An Energy and Performance Exploration of Network-on-Chip Architectures

In this paper, we explore the designs of a circuit-switched router, a wormhole router, a quality-of-service (QoS) supporting virtual channel router and a speculative virtual channel router and accurately evaluate the energy-performance tradeoffs they offer. Power results from the designs placed and routed in a 90-nm CMOS process show that all the architectures dissipate significant idle state power. The additional energy required to route a packet through the router is then shown to be dominated by the data path. This leads to the key result that, if this trend continues, the use of more elaborate control can be justified and will not be immediately limited by the energy budget. A performance analysis also shows that dynamic resource allocation leads to the lowest network latencies, while static allocation may be used to meet QoS goals. Combining the power and performance figures then allows an energy-latency product to be calculated to judge the efficiency of each of the networks. The speculative virtual channel router was shown to have a very similar efficiency to the wormhole router, while providing a better performance, supporting its use for general purpose designs. Finally, area metrics are also presented to allow a comparison of implementation costs.      

WDM光网络非相干串扰实验研究

在WDM光网络系统中,节点如OXC和OADM采用的光器件其非理想性积累构成系统串扰问题。本文提出了WDM光网络系统中串扰问题分析的一般数学模型,根据模型分析比较了WDM光网络中相干串扰和非相干串扰的区别。实验研究了不同非相干串扰代价条件下,对应WDM光网络系统的功率代价。对应1dB功率代价下,实验得出WDM光网络对应非相干串扰代价容限小于－20.8dB的结果。实验结果对于WDM光网络组网规划有一定的参考意义。     

FBG-based bidirectional optical cross connects for bidirectional WDM ring networks

This paper proposes a new bidirectional optical cross connect (BOXC) using fiber Bragg gratings (FBGs) and optical circulators for bidirectional wavelength-division-multiplexing ring networks. Dynamic and independent wavelength routing is achieved by employing cascaded tunable FBGs. The proposed BOXC requires a small tuning range equal to wavelength channel spacing for tunable FBGs. Therefore, the wavelength count of the proposed BOXC could be increased easily without imposing constraint on the tuning range of FBGs. Coherent and incoherent crosstalk arising in the proposed BOXC and their impact on the power penalty are studied in detail. The Monte Carlo simulation is used to characterize the probability distribution of the power penalty due to both coherent and incoherent crosstalk under various conditions. The specification requirements are then obtained for the FBGs used in BOXC nodes with a different number of wavelengths. The study in this paper reveals that the requirement for the isolation of FBGs is more stringent than that for the reflectivity of FBGs at the Bragg wavelength. It is also shown that a large BOXC can be built based on the Benes network structure, and a method to reduce the complexity is also presented.     

Influence of Crosstalk on Signal Power Based on Arrayed-waveguide Grating as N × N Optical Router

The fact that the signal results in signal-crosstalk is confirmed for arrayed-waveguide grating as N × N optical router,and the relation between the crosstalk and power penalty is obtained.The method reveals the random distributions of optical path phase errors in two multiplexers with channel numbers of 10 and 160. It is shown that the crosstalk must be less than -28 dB for a power penalty below 1 dB at a bit error rate of 1×10 -9 .It is found that when N =100,crosstalk power value is -20 dB with compensation power of 2-3 dB,so the compensation power is not ignored.     

Channel‐switching and power control mechanisms for improving network connectivity in wireless mesh networks

A Wireless Mesh Network (WMN) consists of fixed wireless routers, each of which provides service for mobile clients within its coverage area and inter‐connects mesh routers to form a connected mesh backbone. Wireless mesh routers are assigned with a channel or a code to prevent collisions in transmission. With a power control mechanism, each router could be assigned with a power level to control connectivity, interference, spectrum spatial reuse, and topology. Assigning high transmitting power level to a router can enhance the network connectivity but may increase the number of neighbors and worsen the collision problem. How to assign an appropriate power level to each router to improve the network connectivity with a constraint of limited channels is one of the most important issues in WMNs. Given a network topology and a set of channels that has been assigned to mesh routers, the proposed channel‐switching mechanism further reassigns each router with a power level and switches channels of routers to optimize both power efficiency and connectivity. A matrix‐based presentation and operations are proposed to respectively identify and resolve the channel switching problems. Simulation study reveals that the proposed mechanisms increase network throughput and provides a variety of route selection, and thus improves the performance of a given WMN. Copyright © 2009 John Wiley & Sons, Ltd.     

