一种新型的PPS交换机

并行分组交换机(PPS)是一种利用多个低速交换结构实现高速交换的新技术,它是当前高速交换领域的一个研究热点。该文首先分析了PPS交换机的研究现状。然后,提出了一种新型的PPS交换机结构,运用流指数控制的方法保证了信元通过PPS交换机时不会乱序。仿真分析表明该交换结构和算法能提供良好的吞吐率和平均延迟。     

Pipelined Scheduling of Functional HW/SW Modules for Platform‐Based SoC Design

We developed a pipelined scheduling technique of functional hardware and software modules for platform‐based system‐on‐a‐chip (SoC) designs. It is based on a modified list scheduling algorithm. We used the pipelined scheduling technique for a performance analysis of an MPEG4 video encoder application. Then, we applied it for architecture exploration to achieve a better performance. In our experiments, the modified SoC platform with 6 pipelines for the 32‐bit dual layer architecture shows a 118% improvement in performance compared to the given basic SoC platform with 4 pipelines for the 16‐bit single‐layer architecture.     

三级Clos网络中分布式调度算法研究

调度算法用于解决交换网络输出端口竞争问题.鉴于现有三级Clos网络调度算法存在实现复杂、性能无法保证的缺点,本文提出了一种利于分布式调度的三级Clos网络结构和相应的负载均衡调度思想,并给出了一种简单负载均衡调度算法.采用这种新结构和相应的调度思想不仅可以简化三级Clos网络的调度,而且可以充分利用现有单Crossbar网络调度的研究成果,保证算法性能.文中通过理论和仿真分析证明基于这种思想的调度算法比现有算法具有很大的优越性.     

输入队列交换的公平可扩展调度系统

提出了一种基于输入队列交换的公平可扩展网络调度系统FSSA.通过将若干个容量较小的调度器合理连接并使其协同工作,构成多端口大容量网络交换调度系统,解决了单个调度器容量和端口数受集成电路工艺限制的问题.FSSA不仅速度高、规模可扩展而且易于硬件实现.环型连接、管线工作及公平调度技术的采用使FSSA在性能方面得到了进一步优化.仿真结果显示,FSSA的性能可与基于iSLIP、DSRR等算法的单片调度器相比拟,尤其在流量较大时,FSSA的性能明显优于单调度器性能.     

Bandwidth aggregation for real-time applications in heterogeneous wireless networks

A variety of wireless interfaces are available for today's mobile user to access Internet content. When coverage areas of these different technologies overlap, a terminal equipped with multiple interfaces can use them simultaneously to improve the performance of its applications. In this paper, we motivate the advantages that can be had through simultaneous use of multiple interfaces and present a network layer architecture that enables diverse multiaccess services. In particular, we explore in depth one such service provided by the architecture: Bandwidth Aggregation (BAG) for real-time applications. An important aspect of the architecture when providing BAG services for real-time applications is the scheduling algorithm that partitions the traffic onto different interfaces such that the QoS requirements of the application are met. We propose one such algorithm Earliest Delivery Path First (EDPF), that ensures packets meet their playback deadlines by scheduling packets based on the estimated delivery time of the packets. We show through analysis that EDPF performs close to an idealized Aggregated Single Link (ASL) discipline, where the multiple interfaces are replaced by a single interface with same aggregated bandwidth. A prototype implementation and extensive simulations carried using video and delay traces show the performance improvement BAG with EDPF scheduling offers over using just the Highest Bandwidth Interface (HBI) and other scheduling approaches based on weighted round robin.     

一种基于Credit的变长分组并行交换网络调度算法

该文提出了一种新的并行分组交换(PPS)网络调度算法。该算法通过在解复用器处采用以变长分组为业务分配单元的方式消除了信元的乱序问题;通过采用Credit机制进行业务分配,实现了业务到各个交换平面完全公平的分配;各个并行交换单元采用组合输入输出排队,降低了对缓存和交换平面的加速要求,同时可以充分利用现有单Crossbar网络调度算法的研究成果。文中证明了该算法对业务分配的公平性,对高速缓存的需求量以及整个网络的稳定性,仿真进一步证明了该算法具有良好性能。     

Long-range science scheduling for the Hubble Space Telescope

Observations with NASA's Hubble Space Telescope (HST) are scheduled with the assistance of a long-range scheduling system (Spike) that was developed using artificial intelligence techniques. In earlier papers, we have described the system architecture and the constraint representation and propagation mechanisms. In this paper we describe the development of highlevel automated scheduling tools, including tools based on constraint satisfaction techniques and neural networks. The performance of these tools in scheduling HST observations is discussed.     

