Frame bursting: a technique for scaling CSMA/CD to gigabit speeds

Gigabit Ethernet supports the transmission of ordinary Ethernet frames at a data rate of 1000 Mb/s. Both flow-controlled full-duplex point-to-point links and half-duplex shared collision domains are included in the IEEE 802.3z draft standard. The parameters for half-duplex operation were chosen to align with the requirements of current generic building cabling standards, rather than to match the natural way that network size scales inversely with speed, so a star-wired single repeater topology with a maximum diameter of 200 m is permitted. Thus, gigabit Ethernet is the first time that the CSMA/CD medium access control algorithm has been applied to networks in which the round-trip propagation delay can be much greater than the transmission time for a minimum length frame. In this article, we describe the changes made to CSMA/CD that allow it to support large propagation delays without increasing the minimum frame length or changing its existing one-frame-at-a-time service interface. First, carrier extension is used to decouple the slot time from the minimum frame length, so the slot time can be increased without changing the Ethernet frame format. Second, frame bursting is used to reduce the overhead for transmitting small frames by allowing a host to transmit more than one frame without ever releasing control of the channel. Using simulation, we show that CSMA/CD with carrier extension and frame bursting operating on 1000 Mb/s links provides a significant performance increase over 100 Mb/s Fast Ethernet     

自适应p-持续CSMA/CD介质访问控制策略

针对IEEE802．3局域网介质随机访问协议l-持续CSMA／CD，在同一个冲突域内，它的性能在重荷载下有明显缺陷的问题，本文分析了p-持续CSMA／CD协议，提出了一种新的关于概率p的自适应算法。该算法基于对网络流量的分析，在OSI物理层通过信道信息量的在线实时检测，根据反馈控制的原理，在数据链路层动态地调整控制输入参数p，从而控制站点发送分组数据帧的行为，使信道处于最佳的传输状态，提高信道的效率。     

A coded data transceiver for the Star Local Area Network (StarLAN)

The system requirements of the physical layer of a star-connected local area network (StarLAN) are examined. The network forms a 1-Mb/s component of the IEEE 802.3 CSMA/CD (carrier-sense multiple-access with collision detection) system. The new standard allows CSMA/CD network protocols to be used over telephone twisted pairs for office or industrial digital communications. The use of this protocol means that existing local area network controllers and their higher level operations will be immediately compatible with StarLAN operation. The philosophy and functionality of the AM7961 StarLAN coded data transceiver is discussed in the context of the StarLAN environment and multipoint extension (MPE), a bus-oriented extension. MPE is a demanding requirement which needs special consideration in the collision detection system used by the chip. It is graphically shown why these requirements cannot easily be met by conventional techniques. The results on the test demonstrate that the overvoltage sensing system used in the chip overcomes the difficulties encountered     

The 802.3z Gigabit Ethernet Standard

Considers standards development in the IEEE 802.3z Gigabit Ethernet Task Force, part of the IEEE 802.3 CSMA/CD Working Group. IEEE Std 802.3z-1998 was formally approved by the IEEE Standards Board on June 25th, 1998. IEEE Std 802.3z, extends the operating speed of the world's most popular local area network to 1 billion bits per second (1000 Mb/s) for interconnecting high-performance switches, routers, and servers in the backbone of local area networks. Maintaining backward compatibility with the over-100-million-node installed base of 10 Mb/s and 100 Mb/s was a key requirement     

1-persistent DUAL channel local area network for integrated voice and data application

It is a well-known fact that the carrier sense multiple access with collission detection (CSMA/CD) protocol would suffer from throughput degradations and long packet delays at heavy loads due to excessive collisions and retransmissions, and thus is not suitable for voice/data integration. Recently, a new protocol called carrier sense multiple access with time-split collision detection and split channel reservation (CSMA-TCD-SCR) has been proposed¹ to improve the channel throughput of CSMA/CD protocol by splitting the common channel into dual channels, i.e. one for data transfer and the other for reservation. It has been shown that the CSMA-TCD-SCR protocol improves the channel throughput and stability performance significantly.¹ In this paper, we propose a 1-persistent dual-channel LAN (1P-DC-LAN) protocol with the tree algorithm for reservation, which is an extension of the CSMA-TCD-SCR protocol, for voice/data integration applications. Through approximate analytical modelling and computer simulations, it has been demonstrated that the 1P-DC-LAN protocol exhibits nearperfect scheduling throughputs for data transmission and highly bounded voice packet delays. In addition, the results were compared with IEEE 802.3 CSMA/CD and twin-channel CSMA/CD.     