Bandwidth-efficient modular wavelength router for large-scale multihop virtual optical ring networks

A novel wavelength router architecture is proposed to implement large-scale multihop-based virtual optical ring (VOR) networks over a star-configured optical cable plant. A number of cyclic N/spl times/N arrayed waveguide gratings (AWGs) each having an optical loop-back for its own VOR are interconnected together for a two-level VOR network. The proposed router reduces the maximum number of hops and enables the use of wide channel spacing. The design principle is demonstrated through a proof-of-concept experiment and useful features are discussed.     

Design of free space WDM router based on holographic concavegrating

A design for a 91×91 router with 0.33 nm channel spacing using a stationary anastigmatic Littrow mount of a holographic concave grating is reported. A simplified method of concave grating analysis is developed for the evaluation of the wavelength router performance. The coupling losses for all routed channels are within -0.9 dB and the power penalty due to coherent crosstalk caused by aberrations and diffraction at the aperture is negligibly small     

一种动态分配输入队列的片上虫孔路由器结构

为了降低片上网络（NoC）由于虫孔缓冲结构排头（HoL）阻塞导致的性能损失,同时消除虚通道缓冲结构对可变长度报文表现出的缓冲区低利用率现象,本文采用虚拟通道技术提出一种动态分配输入队列（DAIQ）的片上虫孔路由器结构.该结构采用一种令牌表的方式支持虚拟队列深度与数量的动态分配,同时为了支持同一报文微片能够连续调度,本文还提出一种新颖的开关分配机制——SRRM,该机制在高负载下进一步改善了开关的延迟与吞吐率.仿真与综合的结果表明,相比传统虚通道流控的片上路由器结构,DAIQ路由器以50%的缓冲面积获得类似的性能,在0.13微米CMOS工艺下节约了30.18%的标准单元面积与384%的功耗.     

Real-time wavelength routing based on Bragg reflection with integrated forward Raman for long-reach networks

We experimentally demonstrate the all-optical efficient technique for dynamic full-duplex wavelength routing and reach extension through adoption of passive fibre Bragg grating(FBG),energy efficient vertical cavity surface emitting lasers(VCSELs)and high gain forward Raman pump.In this study,two VCSEL channels at 1549.62 nm and 1550.01 nm are directly modulated with a 8.5 Gbit/s data signal each,and transmitted over 25.5 km fibre to a passive single mode FBG-based wavelength routing node.The precise wavelength selectivity of the FBG is exploited to achieve all-optical real-time wavelength routing of the 1549.62 nm VCSEL channel with a maximum insertion loss of 22.1 dB.The 1550.01 nm channel is transmitted through the FBG with a maximum penalty of 1.03 dB.By utilizing the 7.2 dB flat gain of a forward Raman pump,the routed 1549.62 nm VCSEL channel is transmitted over 76.7 km fibre with a maximum penalty of 3.84 dB at bit error rate(BER)of 5-5.Our technique is all-optical,power efficient as it employs passive FBG and circulators and has low cross talk allied to precise wavelength selectivity of the technique.     

一种基于串联双环谐振器的4×4无阻塞波长选择路由器

设计了一种基于串联双环谐振器的4×4无阻塞波长选择路 由器。构成该路由器的串联双环谐振器具有良好的箱型滤波曲线,SOI脊波导截面尺寸为波 导宽度450 nm, 高度220 nm,脊高60 nm,微环半径为10 μm,实验可得其3dB带宽为0.3 nm,消光比为 35 dB,带内功率起 伏只有0.1 dB,为实现该路由器的功能提供了很好的实验基础。该波 长选择路由器有九种非阻塞线路,能够 实现四种广播模式路由,其可作为基本光学交换单元,级联多个路由器来构建更强大的片上 光网络。     

Bidirectional optical cross connects for multiwavelength ringnetworks using single arrayed waveguide grating router

