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.     

一种支持分级服务和最小带宽保证的改进TCP算法

本文提出了一种支持分级服务和最小带宽保证的改进TCP算法－DSAS－TCP，通过对TCP算中慢启动阈值和最小拥塞窗口的分级设置来使不同等级连接的平均吞吐量产生分级，并保证用户的最小带宽需求。通过理论分析和仿真实验验证了该算法的有效性。     

大带宽时延积网络TCP Vegas自适应慢启动算法

TCP Vegas具有比TCP Reno更好的带宽利用能力和稳定性,但是在带宽时延积较大的网络中,TCP Vegas会出现慢启动过早结束、拥塞窗口过小的问题,降低了传输效率.文中在分析慢启动结束的原因和条件的基础上,提出一种对临时性排队时延进行估计,将其排除后再进行慢启动结束条件判断的TCP Vegas慢启动算法,对不同网络条件有自适应能力.仿真结果表明本算法能有效避免慢启动过早结束,使TCP性能明显改善.     

即时恢复--一种新的TCP拥塞控制快速恢复算法

针对TCP连接的网络传输中经常会出现同一窗口多个分组丢失的情况,本文提出了一种新的TCP拥塞控制快速恢复算法:即时恢复算法。即时恢复算法能在恢复阶段探测网络的可用带宽,并将之用于窗口拥塞控制。另外,通过动态设定退出恢复阶段的参数域值,除了能恢复首次丢失的多个分组外,该算法还能很好的解决恢复阶段分组进一步丢失的情况。与SACK、FACK TCP等算法需要对TCP协议的发送、接收部分同时修改不同,该算法只需修改协议的发送部分,而接收部分不需作任何改变。在ns仿真环境的仿真结果表明:在同一窗口有多个分组丢失时,即时恢复算法的性能比TCP New-Reno有较大提高,与 SACK TCP性能相当。     

网格计算中的TCP拥塞控制

目前IP网络所应用的TCP拥塞控制机制是基于1988年Jacobson所设计的算法(慢启动和拥塞避免），虽然TCP在许多不同类型的网络中应用得很好，但在网络计算中。现有的TCP拥塞控制算法已不能有效工作。本分析了TCP传统算法在网格计算中的缺陷，并提出在罔格计算中使用新的TCP拥塞控制算法———个新的带宽增减算法。     

一种基于带宽估计的战术互联网TCP改进机制

无线的战术互联网通常面临复杂的电磁环境,高误码率、终端的频繁移动等恶劣的通信环境使得针对有线网络设计的TCP协议传榆性能急剧下降.针对战术互联网提出了一种基于带宽估计的TCP改进机制,该机制利用TCP确认帧携带的数据包到达时间来估算带宽,并用动态的低通滤波器来平滑带宽的估计值.在此基础上用带宽的估计值更新拥塞窗口,避免在发生链路错误时启动拥塞控制机制,由此提高TCP在战术互联网中的性能.实验结果表明,该算法能减少链路差错对TCP性能带来的影响,提高TCP在战术互联网上的传输性能.     

网格计算中的TCP拥塞控制

目前IP网络所应用的TCP拥塞控制机制是基于1988年Jacobson所设计的算法(慢启动和拥塞避免),虽然TCP在许多不同类型的网络中应用得很好,但在网格计算中,现有的TCP拥塞控制算法已不能有效工作。本文分析了TCP传统算法在网格计算中的缺陷,并提出在网格计算中使用新的TCP拥塞控制算法——一个新的带宽增减算法。     

TCP拥塞控制研究综述

系统深入地研究了TCP协议中的拥塞控制机制,TCP改进协议中研究最多的就是对TCP拥塞窗口调节机制的改进.从这个方面来看主要可以分为三种改进类型,即基于分组丢弃反馈的拥塞窗口调节方式;基于路径延时反馈的拥塞窗口调节方式;基于显式反馈和带宽测量的拥塞窗口调节方式.另外还针对无线网络的固有特性讨论了无线网络上TCP协议的一些改进方案.深入地分析了一些典型的算法,对它们的优劣做了详细的分析和总结,并提出了今后的一些研究方向.     

下一代传输协议SCTP拥塞控制方案浅析

文章介绍了一种新的传输层协议SCTP（流控制传输协议）,该协议的拥塞控制算法主要建立在TCP协议基于窗口的和式增加、积式减少机制之上,由慢启动、拥塞避免、快速重传和快速恢复四个核心算法构成,并且把TCP协议对SACK的可选支持变成必须支持的功能．使其在部分网络失效或者存在突发流量的情况下仍然能够正常工作。     

编码TCP的建模与性能分析

研究了无线网络端到端分组丢失和编码纠错对TCP协议工作过程的影响机制,基于三维Markov链建模描述TCP拥塞窗口和可用窗口的变迁过程,在忽略慢启动阶段的条件下,以最大窗口尺寸、端到端分组丢失率和编码冗余系数为输入参数,通过数值计算求解编码TCP的吞吐率,据此定量分析影响编码TCP性能的因素。基于NS2的模拟实验结果表明基于Markov链的数值计算结果具有较高的准确度。     

