Traffic locality characteristics in a parallel forwarding system

Due to the widening gap between the performance of microprocessors and that of memory, using caches in a system to take advantage of locality in its workload has become a standard approach to improve overall system performance. At the same time, many performance problems finally reduce to cache performance issues. Locality in system workload is the fact that makes caching possible. In this paper, we first use the reuse distance model to characterize temporal locality in Internet traffic. We develop a model that closely matches the empirical data. We then extend the work to investigate temporal locality in the workload of multi‐processor forwarding systems by comparing locality under different packet scheduling schemes. Our simulations show that for systems with hash‐based schedulers, caching can be an effective way to improve forwarding performance. Based on flow‐level traffic characteristics, we further discuss the relationship between load‐balancing and hash‐scheduling, which yields insights into system design. Copyright © 2003 John Wiley & Sons, Ltd.     

粒度自适应的多径流量分割算法

针对多径传输存在的负载均衡和分组乱序之间的矛盾,提出了一种粒度自适应的多径流量分割算法。该算法在“流分片”的基础上,通过对流量特性的分析,将网络流量区分为正常流和强势流,基于“区分服务”的思想,不同流量分别采用适应其流量特性的流片分割粒度。理论分析和仿真结果表明,在考虑路径时延估计误差的情况下,该算法在降低分组乱序的同时具有良好的负载均衡性能。     

Flow-based load-balancing architecture for the agile all-photonic network

A novel routing architecture that balances incoming Internet flows over the agile all-photonic network (AAPN) is proposed. The architecture is based on the adaptive highest random weight (adaptive HRW) algorithm proposed to design load-balanced Internet routers. Extensive numerical evaluation of static and adaptive variations of the routing architecture is studied, and their effect on the network performance in terms of packet drop and flow remapping is presented. The architecture can be seen as a combination of adaptive core node scheduling and adaptive load balancing at the edge nodes. It is stateless and can compute routes quickly based on the packet flow identifier.     

基于Intel网络处理器的微处理器调度算法

针对网络处理器的核心问题一微处理器并行调度，以Intel IXP2400网络处理器为例．提出并实现了一种基于线程池的两阶段微处理器调度算法TS—MPSBPT。该算法将IPv4转发系统分成两个阶段．阶段内微处理器采用线程池工作模式，既解决了微指令空间不足问题，又可以充分发挥系统并行处理特性：该算法通过判断微处理器的空闲线程数和已处理数据包字节数，将数据包分配给线程池中负载最轻的微处理器。实现算法的负载均衡原则：通过将同类型m包或同一个流IP包分配给同一个微处理器，提高eache命中率和局部内存利用率．实现算法的局部性原则。实验结果表明，采用TS—MPSBPT算法的IPv4转发系统负载均衡，与Radisvs公司原IPv4转发程序相比．新系统包转发速率有较大提高．在包长度较小时效果尤其显著。     

基于软件定义网络的服务器集群负载均衡技术研究

在当前的网络体系结构下,采用硬件系统实现服务器集群负载均衡存在着获取负载节点状态困难、流量导向方式复杂等制约因素,不利于提升服务器集群的伸缩性和服务性能。针对此问题,该文提出一种基于软件定义网络(SDN)的负载均衡机制(SDNLB)。该机制借助SDN具有的集中式控制和流量灵活调度优势,利用SNMP协议和OpenFlow协议对服务器的运行状态和全局网络负载信息进行实时监测,并通过权值计算的方式选择出权重最高的服务器作为流处理的目标服务器,在此基础上,采用最优转发路径算法进行流量调度,从而达到提高服务器集群的利用率与处理性能的目的。搭建了实验平台对SDNLB的性能进行仿真测试,实验结果表明:在相同的网络负载条件下,SDNLB与其他负载均衡算法相比,能够有效地降低服务器集群的负载,并能够显著提高网络吞吐量和带宽利用率,缩短流的完成时间和平均时延。     

无线多跳网络的自适应容量感知优化与控制技术

In wireless multimedia communications, it is extremely difficult to derive general end-to-end capacity results because of decentralized packet scheduling and the interference between communi-cating nodes. In this paper, we present a state-based channel capacity perception scheme to provide sta-tistical Quality-of-Service (QoS) guarantees under a medium or high traffic load for IEEE 802.11 wire-less multi-hop networks. The proposed scheme first perceives the state of the wireless link from the MAC retransmission information and extends this information to calculate the wireless channel capaci-ty, particularly under a saturated traffic load, on the basis of the interference among flows and the link state in the wireless multi-hop networks. Finally, the adaptive optimal control algorithm allocates a net-work resource and forwards the data packet by tak-ing into consideration the channel capacity deploy-ments in multi-terminal or multi-hop mesh net-works. Extensive computer simulations demonstrate that the proposed scheme can achieve better per-formance in terms of packet delivery ratio and net-work throughput compared to the existing capacity prediction schemes.     