Bidirectional optical cross connects (BOXCs) using a single arrayed waveguide grating router and tunable fiber Bragg gratings (FBGs) have been proposed for multiwavelength bidirectional WDM ring networks. Two types of structures in fold-back and loop-back configurations have been considered and their feasibility has been demonstrated experimentally. The performances of two proposed structures have been investigated and compared in terms of crosstalk characteristics and backscattering suppression capability. The performance of the fold-back configuration is less vulnerable to the crosstalk of the arrayed waveguide grating (AWG) router. On the contrary, the BOXCs in loop-back configuration suppress backscattered signals effectively without using optical bandpass filters and, thus, have good scalability and are more cost effective. Therefore, the choice between two structures will depend on the crosstalk requirement in the BOXCs and the characteristics of the AWG router. In transmission of 2.5-Gb/s signals, the power penalty of 0.5 dB has been observed at BER of 10^-9 in both structures. Expansion of the proposed 2×2 structure to N × N adopting multistage structure has been discussed. The limitation on the maximum scale due to spectral bandwidth narrowing has been considered     

DWDM设备中波长路由器网络控制管理系统的实现

介绍了DWDM设备的主要功能,波长路由器的基本功能,以及波长路由器的网络控制管理系统的实现.在系统测试中,波长路由器成功实现了其波长路由的功能,能够接收系统西向光信号,对信道波长进行功率监测,将OSC信道波长下载到本地处理后重新上传,发送东向光信号.同时验证了DNE(串口转换器)、单片机组成的代理端系统的可行性.     

CPCA: An efficient wireless routing algorithm in WiNoC for cross path congestion awareness

Wireless network-on-chip (WiNoC) is a new paradigm to mitigate the long-distance transmission latency for conventional wired network-on-chip. The wireless routers in WiNoC have to handle a large number of packets which could cause data congestion, thus reducing the network performance. In this paper, we propose a novel wireless routing algorithm, called CPCA, which exploits the cross path congestion information as hints to route the packets. Under CPCA, the whole network is partitioned into sub-networks. In each subnet, the congestion information of the wireless router is propagated along the cross path. As a result, the routers in the same dimension can get the congestion degree of wireless router within the subnet. Based on the congestion information, CPCA can compute the suitable path for packets routing, which can prominently avoid the congestion aggravation in the wireless router. Experimental results show that our proposed method can effectively improve performance in terms of packets transmission latency and network throughput.     

基于光纤光栅和光环形器的DWDM环网性能分析

对含有光纤光栅和光环形器的DWDM（密集波分复用）环网存在的串扰进行了分析,得到了评价串扰性能的指标（功率恶化、网络吞吐量）的数学模型,并将此模型图形化,联系图形得出了功率恶化、吞吐量与节点个数、节点损耗、串扰系数、发射功率、比特率等因素的关系。得出的结论对于DWDM环网的网络构建和性能改良有着重要的现实指导意义。     

一种故障通道隔离的低开销容错路由器设计

片上网络中路由器发生故障势必会影响整个网络的性能,过大的容错开销也会给网络带来很大的负担.对此,本文提出了一种故障通道隔离的低开销容错路由器架构,该路由器通过减少不必要的交叉开关及合理优化各个端口VC的数目来减小路由器整体开销,同时增加一个冗余通道来达到对路由器容错的目的.当路由器中某个通道发生故障时,通道隔离检测方法使路由器能够在检测故障类型的同时进行数据传输,带回收指针的重传buffer将会进一步减少整个容错结构的开销.实验结果表明在无故障情况下本文设计的路由器较传统路由器平均延时降低45%左右,最大吞吐率提高28%左右,面积开销仅仅增加了18.24%.在故障存在的情况下,本文方案也显现出很大的优越性,能够达到很好的容错效果.     

