A survey of research and standards in high‐speed networks

Abstract

This paper surveys research and standards in high-speed networks. It first classifies high-speed networks according to the area of coverage. Secondly, it discusses the switches. Then, it examines network management protocols: media access control, routeing, flow control and error control. Finally, it investigates current hig-speed network standards: DQDB, FDDI, ATM, and SONET.(d) low error rate (e) abundant bandwidth Owing to the above factors, the architecture and protocols of high-speed networks have the following features:1. Simple algorithms. Since the time available for processing the packets is limited, the network protocols should be simplified as much as possible so as to reduce processing overhead, and the frequency of executing the protocols. . End-to-end protocols.As the number of users in the network is increased by orders of magnitude, it is desirable to shift the processing from the internal nodes of the network to the edge nodes (the end nodes for information transport). . Regular topologies.Regular network topologies simplify routeing, thus reducing network processing. Some regular topologies provide a number of alternative paths between the source and the destination, rendering the network less susceptible to failures. Limiting buffering.In high-speed networks, a large number of packets will flow through the transmission links. Thus, it is impractical to provide enough buffers to store all of them. In addition, current technology requires an optical-to-electrical conversion for buffering packets, which further increases processing overhead. Simplified error control.The error rate of a fibre-optical system is extremely low. This means that end-to-end, forward error control will be preferred over hop-by-hop, feedback control protocols. . Trade-offs between bandwidth and delay.The bandwidth in a high-speed network is not a scarce resource. This creates the opportunity to use some algorithms which squander bandwidth to achieve better performance. . Vulnerability to failures.The large bandwidth of high-speed networks allows many communication connections on a single fibre. Any failure in network hardware or software may cause chaos. Therefore, incorporating reliability features in high-speed networks is crucial. NETWORK ARCHITECTURENetworks may be categorized by their geographical coverage. We distinguish between local area networks (LAN), metropolitan area networks (MAN) and wide area networks (WAN):

Multiple Access Techniques for Fiber-Optic Networks

“…To see this clearly, one can rearrange Eq. (19) to yield immediately 8 Gbps. This ultimate shot-noise capacity limit, however, can never be attained as the other noise sources always dominate.…”

“…Fiber-based LANs and MANs like the Fiber Distributed Data Interface (FDDI) [5], the Fiber Channel Standard (FCS) [6], and the Distributed-Queuing Dual Bus (DQDB) [7] have been in use for some time ago. However, as no attempt was made to take benefit of the particular features offered by fiber-optic technology, the capacity of these networks is restricted by the basic electronic bottleneck to typically a few hundreds of Mbps that must be shared among the many nodes connected to the network [8].…”

Disaster Relief Management Using Reinforcement Learning-Based Routing

“…The shortest path algorithm always selects the shortest path, which is in terms of number of hops (Li, 1991). The shortest path may not be called as optimum path as shortest path routing algorithm never considers the traffic present on the network.…”

Routing equal-size messages on a slotted ring

“…All the nodes have the ability to recognize empty slots or slots addressed to them." A regular topology of the ring and no buffering in the slotted ring are two desirable features of modern high speed networks according to a recent survey, Li et al (2007). Consequently, slotted rings have found numerous real-life implementations.…”