Control of resources in broadband networks with quality of serviceguarantees

The design principles of resource control algorithms based on asynchronous time-sharing (ATS) are addressed, together with their interaction and cooperation in a wide area network environment, and a framework for evaluating the overall performance of the system is presented. The basic concepts of the ATS framework are presented, along with an overview of an architecture for joint scheduling and admission control. Scheduling mediates the low-level competition for service between cells of different classes; admission control regulates the acceptance or blocking of incoming traffic on a call-by-call basis. The performance of the scheduling algorithms is evaluated and the interaction between scheduling and admission control is quantified. A reference model for broadband networks is presented     

A framework for reconfigurable computing: task scheduling and context management-a summary

Reconfigurable computing is consolidating itself as a real alternative to ASICs (Application Specific Integrated Circuits) and general-purpose processors. The main advantage of reconfigurable computing derives from its unique combination of broad applicability, provided by the reconfiguration capability, and achievable performance, through the potential parallelism exploitation. The key aspects of the scheduling problem in a reconfigurable architecture are discussed, focusing on a task scheduling methodology for DSP and multimedia applications, as well as the context management and scheduling optimizations.     

Performance comparison of OBS and SONET in metropolitan ring networks

This paper explores the feasibility of deploying optical burst switching (OBS) in metropolitan area networks (MANs) as an alternative to synchronous optical network (SONET), over wavelength-division multiplexing. We present a comparison between two OBS architectures (with centralized and distributed scheduling schemes), SONET, and next-generation SONET (NG-SONET), respectively. We quantify some of the performance metrics such as end-to-end delay and loss rate when supporting Ethernet traffic in metro ring networks. Our simulation results show that OBS offers significant performance improvement over SONET and NG-SONET. In general, the OBS architecture with distributed scheduling has a superior delay performance, whereas the OBS architecture with centralized scheduling has a better loss metric.     

毫米波通信低复杂度波束选择和用户调度算法

针对毫米波混合波束成形系统中用户调度方案复杂度过高的问题,提出两种低复杂度的波束选择和用户调度联合优化算法。混合波束成形架构使得用户调度问题面临着新的挑战,变成了模拟波束选择和用户调度的联合优化问题。考虑发送端无法获得完美信道状态信息的实用场景,采用基于固定码本的波束训练方案获取等效信道状态信息,引入调用指示函数将联合优化问题建模成非凸组合优化规划,分别以粒子群优化和贪婪算法为核心,提出两种低复杂度的次优解决方法。仿真结果表明,相较穷举搜索,所提算法能在性能和复杂度之间取得很好的折中。      

Variable Partitioning and Scheduling for MPSoC with Virtually Shared Scratch Pad Memory

One of the most critical components that determine the success of an MPSoC based architecture is its on-chip memory. Scratch Pad Memory (SPM) is increasingly being applied to substitute cache as the on-chip memory of embedded MPSoCs due to its superior chip area, power consumption and timing predictability. SPM can be organized as a Virtually Shared SPM (VS-SPM) architecture that takes advantage of both shared and private SPM. However, making effective use of the VS-SPM architecture strongly depends on two inter-dependent problems: variable partitioning and task scheduling. In this paper, we decouple these two problems and solve them in phase-ordered manner. We propose two variable partitioning heuristics based on an initial schedule: High Access Frequency First (HAFF) variable partitioning and Global View Prediction (GVP) variable partitioning. Then, we present a loop pipeline scheduling algorithm known as Rotation Scheduling with Variable Partitioning (RSVP) to improve overall throughput. Our experimental results obtained on MiBench show that the average performance improvements over IDAS (Integrated Data Assignment with Scheduling) are 23.74% for HAFF and 31.91% for GVP on four-core MPSoC. The average schedule length generated by RSVP is 25.96% shorter than that of list scheduling with optimal variable partition.     

Broadcast delivery

The architecture and performance of systems that use a broadcast channel to deliver information to a community of users are discussed. Information is organized into units called pages, and at any instant of time, two or more users may request the same page. Broadcast delivery is attractive for such an environment because a single transmission of a page will satisfy all pending requests for that page. Three alternative architectures for broadcast information delivery systems are considered. They are one-way broadcast, two-way interaction, and hybrid one-way broadcast/two-way interaction. An important design issue is the scheduling of page transmissions such that the user response time is minimized. For each architecture, existing scheduling algorithms are described, and their mean-response-time performance evaluated. Properties of scheduling algorithms that yield optimal mean response time are discussed. A comparative discussion of the performance differences of the three architectures is also provided     

An efficient transmission scheduling algorithm for awavelength-reusable local lightwave network

Although single-hop star networks based on wave-length division multiplexing (WDM) are attractive owing to their all-optical communication features, the throughput of such lightwave networks is limited due to the small number of available wavelengths. In this paper, a wavelength-reusable local lightwave network that consists of two interconnected WDM star networks is proposed. Based on this architecture, the lower bounds for the problems of minimizing the switching duration and the number of switching modes are derived. A transmission scheduling algorithm for this architecture to efficiently reuse the wavelengths is also proposed. The analytical result shows that the proposed scheduling algorithm always produces solutions close to the lower bounds. Simulation results show that given the same number of users and available wavelengths, the solutions (in terms of the average switching duration and the average number of switching matrices) obtained by the proposed scheduling algorithm on the interconnected WDM networks are better than the optimal solution on a single-star WDM network. In most cases, the performance improvement achieves 20 to 45%     

