一种采用公平策略的CSFQ算法

研究一种对核心无状态公平队列调度(CSFQ)算法的改进方法.针对CSFQ缓存资源管理的弊病,造成路由器突发性流响应差及对响应流TCP和非响应流UDP不公平的问题,提出了一种公平策略的CSFQ算法(FCSFQ).FCSFQ提出了基于共享存储器模式的动态缓存管理机制,并以此来改善CSFQ突发性响应差问题,同时采用对响应流与非响应流不同的丢包策略,来解决CSFQ对不同数据流不公平问题.NS仿真实验表明,在现实网络环境下,FCSFQ算法对突发性响应及带宽在TCP、UDP数据流之间的公平分配都有明显改善.     

CSFQ算法分析与改进

核心无状态公平队列调度（CSVQ）算法提供了如同有状态网那样好的公平带宽分配,但它的丢包算法不适用于TCP流。针对TCP流的特点,对CSFQ算法进行如下改进：将缓存队列长度与丢包概率关联起来,用一种类似于RED（random early drop）缓存管理方法解决了缓存频繁溢出导致的一些问题;对TCP流的丢包率进行修正,使用多余带宽来转发TCP包,解决TCP流与UDP流的带宽分配公平性。仿真试验表明,新算法NEW-CSFQ更好地提供数据流公平的频宽共享,对突发流响应较原算法有所提高,且算法复杂度简单,容易在高速核心路由器上实现。     

CSFQ算法的分析与改进

随着 Internet的迅猛发展 ,在 Internet中实现高效、公平、可扩展的业务流控制正变得日趋重要 .CSFQ ( Core-Stateless Fair Queueing)算法无需核心路由器为每一业务流保存状态但仍可实现相当程度的资源公平共享 .本文首先归纳总结了 CSFQ算法 ,然后给出了最大最小公平准则下公平共享速率的理想值、分析并揭示了 CSFQ算法存在的病态情形 ,最后对 CSFQ进行了改进 ,使得改进后的 CSFQ算法对公平共享速率的估算有了自适应机制及对因估算值偏小而丢弃的分组进行补救的机制     

URED:一种基于用户的缓存管理算法

宽带接入是当前热点问题,用户的流量管理是其实现的一个重要方面。根据宽带接入的需要,在分析现有缓存管理算法的基础上,提出了URED基于用户的缓存管理算法,这种算法采用虚拟用户队列和实际业务队列管理分组,将用户流状态引入到分组丢弃概率计算中,在用户区分的粒度上提供了用户流的管理并保证了业务流服务质量。通过分析和仿真,证明了URED算法的性能,比现有缓存管理算法更适合于接入路由器的要求。     

网络相关媒体流缓存管理策略NRC

Internet网络上媒体流的应用受限于应用时的网络状况，如时延、包丢失率等．本文研究通过在网络边缘处设置缓存代理来减小这些影响，提出一种新的缓存管理算法NRC，即接入媒体流服务时，用户以两种方式获取媒体流对象：一部分对象内容从代理缓存中获取，而另一部分对象内容则直接从流媒体源服务器处传输而来；从而加速媒体流接入服务，提高媒体流服务质量,算法同网络特性和媒体流特性相关．最后仿真实验证实同网络和流媒体特性相关的缓存管理算法NRC可以很好地减少服务延迟和提高媒体流的总体服务质量．     

一种基于公平性MPLS的流控制技术研究

多媒体业务在Internet上的广泛应用带来了响应流和非响应流间带宽享用的公平性问题,严重影响网络的性能.提出一种基于公平的MPLS流控制机制,该机制通过在MPLS模型中添加特定的队列管理算法CSFQ,在一定程度上解决了MPLS域中流间的公平性问题,并且其对于MPLS设备的扩展完全可以通过软件的升级来实现,有较好的可行性.     

路由器缓存管理算法之比较研究*

缓存管理是高性能路由器需要解决的技术难题之一,一个好的缓存管理算法可提高路由器的缓存资源利用率并降低分组丢失率.简要介绍了路由器中缓存管理的发展过程,列举了缓存管理一些最主流的算法,并对它们的性质、优缺点作了较为深刻的比较研究.最后利用试验仿真对四种缓存管理算法进行了缓存利用率和分组丢失率方面的评价,并对缓存管理算法的发展作了展望.     