QoS-driven adaptive congestion control for voice over IP in multiservice wireless cellular networks

This work establishes a QoS-driven adaptive congestion control framework that provides QoS guarantees to VoIP service flows in mixed traffic scenarios for wireless cellular networks. The framework is composed of three radio resource management algorithms: admission control, packet scheduling, and load control. The proposed framework is scalable to several services and can be applied in any current or future packet-switched wireless system. By means of dynamic system-level simulations carried out in a specific case study where VoIP and Web service flows compete for shared access in an HSDPA wireless network, the proposed framework is able to increase the overall system capacity twofold depending on the traffic mix, while keeping the system operating optimally in its target QoS profile.     

A Study on Adaptive Time Token Priority-Based Queuing Scheme

Several famous priority-based queuing schemes operated in a gateway to support differentiated services among internet traffic. Examining packet forwarding operations in these queueing schemes, they only support a priority-based service either in a packet enqueuing process or in a packet dequeuing process. If a queuing scheme can support priority-based services in both packet enqueuing/dequeuing processes; it would enhance differentiated service performance for internet traffic. This study proposes a priority-based queuing scheme with an adaptive time token allotment measure to support a differentiated packet forwarding process for different types of IP traffic both in packet enqueuing/dequeuing processes. Depending on packet sizes and packet forwarding priorities of IP traffic, the proposed queuing scheme assigns fix and adaptive time token thresholds dynamically to logical queuing buffers separately. With assigned time tokens, logical queuing buffers allow arrival IP packets to be enqueued in a differentiated way. Moreover, the proposed queuing scheme uses a transferred WRR dequeuing measure to enhance a differentiated packet forwarding process. The simulation results show that the proposed queuing scheme supports a differentiated packet forwarding process for different types of IP traffic. The differentiated packet forwarding performance supported by the proposed scheme is close to the IETF DiffServ scheme; this result shows that the proposed scheme can support differentiated packet forwarding performance for different types of IP traffic with a lower operation cost.     

Domain load balancing routing for multi-gateway wireless mesh networks

As Wireless Mesh Networks (WMNs) are typically used for Internet access, most traffic is routed through the gateways which connect WMN to the wired network. As a result, the gateways tend to get congested and balancing of the traffic load of gateways is critical. In this paper, we consider applications that require continuous provision of a certain bandwidth to a server located at the wired network. If a path that satisfies the bandwidth request cannot be found, the request will be rejected, so that load imbalance will result underutilization of the network capacity. We present a novel load balancing routing algorithm for maximizing the network utilization (i.e., accommodating service requests as many as possible) for multi-gateway WMNs. In the proposed scheme, a WMN is divided into domains. Each domain is served by one gateway, so that all traffic of a domain is served by the corresponding gateway. Our scheme determines routing to balance the traffic load among domains, and then performs load balancing routing within each domain. Simulation results show that in square grid topologies, our intra-domain routing achieves near optimal performance with about 70% less overhead than the existing schemes. Our inter-domain load balancing scheme outperforms the existing heuristics by up to 25% while achieving about 80% performance of the optimal solution.     

Towards a better-than-best-effort forwarding service for multimediaflows

A salient requirement of interactive multimedia applications is that they transmit data continuously at uniform rates with minimum possible end-to-end delay. The majority of these applications do not require hard and fast guarantees of network performance, but the current best-effort forwarding model of the Internet is frequently insufficient for realizing these requirements. Worse still, the requirement of uniform-rate transmission puts many multimedia applications at odds with current and proposed Internet network management practices that assume or require TCP-like reactions to packet loss. We are investigating router-based active queue management, specifically the use of queue occupancy thresholds to isolate TCP flows and to provide a better-than-best-effort forwarding service for flows in need of uniform-rate transmissions. Our current scheme, class-based thresholds (CBT), relies on a packet marking mechanism such as those proposed for realizing differentiated services on the Internet. CBT, when combined with existing active router queue management schemes such as random early detection (RED), provides a performance for TCP that approximates that achievable under a packet scheduling scheme and acceptable performance for multimedia flows. CBT is a simple and efficient mechanism with implementation complexity and run-time overhead comparable to that of RED     

Joint routing and scheduling optimization in arbitrary ad hoc networks: Comparison of cooperative and hop-by-hop forwarding

