Routing in the Manhattan Street Network

Maxemchuk, Nicholas F.

doi:10.1109/tcom.1987.1096802

Cited by 330 publications

(97 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In two-connected networks, for instance, each node has only two optical inputs and two optical outputs, and the routing and switching problem is thus reduced to a binary decision on the state of the switch. Very simple minimum-distance routing algorithms have been found for networks with regular topology [3,4].…”

Section: All-optical Network and Ultrahigh Bit Ratesmentioning

confidence: 99%

Soliton ultrafast all-optical mesh networks

Bononi

Forghieri

Prucnal

1993

IEE Proc. J Optoelectron. UK

View full text Add to dashboard Cite

Channel transmission error arguments show how the size of an all-optical multihop network employing deflection routing is limited for a given optical bit rate. These limits are quantified for nonregenerative all-optical mesh networks such as Manhattan Street network and ShumeNet employing solitons. It is found that the node-to-node fibre span cannot exceed a few kilometres for network sizes up to 400 nodes when the optical bit rate is as high as 100 Gbit/s if the packet error rate is to be bounded below 1 Introduction All-optical networks and ultrahigh bit ratesThe ever-growing need for faster communication speed is leading towards all-optical packet-switching fibre networks, as optical components progressively replace their slower electronic counterparts. As a first step towards higher transmission rates, optical fibres replaced copper wires as network links between nodes. The switching nodes, however, remained electronic, and the optical signal had to be down-converted to electronics for switching, buffering and routing, and then remodulated onto an optical carrier for retransmission. As a next step, in all-optical networks even the switching process will be completely optical. This will allow even higher transmission rates, since the switching can be done much faster than allowed by standard electronic switches, and more flexibility, as the switch is completely transparent to the optical packet, which can thus support any bit rate in the payload. Only the header will need to have a fixed rate since it is the only part of the packet that needs to be read at every intermediate node to correctly route the packet.Multihop mesh networks [l, 23 are attractive in this perspective since they break down the computational complexity of the control of the switching process, which increases exponentially with the number of users, by evenly distributing it among all nodes. In two-connected networks, for instance, each node has only two optical inputs and two optical outputs, and the routing and This proved to be even more attractive in all-optical networks, where one of the technological limitations is the lack of very fast access, flexible, simple optical memories. The limited-time buffering strategy of deflection routing is perfectly suited to be implemented with simple recirculating fibre delay loops, with no need of optical amplification in the loop. Both multihop distributed networks (as opposed to single-hop, centralised networks) and deflection routing (as opposed to store-and-forward (S&F) routing) are solutions which trade efficiency for a much simpler hardware implementation. If ultrahigh bit rates can be sustained by these distributed structures using deflection routing, a net gain in throughput can effectively be achieved.Generation, multiplexing and demultiplexing of optical packets at ultrahigh bit rates is made possible by recent technological breakthoughs in nonlinear optics. Mode-locked semiconductor lasers in the 1.55 pm wavelength region are today capable of producing trains of pulses of widt...

show abstract

Section: All-optical Network and Ultrahigh Bit Ratesmentioning

confidence: 99%

Soliton ultrafast all-optical mesh networks

Bononi

Forghieri

Prucnal

1993

IEE Proc. J Optoelectron. UK

View full text Add to dashboard Cite

show abstract

“…Next, extremely high aggregate network bandwidth can be achieved due to more links used in network spans. A remarkable network presented in the literature is a grid network, e.g., the Manhattan Street Network (MS_Net) [11,12]. It has high connectivity, but su ers from a variety of problems [5,13].…”

Section: Introductionmentioning

confidence: 99%

Design of a large-scale Gbit/s MAN using a cyclic reservation-based MAC protocol

Wang¹,

Hwang

Wang

2000

Journal of Systems Architecture

View full text Add to dashboard Cite

In this paper, a large-scale Gbit/s metropolitan area network (MAN) based on hierarchical ring topologies has been investigated. The network is constituted by backbone and local rings, which are connected by bridges. In the network, tra c congestion may occur in bridges due to the mismatch of transmission speed between backbone and local rings. To cope with the issue, a scheme based on a cyclic reservation-based access method is proposed for the medium access control (MAC) protocol of the network. With the scheme, the protocol can both achieve the fair access of network bandwidth and resolve the tra c congestion in bridges. Particularly, this approach reduces the implementation complexity and cost and enhances e ciency of the global network. Finally, several simulative experiments are performed, and some optimistic results are revealed. Ó 2000 Elsevier Science B.V. All rights reserved.

show abstract

“…A routing scheme, called deflection routing, has been investigated by Greenberg and Goodman [GrG86] for mesh networks (numerically), Greenberg and Hajek [GrH90] for hypercubes (analytically), Maxemchuk [Max87], [Max89] for the Manhattan network and the Shuffle Exchange (numerically and through simulations), and Varvarigos [Var90] for hypercubes (for a priority deflection scheme, through the formulation of a Markov chain). The analysis in these works was either approximate, or it involved simulations.…”

Section: Introductionmentioning

confidence: 99%

Performance of hypercube routing schemes with or without buffering

Varvarigos

Bertsekas

1994

IEEE/ACM Trans. Networking

View full text Add to dashboard Cite

We consider two different hypercube routing schemes, which we call the simple and the proerity scheroew We evaluate the throughput of both the unbuffered and the buffered version of these Lchems fosr 4rrdom multiple node to node communications. The results obtained are approximate, but very accurate as simulation results indicate, and are given in particularly interesting forms. We find that little buffer space (between one and three packets per link) is necessary and adequate to achieve throughput close to that of the infinite buffer case. We also consider two deflection routing schemes, called the simple deflection and the priority deflection schemes. We evaluate their throughput using simulations, and compare it to that of the priority scheme.

show abstract

Routing in the Manhattan Street Network

Cited by 330 publications

References 8 publications

Soliton ultrafast all-optical mesh networks

Soliton ultrafast all-optical mesh networks

Design of a large-scale Gbit/s MAN using a cyclic reservation-based MAC protocol

Performance of hypercube routing schemes with or without buffering

Contact Info

Product

Resources

About