OBS网络的TCP性能分析与改进

光突发交换(OBS，Optical Burst Switch)是下一代IP-over-DWDM光网络最有前途的解决方案之一。Burst组装算法和数据通道调度算法是OBS网络的关键技术，TCP协议是当前主要的传输层协议。该文对OBS网络的TCP性能进行分析，并提出了基于源地址集的组装算法和ACK优先调度算法。仿真结果表明使用该算法可以显著改善OBS网络的TCP性能。     

移动网络中基于两级缓存的流媒体调度算法

提出一种适用于移动通信网的两级缓存流媒体系统结构2CMSA（two—level cache mobile streaming architecture），它突破了移动流媒体系统中终端缓存空间小、无线接入网带宽窄的局限；针对2CMSA结构设计了基于两级缓存的移动流媒体调度算法2CMSS（two—level cache based mobile streaming scheduling algorithm），建立数学模型分析了其性能；仿真实验证明，与原有的移动流媒体系统相比，使用2CMSS调度算法能够有效地节省网络传输开销，降低用户启动时延。     

A Dynamic Resource Allocation Scheme for the Downlink Heterogeneous Traffic of Internet Protocol (IP) Based OFDM networks

Radio Resource management mechanisms such as physical-centric radio resource allocation and medium access control (MAC)—centric packet scheduling are expected to play a substantial role in the performance of orthogonal frequency division multiplexing (OFDM) based wireless networks. OFDM provide fine granularity for resource allocation since they are capable of dynamically assigning sub-carriers to multiple users and adaptively allocating transmit power. The current layered networking architecture, in which each layer is designed and operated independently, results in inefficient resource use in wireless networks due to the nature of the wireless medium, such as time-varying channel fading. Thus, we need an integrated adaptive design across different layers, allowing for a cross-layer design. In this paper, a scheduling scheme is proposed to dynamically allocate resources for the downlink data transmission of internet protocol based OFDM networks. Generally to maximize the capacity and user satisfaction improvements in packet data admission, scheduling and policing are necessary. Of the three, efficient scheduling has the greatest impact on increased system capacity or effective spectrum usage. In addition, proper scheduling can greatly improve user satisfaction. The contribution of this work is twofold: first we evaluate current allocation schemes by exploiting the knowledge of channel sate information (CSI) and traffic characteristics in terms of queue state information (QSI) to acquire the system performance on a real time network. Second, the resource allocation scheme is extended by incorporating MAC layer information as well as opportunistic packet scheduling in the time-domain-based on minimum weight cost function. The key factors that affect the overall system performance in terms of system average throughput and delay are identified, evaluated and discussed.     

Efficient variable partitioning and scheduling for DSP processors with multiple memory modules

Multiple on-chip memory modules are attractive to many high-performance digital signal processing (DSP) applications. This architectural feature supports higher memory bandwidth by allowing multiple data memory accesses to be executed in parallel. However, making effective use of multiple memory modules remains difficult. The performance gain in this kind of architecture strongly depends on variable partitioning and scheduling techniques. In this paper, we propose a graph model known as the variable independence graph (VIG) and algorithms to tackle the variable partitioning problem. Our results show that VIG is more effective than interference graph for solving variable partitioning problem. Then, we present a scheduling algorithm known as the rotation scheduling with variable repartition (RSVR) to improve the schedule lengths efficiently on a multiple memory module architecture. This algorithm adjusts the variable partitions during scheduling and generates a compact schedule based on retiming and software pipelining. The experimental results show that the average improvement on schedule lengths is 44.8% by using RSVR with VIG. We also propose a design space exploration algorithm using RSVR to find the minimum number of memory modules and functional units satisfying a schedule length requirement. The algorithm produces more feasible solutions with equal or fewer number of functional units compared with the method using interference graph.     

MCBF: a high-performance scheduling algorithm for buffered crossbar switches

The buffered crossbar architecture is becoming very attractive for the design of high performance routers due the unique features it offers. Many distributed scheduling algorithms have been proposed for this architecture. Despite their distributed nature, the existing schemes require quite a bit of hardware and timing complexity. We propose a novel scheduling scheme named the most critical buffer first (MCBF). This scheme is based only on the internal buffer information and requires much less hardware than the existing schemes. Yet, it exhibits good performance and outperforms all its competitors. More interestingly, MCBF shows optimal stability performance while being almost a stateless algorithm.     

移动核心网应用NFV的关键问题探讨及实践

基于网络功能虚拟化的背景,探讨了移动核心网应用网络功能虚拟化(NFV)需要解决的3大关键技术:基础设施平台选择、核心网云管理架构设计和转发面性能加速,并基于NFV架构定义了一种移动核心网的云管理系统架构。与传统电信核心网管理相比,该系统架构增加了对硬件资源、虚拟资源层、虚拟化网元以及完整网络功能的管理、编排和调度,增加了虚拟网元管理和虚拟网元之间的接口,增强了对通用硬件运行状态的管理能力,以便能更及时、更全面地管理通用硬件和虚拟网元,实现实时的资源和功能调度。