Cooperation schemes form a key aspect of infrastructure-less wireless networks that allow nodes that cannot directly communicate to exchange information through the help of intermediate nodes. The most widely adopted approach is based on hop-by-hop forwarding at the network layer along a path to destination. Cooperative relaying brings cooperation to the physical layer in order to fully exploit wireless resources. The concept exploits channel diversity by using multiple radio units to transmit the same message. The underlying fundamentals of cooperative relaying have been quite well-studied from a transmission efficiency point of view, in particular with a single pair of source and destination. Results of its performance gain in a multi-hop networking context with multiple sources and destinations are, however, less available. In this paper, we provide an optimization approach to assess the performance gain of cooperative relaying vis-a-vis conventional multi-hop forwarding under arbitrary network topology. The approach joint optimizes packet routing and transmission scheduling, and generalizes classical optimization schemes for non-cooperative networks. We provide numerical results demonstrating that the gain of cooperative relaying in networking scenarios is in general rather small and decreases when network connectivity and the number of traffic flows increase, due to interference and resource reuse limitations. In addition to quantifying the performance gain, our approach leads to a new framework for optimizing routing and scheduling in cooperative networks under a generalized Spacial Time Division Multiple Access (STDMA) scheme.     

MPLS自适应流量工程的随机队列网络模型

负载均衡是改进Internet性能和可扩展性的关键技术,本文采用自适应流量工程方法优化网络资源利用率、最小化网络拥塞。假定Internet流量模型具有平稳随机过程特性,其统计变量具有长时间尺度下的恒定不变性,据此提出了一种基于流的多径负载均衡随机队列网络模型。基于M/Pareto/1队列模型的优化公式对减小网络拥塞、改善网络性能具有统计意义上的指导作用。     

A packet scheduling scheme for 4G wireless access systems aiming to maximize revenue for the telecom carriers

Efficient utilization of network resources is a key goal for emerging BWAS. This is a complex goal to achieve due to the heterogeneous service nature and diverse QoS requirements of various applications that BWAS support. Packet scheduling is an important activity that affects BWAS QoS outcomes. This paper proposes a new packet scheduling mechanism that improves QoS in mobile wireless networks which exploit IP as a transport technology for data transfer between BWAS base stations and mobile users at the radio transmission layer. In order to improve BWAS QoS the new packet algorithm makes changes at both the IP and the radio layers. The new algorithm exploits handoff priority scheduling principles and takes into account buffer occupancy and channel conditions. The packet scheduling mechanism also incorporates the concept of fairness. The algorithm offers an opportunity to maximize the carriers’ revenue at various traffic situations. Simulation results were compared to well-known algorithms which demonstrated the new packet scheduling algorithm is able to provide a low handoff packet drop rate, low packet forwarding rate, low packet delay, ensure fairness amongst the users of different services and generates higher revenue. Furthermore this research proposes a new and novel measure named “Satisfaction Factor” to measure the efficacy of various scheduling schemes and finally proposes four performance metrics for NodeB’s of in Next Generation Wireless Networks.     

FALB:一种FC协议链路聚合算法

链路聚合是一种通过将多条物理链路捆绑成一条逻辑链路来增加带宽的常用技术,但是在基于光纤通道（fibre channel,FC）协议的交换网络中,随着转发流量区域的不断集中,传统静态链路聚合算法存在负载均衡性差、可靠性低的问题。因此提出了一种基于FC协议的动态负载均衡（flow adaptive load balancing, FALB）链路聚合算法,可以实现对流量的动态自适应匹配以及断链可重连保护机制。结果表明,该算法机制相较于静态负载均衡的算法设计,可以将有效带宽比最终收敛到 95%以上,有效提升聚合链路的负载均衡特性和可靠性。     

A Multi-Server Scheduling Framework for Resource Allocation in Wireless Multi-Carrier Networks

Multiuser resource allocation has recently been recognized as an effective methodology for enhancing the power and spectrum efficiency in OFDM (orthogonal frequency division multiplexing) systems. It is, however, not directly applicable to current packet-switched networks because most existing packet- scheduling schemes are based on a single-server model and do not serve multiple users at the same time. In this paper, we propose a cross-layer resource allocation algorithm based on a novel multi-server scheduling framework to achieve overall high system power efficiency in packet-switched OFDM networks. Our contribution is four fold: (i) we propose and analyze a MPGPS (multi-server packetized general processor sharing) service discipline that serves multiple users at the same time and facilitates multiuser resource allocation; (ii) we present a MPGPS-based joint MAC-PHY resource allocation scheme that incorporates packet scheduling, subcarrier allocation, and power allocation in an integrated framework; (iii) by investigating the fundamental tradeoff between multiuser-diversity and queueing performance, we present an A-MPGPS (adaptive MPGPS) service discipline that strikes balance between power efficiency and queueing performance; and (iv) we extend MPGPS to an O-MPGPS (opportunistic MPGPS) service discipline to further enhance the resource utilization efficiency.     