TCP with sender-side intelligence to handle dynamic, large, leaky pipes

Transmission control protocol Westwood (TCPW) has been shown to provide significant performance improvement over high-speed heterogeneous networks. The key idea of TCPW is to use eligible rate estimation (ERE) methods to intelligently set the congestion window (cwnd) and slow-start threshold (ssthresh) after a packet loss. ERE is defined as the efficient transmission rate eligible for a sender to achieve high utilization and be friendly to other TCP variants. This work presents TCP Westwood with agile probing (TCPW-A), a sender-side only enhancement of TCPW, that deals well with highly dynamic bandwidth, large propagation time/bandwidth, and random loss in the current and future heterogeneous Internet. TCPW-A achieves this goal by adding the following two mechanisms to TCPW. 1) When a connection initially begins or restarts after a timeout, instead of exponentially expanding cwnd to an arbitrary preset sthresh and then going into linear increase, TCPW-A uses agile probing, a mechanism that repeatedly resets ssthresh based on ERE and forces cwnd into an exponential climb each time. The result is fast convergence to a more appropriate ssthresh value. 2) In congestion avoidance, TCPW-A invokes agile probing upon detection of persistent extra bandwidth via a scheme we call persistent noncongestion detection (PNCD). While in congestion avoidance, agile probing is actually invoked under the following conditions: a) a large amount of bandwidth that suddenly becomes available due to change in network conditions; b) random loss during slow-start that causes the connection to prematurely exit the slow-start phase. Experimental results, both in ns-2 simulation and lab measurements using actual protocols implementation, show that TCPW-A can significantly improve link utilization over a wide range of bandwidth, propagation delay, and dynamic network loading.     

Cross-Layer-based Local Repair for Maximizing Goodput and Minimizing Control Messages in Multicasting MANET

The Multicast Ad hoc On-Demand Distance Vector (MAODV) protocol achieves multicast routing in self-organizing wireless mobile on-demand networks, e.g., Mobile Ad-hoc Networks (MANETs). However, unreliable wireless links degrade network reliability and network goodput, and the unreliable link problem becomes worse in multicasting because a multicast tree consists of more number of wireless links. MAODV adopts a broadcast-type local repair, and thus yields a large number of broadcast-type repair messages, increases extensive control overhead, and involves largely power consumption. Thus, a cross-layer unicast-type multihop local repair approach is proposed to recover broken links in multicasting MANETs. Additionally, the cross-layer mechanism provides mobile nodes to send a cross-layer message to the TCP sender to keep current congestion window (cwnd) and slow start threshold (ssthresh) when downstream links are temporarily broken, and then increases network goodput. Finally, the optimal number of neighbor-tiers is analyzed and the optimal substitute node is identified. Numerical results demonstrate that the proposed approach outperforms other approaches in successful repair rate, control message overhead, packet delivery ratio, and network goodput.     

Improving fairness of PGMCC

PGM congestion control (pgmcc), a single‐rate multicast congestion control scheme, is one of the most promising schemes which aim to solve the problem of multicast fairness with TCP. However, by deep investigation, we find that there exist two inappropriate mechanisms in this scheme, which are the fixed slow start threshold (ssthresh) and the mechanism of changing a group representative (acker). Our experiments have revealed the fact that these mechanisms can lead to the unfairness of pgmcc under some network conditions. In this paper, two new mechanisms have been proposed to replace the original ones. One mechanism is to make the sender adapt the value of ssthresh to the network conditions to mimic the action of TCP, and in order to avoid being more aggressive than TCP, the other one is to make the sender reduce the congestion window by half when the acker changes. Our experiment results, parts of which are discussed in this paper, show that the modified pgmcc can achieve better performance than the original one. Copyright © 2005 John Wiley & Sons, Ltd.     

The slow start power controlled MAC protocol for mobile ad hoc networks and its performance analysis

We propose and evaluate the performance of a new MAC-layer protocol for mobile ad hoc networks, called the Slow Start Power Controlled (abbreviated SSPC) protocol. SSPC improves on IEEE 802.11 by using power control for the RTS/CTS and DATA frame transmissions, so as to reduce energy consumption and increase network throughput and lifetime. In our scheme the transmission power used for the RTS frames is not constant, but follows a slow start principle. The CTS frames, which are sent at maximum transmission power, prevent the neighbouring nodes from transmitting their DATA frames at power levels higher than a computed threshold, while allowing them to transmit at power levels less than that threshold. Reduced energy consumption is achieved by adjusting the node transmission power to the minimum required value for reliable reception at the receiving node, while increase in network throughput is achieved by allowing more transmissions to take place simultaneously. The slow start principle used for calculating the appropriate DATA frames transmission power and the possibility of more simultaneous collision-free transmissions differentiate the SSPC protocol from the other MAC solutions proposed for IEEE 802.11. Simulation results indicate that the SSPC protocol achieves a significant reduction in power consumption, average packet delay and frequency of RTS frame collisions, and a significant increase in network throughput and received-to-sent packets ratio compared to IEEE 802.11 protocol.     

