Toward customer-usable all-optical networks

Green, Jr. P.E.

doi:10.1109/35.335999

Cited by 11 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is particularly true in packet switched systems. Wavelength division multiplexing (WDM) has been used to reduce the impact of the speed mismatch by partitioning the enormous bandwidth into multiple, multi-access, more manageable channels that operate at speeds compatible with the electronic interface [1,2]. This paper investigates media access protocols for WDM star-coupled hierarchical networks.…”

Section: Introductionmentioning

confidence: 99%

Media access protocols for WDM networks with on-line scheduling

Sivalingam

Wang

1996

J. Lightwave Technol.

View full text Add to dashboard Cite

This paper studies media access protocols which support collision-free broadcast/multi-cast communication for optically connected star-coupled systems with Wavelength Division Multiple Access channels. An early hybrid access protocol, consisting of reservation and of receiver channels pre-allocation, allows reservation for exactly one WDM channel per node during the reservation phase. We extend this protocol by allowing reservations on multiple channels and applying scheduling algorithms to improve network performance. Existing scheduling algorithms for similar reservation problems, although providing optimal scheduling for network utilization, have unacceptable computational cost and high implementation complexity. We propose two on-line scheduling algorithms which run in linear time and are simple, making them amenable for hardware implementation. Performance of the protocol using scheduling with varying system parameters is evaluated through discrete-event simulation under both uniform and non-uniform traffic patterns. Our simulation results show that our approach achieves lower average packet latency and higher network utilization compared to the early hybrid access protocol with single channel, especially in the client-server environment.

show abstract

Section: Introductionmentioning

confidence: 99%

Media access protocols for WDM networks with on-line scheduling

Sivalingam

Wang

1996

J. Lightwave Technol.

View full text Add to dashboard Cite

show abstract

“…System Cost: The system cost usually is assumed to be determined by the end station cost, in particular by the number of transmitters/receivers. For example, the Rainbow project at IBM supported circuit-switched traffic [5]. Unfortunately, relative cost factor is unclear at this stage.…”

Section: Introductionmentioning

confidence: 99%

<title>Channel access protocols for high-speed LANs using WDM: a comparative study</title>

Sivalingam

1995

SPIE Proceedings

View full text Add to dashboard Cite

Extensive research and considerable progress in the dense optical wavelength division multiplexing (WDM) technology has made it the most likely candidate for implementation in next-generation high peiformance switches, tANs and MANs. The first key challenge is the continuous development of costeffective optical components in order to realize practical networks. The next main challenge is to design an effective channel access protocol that will: (a) takefull advantage ofthe existing and emerging optical technologies, (b) fairly and successfully coordinate transmissions between the networks nodes, and (c) efficiently manage the enormousfiber bandwidth.There have been many access protocols proposedfor WDM networks including quite a few proposals that require many more channels than the number ofnodes. The practical device characteristics impose an upper limit on the number of available WDM channels. Ourfocus in this paper is on systems with many nodes and small number of channels. We study the use ofin-band signaling mechanism compared to the separate control channel (out-band signaling) approach. Signaling is used to achieve coordination between source and destination with respect the communication channel. The objective of this paper is to understand the key trade-offs in the protocol design for such networks through close examinations of the two approaches. Our intent is not to claim the superiority ofeither approach, rather to stimulatefurther studies on such systems. Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/25/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx SP!E Vol. 2614 1291 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/25/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

show abstract

“…It specifies the maximum number of packets per time slot that can be sent through the network in a non-blocking fashion. It is given by (3) z,,, = (Euve + l)(zH + zD + z, ) time slots (4) This equation assumes that the traffic statistics are such that packets on average will require E,,, hops to reach their destination. If the traffic statistics are such that the majority of packets require the maximum number of hops to reach their destination, then Eq.…”

Section: 3mentioning

confidence: 99%

A scalable recirculating shuffle network with deflection routing

Monacos

Sawchuk²

Proceedings of Massively Parallel Processing Using Optical Interconnections

View full text Add to dashboard Cite

We describe a high-speed optical wide area network (WAN) architecture called the multi-cylinder ShuffleNet (MCSN). Its basic topology is a recirculating shuffle exchange network with an all-optical data path from source to destination. The network capacity and performance is improved by augmenting it with multiple parallel copies of the original topology called routing cylinders. This approach avoids the use of a global request/acknowledge mechanism by matching the rate at which packets can be input into the network to the rate at which packets can be transmitted through the network (without congestion). Packets are dynamically stored within the switching nodes (using only fixed delay line elements) and the links of the network, and congestion control within the network is handled by routing blocked packets onto alternative routing cylinders. The architecture is easily scalable, designed for bit-synchronous, packet asynchronous traffic and utilizes a simple distributed control scheme. Simulation results which demonstrate these features are given. IntroductionOptical interconnection networks offer the potential for bandwidth in the tens to hundreds of gigabitslsec over a given data path. Their primary limitation is that complex logic, data buffering and routing is difficult to perform in the optical domain. Our goal is to perform these functions as fast as possible while maintaining the data in all-optical form as much as possible to avoid electronicloptical conversions.Current optical communication networks fall into several categories: i) ring or bus topologies with optical interconnects [l]-[S]; ii) circuitlpacket switched networks for packet asynchronous traffic [9]-[ 111; iii) circuit switched networks using global electronic control with optical data paths in the switching nodes and links [12], 1131; and iv) packet switched optical routing networks for packet synchronous traffic [14]-[24]. Ring or bus topologies are simple to implement and maintain data in optical form while in the network but have only one data path for all communications. Circuitlpacket switched networks with optical point-to-point links and electronic paths within the switching nodes [9], [ 101 provide multiple communication paths 0-8186-7591-8196 $5.00 0 1996 IEEE Proceedings of MPPOI '96and allow for packet asynchronous traffic, but the link data rate on each path is limited by optoelectronic conversions. Additionally, circuit switched networks are limited by the requestlacknowledge cycle time for path establishment. The packet switched network described in [ 111 routes packet asynchronous traffic with optical switching nodes but requires monitoring the status of a switching node before sending a new packet into the network. Synchronously switched optical schemes provide multiple data paths with the enhanced throughput of optics, but these networks are not suitable for distributed networks where packet synchronization is not possible without the use of variable delay electronic buffers. Current optical network implementations do not off...

show abstract

Toward customer-usable all-optical networks

Cited by 11 publications

References 11 publications

Media access protocols for WDM networks with on-line scheduling

Media access protocols for WDM networks with on-line scheduling

<title>Channel access protocols for high-speed LANs using WDM: a comparative study</title>

A scalable recirculating shuffle network with deflection routing

Contact Info

Product

Resources

About