一种适用于星地网络的流量负载均衡算法

在星地网络的背景下,为了降低由于地面网络链路负载过高导致的传播时延,提出了一种适用于星地网络的负载均衡算法。算法基于网络的链路容量设置一个链路资源利用率作为阈值,当发现路径出现过载的时候分成两步进行路径的预规划：第一步,运用蚁群算法为链路过载的数据流选择其他的路径,以满足地面网内部的负载均衡;第二步,如果仍出现链路超阈值的情况,通过卸载延迟容忍数据流来执行地面与卫星网络之间的负载均衡。仿真结果表明,设定自适应的过载阈值可以有效降低链路的拥塞,与现有算法相比该算法不仅可以更加均匀地分配地面网络中的流量,而且借助卫星网络使地面网络的平均时延降低了18.3%,提高了网络的服务质量。     

Resource prediction and management in active networks

Active networks provide a programmable user–network interface that supports dynamic modification of the network's behavior. Network nodes, in addition to forwarding packets, perform customized computation on the messages flowing through them. Resources in an active network mainly consist of CPU and bandwidth. The inherent unpredictability of processing times of active packet poses a significant challenge in CPU scheduling. It has been identified that prior estimation of the resource requirements of a packet is very difficult since it is platform dependent and also depends on processing load at the time of execution, operating system scheduling, etc. An efficient allocation is required for the optimal utilization of resources. In this paper, resources are estimated using prediction techniques such as single exponential smoothing (SES), adaptive‐response‐rate single exponential smoothing (ARRSES) and Holt's two‐parameter estimation models. The estimated results agreed most with the actual requirements. The estimation models were compared with model criteria. An algorithm was also designed for the allocation of resources. Effectiveness of the algorithm was measured through simulation and achieved almost perfect fairness for all flows and also provided much superior delay guarantees under a highly dynamic environment. Copyright © 2008 John Wiley & Sons, Ltd.     

面向异构信号处理平台的负载均衡算法

针对当前异构信号处理平台中信号处理应用的调度算法优化目标单一且调度结果中处理器负载不均衡的问题，提出了一种基于蚁群优化算法的负载均衡算法。该算法结合蚁群优化算法的快速搜索能力和组合优化能力，以信号处理应用的调度长度和处理器负载均衡为优化目标，对初始信息素矩阵和蚂蚁的遍历顺序进行改进，提出调度长度启发因子和负载均衡启发因子对处理器选择公式进行改进，利用轮盘赌策略确定信号处理应用各子任务分配的处理器，完成信号处理应用的调度。仿真结果表明，该算法得到调度结果在调度长度和负载均衡方面均有改进，可以充分发挥各处理器性能，提高异构信号处理平台的整体效率。     

Efficient policies for carrying Web traffic over flow-switchednetworks

To efficiently transfer diverse traffic over high-speed links, modern integrated networks require more efficient packet-switching techniques that can capitalize on the advances in switch hardware. Several promising approaches attempt to improve the performance by creating dedicated “shortcut” connections for long-lived traffic flows, at the expense of the network overhead for establishing and maintaining these shortcuts. The network can balance these cost-performance tradeoffs through three tunable parameters: the granularity of flow end-point addresses, the timeout for grouping related packets into flows, and the trigger for migrating a long-lived flow to a shortcut connection. Drawing on a continuous one-week trace of Internet traffic, we evaluate the processor and switch overheads for transferring HTTP server traffic through a flow-switched network. In contrast to previous work, we focus on the full probability distributions of flow sizes and cost-performance metrics to highlight the subtle influence of the HTTP protocol and user behavior on the performance of flow switching. We find that moderate levels of aggregation and triggering yield significant reductions in overhead with a negligible reduction in performance. The traffic characterization results further suggest schemes for limiting shortcut overhead by temporarily delaying the creation of shortcuts during peak load and by aggregating related packets that share a portion of their routes through the network     

An analytical model for all-optical packet switching networks with finite FDL buffers

Signal loss and noise accumulation can cause fading in optical buffers implemented by fiber delay lines (FDLs). Optical packets that excessively recirculate through FDLs are easily dropped from their routing paths. Therefore, analytical models and packet scheduling schemes require additional considerations for FDL buffers. This work proposes an analytical model for all-optical packet switching networks with finite FDL buffers and a general class of scheduling schemes including many basic scheduling schemes. We intend to minimize the packet loss probability by ranking packets to achieve an optimal balance between latency and residual lifetime in the general class of scheduling schemes. The analytical model is based on a non-homogeneous Markovian analysis to study the effects of various scheduling schemes on packet loss probability and average latency. Analytical results show how various network parameters affect the optimal balance, and illustrate how properly balancing latency and residual distance can significantly improve network performance.