A Game-Theoretic Study of CSMA/CA Under a Backoff Attack

CSMA/CA, the contention mechanism of the IEEE 802.11 DCF medium access protocol, has recently been found vulnerable to selfish backoff attacks consisting in nonstandard configuration of the constituent backoff scheme. Such attacks can greatly increase a selfish station's bandwidth share at the expense of honest stations applying a standard configuration. The paper investigates the distribution of bandwidth among anonymous network stations, some of which are selfish. A station's obtained bandwidth share is regarded as a payoff in a noncooperative CSMA/CA game. Regardless of the IEEE 802.11 parameter setting, the payoff function is found similar to a multiplayer Prisoners' Dilemma; moreover, the number (though not the identities) of selfish stations can be inferred by observation of successful transmission attempts. Further, a repeated CSMA/CA game is defined, where a station can toggle between standard and nonstandard backoff configurations with a view of maximizing a long-term utility. It is argued that a desirable station strategy should yield a fair, Pareto efficient, and subgame perfect Nash equilibrium. One such strategy, called CRISP, is described and evaluated     

CSMA/CD with Deterministic Contention Resolution

This paper presents a novel media access protocol CSMA/CD with deterministic contention resolution (DCR) for a local area network. It usually operates as CSMA/CD, but once a collision occurs it resolves the collision using a kind of implicit token passing. An analysis was conducted on DCR performance characteristics based on simulation studies and in comparison to conventional CSMA/CD and implicit token passing. It was found that they were very satisfactory in terms of throughput, message delay, and delay standard deviation, these performance characteristics make DCR attractive as a media access protocol for combined voice and data traffic. It has also been shown that this protocol assures robustness against various kinds of transmission errors and station failures. A prototype of a local area network using this protocol has been developed. It consists of a pair of optical fiber buses to which each station is attached via a pair of directional couplers.     

基于预警优先级的ZigBee传感网络MAC层退避算法

针对无线传感器网络中有预警信息的高优先级数据包需要尽快传输,且IEEE 802.15.4协议本身不支持任何优先级传输机制的情况,提出了一种基于预警优先级的非时隙CSMA/CA自适应调整阶梯退避算法,并建立了离散时间马尔可夫分析模型,比较分析了网络中不同优先级节点的信道接入概率、网络吞吐量和传输延时,仿真结果表明,本文改进的自适应调整阶梯退避机制对提高无线传感器网络中高优先级数据包的实时传输性能具有积极的作用。     

The Demand Priority MAC protocol

A new technology needs to offer more than just 100 Mb/s. To succeed in the LAN marketplace, a new LAN technology must be very cost competitive with the established LANs, such as Ethernet and Token Ring, while also providing backwards compatibility with existing network software. We describe a new 100-Mb/s LAN technology that has these characteristics This technology is being defined as an open standard within the IEEE Project 802.12 Demand Priority group. Two important objectives were established for this LAN technology: first, it should be able to use the unshielded twisted pair (UTP) wiring found in a large number of installations and, in particular, to use the same wiring as defined for use in 1OBase-T. This objective was later extended to encompass support for the shielded twisted pair (STP) used for IEEE 802.5. This will enable the majority of current LAN users to benefit from their enormous investment in cable plant. The second objective was that the network should support new applications, such as video conferencing and remote training, while also providing backwards compatibility with the massive installed software base. Both objectives have been met. The Demand Priority MAC protocol currently being standardized in IEEE 802.12, offers substantial benefits over the CSMA/CD protocol of IEEE 802.3. By preserving both the current wiring infrastructure and investment in software, and by using the very simple Demand Priority MAC protocol, 100 Mb/s LANs could soon be as low-cost as 10Base-T is today     

Setting target rotation times in an IEEE token bus network

The IEEE standard 802.4 token bus protocol requires each network station to implement a synchronous (highest priority) message class, and permits a station to implement three lower priority classes: urgent asynchronous, normal asynchronous, and time available. Each of the lower three priorities (called access classes) is assigned a target token rotation time that limits the amount of time that a station can use to service lower priority traffic. A formulation of the problem is presented in which messages are transmitted from an access class as long as network throughput remains below a user-specified threshold. Formulas are derived that transform this priority scheme, based on network throughput limits, into the proper target rotation time settings that the token bus protocol actually requires. The analytical model is compared with a computer simulation of the token bus protocol and shows close agreement     

The IEEE 802.11 Distributed Coordination Function in Small-Scale Ad-Hoc Wireless LANs

