Showing papers on "Routing protocol published in 1984"

Second Derivative Algorithms for Minimum Delay Distributed Routing in Networks

Abstract: We propose a class of algorithms for finding an optimal quasi-static routing in a communication network. The algorithms are based on Gallager's method [1] and provide methods for iteratively updating the routing table entries of each node in a manner that guarantees convergence to a minimum delay routing. Their main feature is that they utilize second derivatives of the objective function and may be viewed as approximations to a constrained version of Newton's method. The use of second derivatives results in improved speed of convergence and automatic stepsize scaling with respect to level of traffic input. These advantages are of crucial importance for the practical implementation of the algorithm using distributed computation in an environment where input traffic statistics gradually change.

Datagram routing for internet multicasting

Abstract: We present a solution to the problem of multidestination routing in internetworks. The component subnets of these internets share a common datagram internet layer, and the gateways and hosts can determine the next gateway en route to a foreign net. Our datagram routing offers high resilience to network failures, major reductions in network traffic, and no changes whatsoever to the subnetwork routing. The routing follows “shortest” paths as defined by the distance criteria of an internet. We intend to use the algorithm as an option of the DoD Internet Protocol with only minor changes to IP while preserving interoperability with IP modules not supporting multidestination.

Alternate paths in a self-routing packet switching network

Abstract: A packet switching architecture in which switching network nodes (200-206) automatically determine alternate routes through a switching network (107) so as to increase reliability and distribute traffic. The switching network comprises stages of distribution (D) and routing nodes (R). The routing nodes are responsive to physical addresses associated with the packets to communicate those packets to address designated downstream nodes. The distribution switching nodes statistically communicate packets to downstream switching nodes on the basis of an alternate routing algorithm and availability of downstream nodes. The initial network stages are alternate distribution and routing stages followed by only routing stages for the remaining stages of the network. Both the routing and the distributing nodes are identical in design and are responsive to an input signal from the network to assume either routing or distributing functions.

An Efficient Channel Router

Abstract: In the LSI chip layout design, channel routing is one of the key problems. The problem is to route a spcified net list between two rows of terminals across a two layer channel. This paper presents a new routing algorithm, which is an improved version of the classical "left edge algorithm". The new algorithm uses a row by row approach, calculating an optimum net assignment to each row. The algorithm was implemented for examples in previously published papers. Experimental results show that the new algorithm produces optimum solutions in most cases.

Routing Strategies in Computer Networks

Abstract: A principal function of distributed computer networks is message routing in which a data unit is moved along a network path from where it is first accepted (source node) to where it is to be delivered (destination node), through possibly one or more network nodes. In the OS1 protocol architecture, routing is the principal function of Level 3, the Network Layer. 1 In most operational networks, the network automatically generates the routes used to forward data units. To use network resources efficiently, selected routes must involve minimum cost, according to some specific cost criteria, such as the number and type of communications channels on the route, the utilization levels of the channels, and the average delay incurred in transporting a data unit along the route. The cost of a route is the sum of the costs of the constituent links. Once we know the cost of the network links, we can calculate the minimum-cost path for any source-destination pair using algorithms developed in operations research. Major issues related to routing include

Sequencial Routing Optimization for Circuit Switched Networks

Abstract: Nonhierarchical alternate routing is the simplest extension to the hierarchical routing currently used in telephone networks. In order to evaluate the merits of such a technique, it is necessary to be able to optimize some performance measure over the routing variables. In a first step, a short discussion is presented of the evaluation algorithms used for computing the end-to-end blocking probabilities in networks operating under nonhierarchicai alternate routings. Then the results of a statistical sampling of the routing space for a small network are presented, to illustrate important characteristics of the routing problem. Next, two heuristic optimization techniques are discussed, one based on classical steepest descent and the other inspired from classical routing optimization of traffic flows in transportation and communication networks. Acceleration techniques are discussed in both cases, and the convergences rates of the alorgithms are presented for a number of networks. A brief discussion is finally given of the usefulness of nonhierarchical alternate routing in reducing the blocking probability in already existing networks, and in decreasing the number of trunks required to maintain a prespecified grade of service, both in real and artificially generated networks.

Routing of telephone traffic as a controlled Markov process

Abstract: In a modern stored-program-controlled telephone network, a considerable amount of information regarding the state of the network can be made available for operational decisions such as traffic routing. This paper is an attempt to obtain a computable solution to the problem of optimum traffic routing with state information, in the framework of the theory of Markov decision processes.

Distributed asynchronous optimal routing in data networks

Abstract: We prove convergence of a distributed gradient projection method for optimal routing in a data communication network. The analysis is carried out without any synchronization assumptions and takes into account the possibility of transients caused by updates in the routing strategy being used.

Fibre-optic local-area network with arbitrary topology

Abstract: The lack of satisfactory passive connectors, of standardisation in fibre-optic devices and the unidirectional light transmission employed in most fibre links prohibits the direct translation of networking techniques from the better developed wire or cable systems. A point-to-point network with nodes connected in an arbitrary fashion can provide an ideal solution to these problems while maintaining high network reliability and permitting easy reconfiguration, but has associated problems in the routing of the data. A possible solution is to use the flood routing technique as proposed by Neff and Senzig in their Anarchy network. Unfortunately, this network exhibits very low channel utilisation and has packet looping problems. In the paper we discuss methods of removing the looping problem and then propose a modified Anarchy flood routing (MAFR) protocol, for which a simplified performance analysis and simulation studies indicate that better network performance can be achieved with little additional hardware. However, we believe that a hybrid flood routing scheme (HFR) involving MAFR and a virtual circuit approach will produce a better system.

Local Routing of Two-terminal Nets is Easy

Abstract: A local routing problem is given by a routing region (a subgraph of the planer grid) and a set of nets. For each net a global routing is also given. The problem is to find a local routing which is consistent with the global routing (if there is one). In this paper we show that local routing problems can be sloved in time O(n(log n)2).

RMCNET-4: A Survivable Local Area Network

Abstract: connectivity of four. This paper describes a new topology for a network with connectivity of four. The topology has a regular structure and can accommodate any number (>8) of nodes. The paper also describes the proposed routing and protocol for data transfer between any two nodes of the network. Beforetransferring data the sending node must establish connection to the destination node. This is done by use of a special packet called 'Interconnection Packet'. The performance of the network has been evaluated by simulations done ona Honeywell DPS8 computer using Pascal compiler. Results of the simulations are presented.