24％。     

Analysis and enhancement of TCP Vegas congestion control in a mixed TCP Vegas and TCP Reno network scenario

Its more refined congestion control mechanisms, also based on the estimation of round trip delays, allow TCP Vegas to outperform the more widespread TCP Reno congestion control, based only on the packet loss detection, in a number of network environments. However, these mechanisms make TCP Vegas less aggressive with respect to TCP Reno; thereby TCP Vegas sources show high weakness in taking the available bandwidth when competing with other TCP Reno sources. This is a major reason that hinders the spread of TCP Vegas among Internet users. In this work, after a preliminary analytic study about the limits of TCP Vegas in mixed network environments, we describe a new adaptive mechanism for TCP Vegas, called TCP NewVegas, designed in order to improve its performance even in heterogeneous network scenarios. The large number of simulations, presented in this paper, show that TCP NewVegas guarantees good performance even in mixed network environments, without canceling the desirable features (e.g. fairness) that TCP Vegas exhibits in homogeneous environments.     

具有反馈时延的TCP Vega s 拥塞控制算法的稳定性分析

将Internet拥塞控制算法看作一个复杂的动态非线性反馈系统，借助多变量频率域控制系统的广义Nyquist判据和时延微分方程的Pontryagin判据，研究了各通信回路时延相同和时延不同条件下Internet网络拥塞控制算法，得到了具有反馈时延的TCP Vegas拥塞控制算法在平衡点渐近稳定的多个判据。仿真结果说明了稳定性判据的有效性和实用性。     

Bacterial contamination due to point and nonpoint source pollution in a rapidly growing urban center in an arid region

The distribution of bacterial contaminants (Escherichia coli and total coliforms) and nutrients (nitrogen and phosphorus) in a rapidly growing metropolitan area, the Las Vegas Valley, were studied to assess the contributions of point and nonpoint pollution sources under dry weather flow (pre rainfall) and wet weather flow (post rainfall) events. Nonpoint sources emanating from urban centers and septic seepage were found to be a primary source of bacterial contamination in the Las Vegas Wash, a mainstream wash channel where the Las Vegas Valley watershed drains. Contribution of point sources (wastewater treatment plants) to E. coli and total coliforms were found to be negligible. However, data in the Tropicana/Flamingo Wash, a tributary of the Las Vegas Wash, showed high bacterial activities in the wet antecedent rainfall condition as compared to the pre-rain dry antecedent condition suggesting that the rainfall increased bacterial activities.     

Filtration effectiveness of HVAC systems at near‐roadway schools

Concern for the exposure of children attending schools located near busy roadways to toxic, traffic‐related air pollutants has raised questions regarding the environmental benefits of advanced heating, ventilation, and air‐conditioning (HVAC) filtration systems for near‐road pollution. Levels of black carbon and gaseous pollutants were measured at three indoor classroom sites and at seven outdoor monitoring sites at Las Vegas schools. Initial HVAC filtration systems effected a 31–66% reduction in black carbon particle concentrations inside three schools compared with ambient air concentrations. After improved filtration systems were installed, black carbon particle concentrations were reduced by 74–97% inside three classrooms relative to ambient air concentrations. Average black carbon particle concentrations inside the schools with improved filtration systems were lower than typical ambient Las Vegas concentrations by 49–96%. Gaseous pollutants were higher indoors than outdoors. The higher indoor concentrations most likely originated at least partially from indoor sources, which were not targeted as part of this intervention.     

美国建筑考察之旅

本文对美国城市旧金山、洛杉矶、纽约、华盛顿、拉斯维加斯、圣地亚哥的特色建筑进行描述,揭示了隐藏在建筑表象下的美国多元文化.     

针对逆向流的Reno和Vegas算法性能的仿真分析

网络拥塞已经成为制约网络发展和应用的一个瓶颈.在TCP中已经实现的Reno和Vegas均是基于端到端的拥塞控制算法,它们虽然减轻了前向路径上的拥塞程度,但不能改善后向路径上数据流拥塞所导致的诸如吞吐率、丢包率等网络性能降低这一状况.基于"包丢失是由于网络拥塞引起"的假定,利用OPNET网络仿真软件,建立了单个瓶颈网络,设定了吞吐量、丢包率以及拥塞窗口(cwnd)等参数,对Reno和Vegas两种算法在逆向流中的性能进行比较,分析影响网络吞吐量的因素.仿真结果表明,在非对称网络中,Reno比Vegas更灵活且更实用,后向路径上逆向流的拥塞会极大地降低Vegas的吞吐量,使得Vegas和Reno不能很好地兼容.     

TCP Vegas-O:一种新的基于延迟估计的TCP Vegas改进算法

针对TCP Vegas会出现慢启动过早结束、拥塞窗口过小导致带宽利用率下降的问题,以及在与Reno等基于丢包来判断拥塞的算法竞争带宽时的公平性差等问题,文中分别就慢启动和拥塞避免阶段进行了相应的改进,最后将其结合.仿真结果表明,该算法对慢启动过早结束、带宽公平性等TCP Vegas协议的缺陷有了明显的改善,特别是在高带宽时延乘积网络中.