The IEEE 802.11 standards for wireless local area networks define how the stations of an ad-hoc wireless network coordinate in order to share the medium efficiently. This work investigates the performance of such a network by considering the two different access mechanisms proposed in these standards. The IEEE 802.11 access mechanisms are based on the carrier sense multiple access with collision avoidance (CSMA/CA) protocol using a binary slotted exponential backoff mechanism. The basic CSMA/CA mechanism uses an acknowledgment message at the end of each transmitted packet, whereas the request to send/clear to send (RTS/CTS) CSMA/CA mechanism also uses a RTS/CTS message exchange before transmitting a packet. In this work, we analyze these two access mechanisms in terms of throughput and delay. Extensive numerical results are presented to highlight the characteristics of each access mechanism and to define the dependence of each mechanism on the backoff procedure parameters.     

A new adaptive receiver-initiated scheme for mitigating starvation in wireless networks

Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) has been adopted by the IEEE 802.11 standard and provides good performance when all transmitters are within the range of each other. Unfortunately, in multi-hop topologies, the asymmetric view of the channel state leads to a throughput distribution where a few flows may capture all the available bandwidth while many other flows get very low throughput and sometime meet starvation. To address this problem, in this paper we describe a solution called Carrier Sense Multiple Access with Collision Avoidance by Receiver Detection (CSMA/CARD) which makes use of collisions sensed by a receiver at the physical layer to help the handshake mechanism and mitigate the effect of such problem. More specifically, we propose a mechanism based on historical observations, where collisions can be used by the receiver to predict whether some sender attempted to initiate a transmission. The receiver then reacts accordingly by participating itself in a handshake sequence. We show some interesting results, obtained through analysis and simulations, when the CSMA/CARD is compared to the IEEE 802.11 protocol.     

A multistage hierarchical distributed arbitration technique for priority-based real-time communication systems

ISO 11898 is a communication protocol based on the carrier sense multiple access with collision detection and arbitration on message priority (CSMA/CD+AMP) technique, which at present is largely used as a real-time network for industrial environments. Unfortunately, because of the peculiarities of the arbitration technique it adopts, it suffers from severe limitations on the maximum extension of the network, which cannot be overcome simply by means of improvements in the transceiver's technology as they depend on the limited propagation speed of the signals on the communication support. In this paper, a new kind of network is presented that features a behavior very similar to ISO 11898, but which achieves noticeably larger areas to be covered without having to reduce the bit rate. It relies on a tree topology and adopts a brand new multistage hierarchical distributed arbitration technique, which takes the increased propagation delays into account properly.     

The CSMA/DM LAN protocol

We propose an extension of the classical CSMA/CD protocol that eliminates its three main drawbacks. The new protocol, called dual‐mode CSMA/CD (CSMA/DM), operates in two modes, light mode and heavy mode. The light mode of CSMA/DM is almost the same as the original CSMA/CD protocol and is primarily used when the LAN load is light. The heavy mode is a collision free mode and is applied when the LAN load is heavy. The proposed modification to CSMA/CD is minimal while performance gain is significant. CSMA/DM automatically switches between its two modes based on the observed LAN load. Under heavy mode, the monitor station in a CSMA/DM LAN allocates bandwidth to those busy stations without collisions. The newly added priority scheme in the heavy mode allows stations to reserve bandwidth for their high priority frames. Compared with the current collision‐free fast Ethernet and Gigabit Ethernet, CSMA/DM has the advantage of not needing any extra switches or hubs while maintaining comparable performance. Its priority scheme provides more flexibility on bandwidth distribution than in fast Ethernet and Gigabit Ethernet. It can also be easily adapted for high‐speed wireless LANs. More importantly, CSMA/DM should be a good alternative of the widely used CSMA/CA (collision avoidance) in mobile ad hoc networks (MANET) and sensor networks. Copyright © 2005 John Wiley & Sons, Ltd.     

Fiber optic LANs for the manufacturing environment

The physical layer requirements for local area networks (LANs) that satisfy the set of environmental requirements unique to factory automation applications are discussed. In many cases these requirements are best met with fiber-optic LANs compatible with the IEEE 802.3 (CSMA/CD) or 802.4 (token bus) network standards. Implementation of a passive fiber-optic star-coupler-based LAN is presented in terms of two simple fiber-optic system design rules. A number of applications of the passive star-coupler-technology LAN are briefly described to highlight the benefits of fiber optics as a data communication medium     