客户端缓存容量的分区流切入算法

流切入算法对视频点播研究有重要影响。用流切入算法可以实现真视频点播系统,因此该算法实现的系统的用户满意程度很高,可以有效地降低用户的食言概率。该文提出了一种基于客户端缓存容量的分区流切入算法。通过以客户端缓存容量的大小作为分区依据将原始流分成若干切入区,进一步提高了通道利用率。文章最后进行了模拟实验,结果表明在一定的资源利用情况和视频点播服务器工作负荷下,分区流切入算法比流切入算法在通道利用率上效果更优,算法性能提高将近50%。     

基于路由器的公平带宽分配算法研究

本文介绍了目前在路由器上存在的几种带宽分配的不公平性问题,阐述了在路由器上如何避免这种不公平性,并对核心无状态公平队列调度(CSFQ)算法进行详尽分析,指出了CSFQ算法的问题并总结了目前的解决方法以达到带宽分配的尽可能公平.     

应用Best-Effort思想的公平排队算法

研究了以CSFQ为主要代表的核心路由器中无状态公平排队技术，针对其在吞吐量等性能上的不足，提出了一种结合CSFQ与FIFO两种技术的公平排队算法CSFQ—β．算法能够达到近似公平的带宽分配，在保持了CSFQ的其它优点基础上．更进一步地改善了总体吞吐量．减少了包的转发时延，并更有效地利用了链路带宽，且仍能避免拥塞的产生．     

TCP performance optimization in multi-cell WLANs

Though significant attention has been given to understanding the performance of a single-cell WLAN, performance evaluation of a group of interfering basic service sets (BSSs) within an extended service set (ESS) is still an open area. In this paper, we first demonstrate that a severe throughput imbalance occurs between downlink TCP flows even in the simplest of multi-cell WLANs via simulation and real world experiments; then, to solve this unfairness problem, we derive an analytical model that describes the interaction between TCP flows at the MAC layer, and formulate a throughput allocation problem as a nonlinear optimization problem subject to certain fairness requirements. Our formulation considers real world complexity such as hidden terminals, packet transmission retry limit, and the unique characteristics of TCP traffic. Solving our optimization problem yields the optimal MAC layer contention window settings that can lead each TCP flow to its target end-to-end throughput without the need for any per-flow queuing nor modification of the TCP sender. Simulation results show that our approach can achieve a fair allocation on the end-to-end throughput and attest to the accuracy of our proposed method.     

数据中心网络中基于传输速率分配算法的TCP协议

传输控制协议(Transmission Control Protocol,TCP)是数据中心网络中常用的传输协议。然而,同一个网络环境下,由于TCP协议的公平共享原则,TCP无法保障不同优先级业务的服务质量。针对该问题,提出了基于传输速率分配算法的TCP(Transmit Rate Allocation based TCP,TRA-TCP)。汇聚节点实时监测记录不同优先级业务的数据传输速率,并为不同优先级的业务流分配不同的传输速率,以保障高优先级业务的QoS。实验表明,与已有协议相比,TRA-TCP协议最高能够将高优先级业务的吞吐量提升50%,并且将高优先级业务的时延保持在0.1秒以下。     

Performance evaluation of bandwidth allocation methods in a geostationary satellite channel in the presence of internet traffic

Resource allocation represents an important issue for the next generation TCP/IP Quality of Service-based satellite networks. Many schemes, proposed in the recent literature, consider Internet traffic as the superposition of traffic sources without distinguishing between User Datagram Protocol (UDP) and Transmission Control Protocol (TCP) flows, even if UDP and TCP imply very different traffic characteristics. The basic idea of this work is that a resource allocation algorithm which is conscious of the difference may be more efficient because it can make use of the different behaviour of TCP and UDP in the presence of network congestion. Actually TCP reduces the source flow rate and, as a consequence, also the bandwidth occupancy when there is network congestion. The use of this feature within the bandwidth allocation scheme allows reducing the bandwidth waste due to over provisioning and using the residual bandwidth for other sources. The advantage is particularly clear over satellite channels where fading often affects the communication: having some residual bandwidth available for stations which have experienced fading can improve the satellite network performance.This paper presents a detailed performance evaluation of a bandwidth allocation scheme, called E-CAP-ABASC and studied for the satellite environment. The bandwidth is assigned to the earth stations that compose the network by a master station on the basis of a cost function whose main part is represented by a closed-form of the packet loss probabilities for the TCP and UDP traffic. The use of two different packet loss probability models for TCP and UDP allows exploiting the different features of the two traffic types, so improving the overall performance either in terms of packet loss or, on the other hand, in terms of the traffic admitted.The performance evaluation is carried out by varying the link degradation due to fading, the traffic load, and the flow balance between UDP and TCP. The results show a good performance of E-CAP-ABASC, compared with two other schemes. Advantages and drawbacks are discussed.     

TCP/IP traffic over ATM networks with FMMRA ABR flow and congestion control

In this paper, we study and compare the performance of TCP/IP traffic running on different rate based ABR flow control algorithms such as EFCI, ERICA and FMMRA by extensive simulations. The FMMRA algorithm is shown to exhibit the favorable features of least buffer requirement, fair bandwidth allocation to TCP connections, fast and accurate ACR rate adjustment according to the changes of network traffic, and the highest effective TCP throughput.     