All-Optical Wavelength-Bypassing Ring Communications Networks

We introduce an all-optical WDM packet communication network that performs wavelength bypassing at the routers. Packets that arrive at a wavelength (optical cross-connect) router at designated wavelengths are switched by the router without having their headers examined. Thus, the processing element of the router is bypassed by such packets. For packet traffic that uses wavelengths that do not bypass a switch, the headers of such packets are examined to determine if this switch is the destination for the flow. If latter is the case, the packet is removed. Otherwise, the packet is switched to a pre-determined output without incurring (network internal) queueing delays. We study a ring network with routers that employ such a WDM bypassing scheme. We present methods to construct wavelength graphs that define the bypassing pattern employed by the routers to guide the traffic flows distributed at each given wavelength. Performance is measured in terms of the network throughput and the average processing path length (i.e., the average number of switches not being bypassed). For a fixed total processing capacity, we show that a WDM bypassing ring network provides a higher throughput level than that exhibited by a non-bypassing ring network, using the same value of total link capacity. By using WDM bypassing, the average processing path length (and thus the packet latency) is reduced. We study a multitude of network loading configurations, corresponding to distinct traffic matrices and client-server scenarios. Higher throughput levels are obtained for network configurations driven by non-uniform traffic matrices. The demonstrated advantages of WDM bypassing methods shown here for WDM ring networks are also applicable to more general network topological layouts.     

Low-power low-area network-on-chip architecture using adaptive electronic link buffers

In the deep sub-micron regime, the performance of network-on-chip (NoC) architectures is bound by the limited power and area budget. Proposed is a low-power low-area NoC architecture using a novel power-efficient control circuit that enables repeaters along the inter-router links to function as adaptive link buffers, thereby reducing the number of buffers required in the router. Simulation results in the 90 nm technology show power savings of nearly 45% and area savings of 50% for the proposed technique.     

Theoretical simulation and experimental investigation on a WDMsurvivable unidirectional open ring network using tunable channelselecting receivers

In this paper theoretical and experimental investigations on a wavelength-division-multiplexed (WDM) reconfigurable open ring network are reported. The theoretical study is focused on network component characteristics and ring network structure. These aspects will form the basis of numerical modeling. A powerful computer aided design software “COMSIS” is used for the simulation. The simulation takes into account the most important parameters: node structure, link losses, EDFA characteristics, optical powers of the channels, and signal wavelength arrangements. The power of a new channel added at each node and its wavelength with respect to those present before the node are two important parameters in the performance analysis of this open ring network. If the performance criterion is to receive the channels with high signal to noise ratios (SNR) and a narrow spread ΔSNR, the optimal length of this ring network can reach 366 km for 4 nodes and 565 km for 8 nodes. The experimental demonstrator is composed of 4 secondary nodes incorporating tunable channel selecting receivers (TCSR's) and erbium doped fiber amplifiers (EDFA's). The reconfiguration and the protection of the ring is computer controlled. Experimental results of a fully connected testbed demonstrator are also presented     

Cross-Layer Fair Bandwidth Sharing for Multi-Channel Wireless Mesh Networks

In a wireless mesh network (WMN) with a number of stationary wireless routers, the aggregate capacity can be increased when each router is equipped with multiple network interface cards (NICs) and each NIC is assigned to a distinct orthogonal frequency channel. In this paper, given the logical topology of the network, we mathematically formulate a crosslayer fair bandwidth sharing problem as a non-linear mixedinteger network utility maximization problem. An optimal joint design, based on exact binary linearization techniques, is proposed which leads to a global maximum. A near-optimal joint design, based on approximate dual decomposition techniques, is also proposed which is practical for deployment. Performance is assessed through several numerical examples in terms of network utility, aggregate network throughput, and fairness index. Results show that our proposed designs can lead to multi-channelWMNs which are more efficient and fair compared to their singlechannel counterparts. The performance gain on both efficiency and fairness increase as the number of available NICs per router or the number of available frequency channels increases.     

Interference and traffic aware channel assignment in WiFi-based wireless mesh networks

Wireless mesh networks (WMN) typically employ mesh routers that are equipped with multiple radio interfaces to improve network capacity. The key aspect is to cleverly assign different channels (i.e., frequency bands) to each radio interface to form a WMN with minimum interference. The channel assignment must obey the constraints that the number of different channels assigned to a mesh router is at most the number of interfaces on the router, and the resultant mesh network is connected. This problem is known to be NP-hard. In this paper we propose a hybrid, interference and traffic aware channel assignment (ITACA) scheme that achieves good multi-hop path performance between every node and the designated gateway nodes in a multi-radio WMN network. ITACA addresses the scalability issue by routing traffic over low-interference, high-capacity links and by assigning operating channels in such a way to reduce both intra-flow and inter-flow interference. The proposed solution has been evaluated by means of both simulations and by implementing it over a real-world WMN testbed. Results demonstrate the validity of the proposed approach with performance increase as high as 111%.