A dynamic ID management protocol for CSMA/IC in ad hoc networks

Contention based MAC protocols are widely used in ad hoc networks because they are suitable, where no central control node exists. However, contention based MAC protocols waste much time because of frequent collisions and long contention times. Moreover, it is hard for them to fairly distribute medium access opportunities. As a result, the problem of unfair medium access may arise under normal network conditions. Recently, another contention based MAC protocol, named the Carrier Sense Multiple Access/ID Countdown (CSMA/IC) was proposed. CSMA/IC resolves medium access contention by comparing the IDs of contending nodes with a simple signaling process. Therefore, medium access collisions never happen as long as each node possesses a unique ID, and the time cost for contention may be smaller than any other contention based MAC protocols if the number of IDs is managed so as to be as small as possible. Furthermore, CSMA/IC may support fair medium access by manipulating the ID of each node properly. In this paper, we propose a novel dynamic ID management protocol which enables a node to acquire a unique ID without any message exchanges and fairly distributed the number of medium access opportunities to all contending nodes. The proposed protocol also makes the contention process of CSMA/IC efficient by dynamically managing the length of the ID field according to the network traffic. The simulation results show that the proposed ID management protocol significantly improves the aforementioned aspects of CSMA/IC MAC protocol compared to previous ID management schemes.     

AFDX的分布式仿真

航空电子全双工交换式以太网(AFDX,Avionics Full Duplex Switched Ethernet)技术基于IEEE802.3以太网和TCP/IP通用原理,对于商用货架(COTS)网络化技术如何应用于未来飞机数据网络给出了定义[1],已经成为了一种新的航空电子数据网络标准。由于AFDX协议以及航空总线布局的复杂性,采用网络仿真技术可以大大缩短网络建设周期,减小网络建设的投资风险。为了模拟AFDX的网络性能,尤其是其在数据通信方面的高可靠性和高实时性,根据AFDX协议及其传输和交换原理,建立了基于高级仿真体系结构(HLA)的AFDX分布式仿真体系,并对各子系统的构成以及仿真执行流程进行了研究。通过对仿真结果的分析,验证了仿真模型的正确性,并表明仿真设计满足系统性能要求。     

基于FMC标准的万兆以太网卡

结合FMC标准与IEEE 802.3ae协议标准,设计了一种基于FPGA夹层卡(FPGA Mezzanine Card,FMC)标准的万兆以太网卡,满足现代工业大数据量传输应用的要求。FMC标准接口可实现多通道高速接口,解决应用母板与网卡之间的数据传输瓶颈。使用Verilog硬件描述语言设计了地址解析协议(Address Resolution Protocol,ARP)与用户数据报协议(User Datagram Protocol,UDP)的硬件协议栈,实现开放式系统(Open System Interconnect,OSI)模型的传输层;Xilinx 10 G Ethernet subsystem IP与小型化可热插拔(Small Form-factor Pluggable,SFP)光接口实现OSI模型的网络层、数据链路层、物理层。通过FMC母板、万兆以太网卡、PC上位机组建的网络成功测试了万兆以太网通信。     

Prioritized-virtual-time CSMA: head-of-the-line priority classeswithout added overhead

The prioritized-virtual-time carrier sense multiple access (PVT-CSMA) protocol is a media access algorithm for cable- and radio-based local area networks that incorporates message-based priority classes. PVT-CSMA implements the head-of-the-line (HOL) priority queuing discipline, in which higher priority messages are always transmitted in preference to lower priority messages, and messages are transmitted first-come-first-served within each priority class. Expressions are obtained for both throughput as a function of offered load and mean delay as a function of throughput for asynchronous (unslotted) PVT-CSMA, which are easily solvable for any number of priority classes and for class-dependent message lengths, arrival rates, retransmission rates, etc. Specialization of the delay analysis to the single class case yields a substantial improvement over previous results for asynchronous virtual-time CSMA. Comparisons are made between this analysis and simulations of PVT-CSMA, and with simulations of Tobagi's P-CSMA (1982) protocol     

A carrier sense multiple access protocol with power backoff (CSMA/PB)

In this paper, we exploit space as a new dimension in collision resolution for a carrier sense multiple access (CSMA) protocol. Most contention-based medium access control protocols resolve collisions by backing off in time. We introduce power backoff (PB), the use of transmission power control to resolve collisions by backing off in space, and incorporate it into a CSMA protocol as CSMA/PB. Through analysis and simulation, we show that collision resolution using power backoff can be remarkably successful. Simulation results show that CSMA/PB outperforms IEEE 802.11 in both static and mobile ad hoc network scenarios. CSMA/PB improves end-to-end throughput and uses less energy. The resulting gains in throughput per unit energy can be substantial.