基于RTT自适应的无线TCP改进算法

针对TCP Reno在无线环境下的性能恶化问题,在研究分析TCP Reno拥塞控制算法问题的基础上,提出一种基于RTT自适应的改进算法.该算法实现了丢包区分的拥塞窗口与慢启动门限调整,减轻了传统TCP由于无法区分拥塞丢包与误码丢包、盲目将拥塞窗口减半带来的性能下降.分析了该算法的可行性,并通过NS仿真对其吞吐量、带宽利用率、公平性等指标进行评估.仿真结果表明,相对TCP Reno,改进算法实现了无线环境下的TCP性能改善,同时具有一定的友好性与公平性.     

一种非TCP友好流惩罚算法

为了改善Internet中各流量带宽分配的不公平性,提出一种适于缓存管理的非TCP友好流的惩罚算法（NFTFP算法）。通过在缓冲区记录每个连接所占用的带宽,对于非TCP友好流量进行识别和惩罚,从而提高不同数据流享用带宽的公平性。仿真实验表明在标准TCP流量和非TCP友好流共存的环境下,NFTFP算法能提高对非TCP友好流的惩罚力度,在保持较高的链路利用率的同时,获得比RED算法更好的公平性。     

Packet Spacing: An Enabling Mechanism for Delivering Multimedia Content in Computational Grids

Streaming multimedia with UDP has become increasingly popular over distributed systems like the Internet. Scientific applications that stream multimedia include remote computational steering of visualization data and video-on-demand teleconferencing over the Access Grid. However, UDP does not possess a self-regulating, congestion-control mechanism; and most best-effort traffic is served by congestion-controlled TCP. Consequently, UDP steals bandwidth from TCP such that TCP flows starve for network resources. With the volume of Internet traffic continuing to increase, the perpetuation of UDP-based streaming will cause the Internet to collapse as it did in the mid-1980's due to the use of non-congestion-controlled TCP.To address this problem, we introduce the counter-intuitive notion of inter-packet spacing with control feedback to enable UDP-based applications to perform well in the next-generation Internet and computational grids. When compared with traditional UDP-based streaming, we illustrate that our approach can reduce packet loss over 50% without adversely affecting delivered throughput.     

A Lightweight Fairness-Driven AQM for Regulating Bandwidth Utilization in Best-Effort Routers

The end-to-end congestion control mechanism of transmission control protocol (TCP) is critical to the robustness and fairness of the best-effort Internet. Since it is no longer practical to rely on end-systems to cooperatively deploy congestion control mechanisms, the network itself must now participate in regulating its own resource utilization. To that end, fairness-driven active queue management (AQM) is promising in sharing the scarce bandwidth among competing flows in a fair manner. However, most of the existing fairness-driven AQM schemes cannot provide efficient and fair bandwidth allocation while being scalable. This paper presents a novel fairness-driven AQM scheme, called CHORD (CHOKe with recent drop history) that seeks to maximize fair bandwidth sharing among aggregate flows while retaining the scalability in terms of the minimum possible state space and per-packet processing costs. Fairness is enforced by identifying and restricting high-bandwidth unresponsive flows at the time of congestion with a lightweight control function. The identification mechanism consists of a fixed-size cache to capture the history of recent drops with a state space equal to the size of the cache. The restriction mechanism is stateless with two matching trial phases and an adaptive drawing factor to take a strong punitive measure against the identified high-bandwidth unresponsive flows in proportion to the average buffer occupancy. Comprehensive performance evaluation indicates that among other well-known AQM schemes of comparable complexities, CHORD provides enhanced TCP goodput and intra-protocol fairness and is well-suited for fair bandwidth allocation to aggregate traffic across a wide range of packet and buffer sizes at a bottleneck router.     

REBOOK: A Deterministic,Robust and Scalable Resource Booking Algorithm

This paper addresses the resource reservation problem in a packet switching network. REBOOK, the presented algorithm, provides deterministic, fast (real-time) dynamic resource allocation and release. Based on a stateful approach, it handles faults and network errors, and recovers from route changes and unexpected flows shutdown. The distributed scheme used to store flows information avoids the need of searching for entries within the routers’ control memory when packets are received and guarantees constant complexity. REBOOK can be implemented in hardware and is compatible with any packet switching network. In the Internet, it can be integrated in TCP or used with UDP to make it network friendly. Moreover, a slightly extended implementation of RSVP can be used as signalling and hosting protocol. A software implementation as standalone protocol has been developed to prove its effectiveness, robustness, and performance.