A Scalable Priority-Based Multi-Path Routing Protocol for Wireless Sensor Networks

In this paper, a scalable priority-based multi-path routing protocol (PRIMP) is proposed for wireless sensor networks to offer extended network lifetime and robust network fault tolerance, under the context of stringent energy constraint and vulnerability of sensors to dynamic environmental conditions. A novel interest dissemination strategy which invokes an on-demand virtual source technique is designed in PRIMP to minimize communication overheads and energy wastage. In routing, data traffic is distributed over multiple braided data paths simultaneously by a priority-based probabilistic approach at each hop to achieve the robustness against the unreliable transmission due to frequent node failures. Extensive simulations validate that PRIMP exhibits significantly better performance in energy conservation, load-balancing and data delivery than comparable schemes, while at the same time PRIMP achieves a nice scalability feature in terms of energy dissipation with various network sizes and network densities. Last but not least, PRIMP addresses the slow startup problem that is prevalent in data-centric routing schemes.     

Simple starting scheme for a parallel resonance inverter forinduction heating

A starting scheme for thyristor-based parallel resonant current source inverters, with particular reference to induction heating applications, is presented. The input stage to the DC-link is a three-phase phase controlled rectifier. It is shown that instead of using the usual startup circuits providing forced commutation, it is possible to start the system by using a single additional thyristor across the DC-link inductor, together with a special timed gating at the input and output converter. The scheme provides startup for a wide range of loads. The paper explains the principle starting concept and derives a simplified model to quantify the important parameters that govern the startup. The theoretical concept is verified by experimental data demonstrating the start of such an inverter     

Fast start crystal oscillator design with negative resistance control

Clock is an essential part of most of the integrated circuits as time base reference. It must be very accurate, highly reliable and readily available as soon as it is enabled. In all types of oscillator architectures, crystal oscillator is the most accurate and stable clock generator. But usually the crystal oscillator circuit suffers from slow startup. Therefore, it is essential to improve the startup time with optimally controlled crystal drive such that crystal drive power rating is not compromised. We propose a method that discusses about increasing the negative resistance during startup, using a startup circuit for fast start. Once the reliable startup is achieved, the negative resistance is decreased and the startup circuit is disconnected. The reduction in negative resistance is done with current steps and there are two ways in which it is achieved, the Digital control method and Analog control method. In digital control method, the current steps are timed at regular intervals and in analog control method, oscillator output amplitude is given as feedback to the startup circuit there by reducing the negative resistance. In the 32768 Hz real time clock generating oscillator, the startup time can be improved from 330 ms to 220 ms using the conventional startup method. With the proposed digital control method, lesser startup time of 160 ms is achieved and in analog control method it is further reduced to 120 ms.     

Dropping Policy with Burst Retransmission to Improve the Throughput of TCP Over Optical Burst-switched Networks

A major concern in optical burst-switched (OBS) networks is contention, which occurs when more than one bursts contend for the same data channel at the same time. Due to the bufferless nature of OBS networks, these contentions randomly occur at any degree of congestion in the network. When contention occurs at any core node, the core node drops bursts according to its dropping policy. Burst loss in OBS networks significantly degrades the throughput of TCP sources in the local access networks because current TCP congestion control mechanisms perform a slow start phase mainly due to contention rather than heavy congestion. However, there has not been much study about the impact of burst loss on the performance of TCP over OBS networks. To improve TCP throughput over OBS networks, we first introduce a dropping policy with burst retransmission that retransmits the bursts dropped due to contention, at the ingress node. Then, we extend the dropping policy with burst retransmission to drop a burst that has experienced fewer retransmissions in the event of contention at a core node in order to reduce the number of events that a TCP source enters the slow start phase due to contention. In addition, we propose to limit the number of retransmissions of each burst to prevent severe congestion. For the performance evaluation of the proposed schemes, we provide an analytic throughput model of TCP over OBS networks. Through simulations as well as analytic modeling, it is shown that the proposed dropping policy with burst retransmission can improve TCP throughput over OBS networks compared with an existing dropping policy without burst retransmission.     

基于RBF神经网络的光纤陀螺启动补偿及应用

光纤陀螺对温度比较敏感,由于温度引起的零偏漂移是光纤陀螺工作尤其是启动过程中的一种较大误差。文中为了减小光纤陀螺启动过程的零偏漂移、缩短启动时间,提出了对光纤陀螺启动过程进行补偿的方案。该方案以光纤陀螺温度和温度变化率为输入、光纤陀螺漂移为输出建立二输入单输出的RBF神经网络,用于陀螺启动过程补偿。在室温下对某型号光纤陀螺启动漂移进行了补偿,试验结果表明该方法能有效减小陀螺的启动温度漂移,缩短陀螺启动时间。将该方案运用到某型号的光纤陀螺寻北仪上,常温试验表明,该方案大大缩短了寻北仪的准备时间,提高了寻北精度。