Showing papers on "Node (networking) published in 1973"

Network Design with Fixed and Variable Cost Elements

Abstract: A route selection algorithm is presented far designing transportation networks. The algorithm balances fixed construction costs and variable user costs in a network having a fixed set of nodes and a known demand, for internode service. The problem solved is a special case of the fixed-cost, multicommodity transshipment problem in which each commodity has a single, unique source node. The route selection algorithm alternatively applies link elimination and link insertion criteria that converge to a local optimum. Upper and lower bounds on the fixed and variable portions of the globally optimum solution are determined and the sensitivity of the solution is estimated. Unique rules are formulated for identifying links that must or must not appear in the globally optimum solution. The solution procedure has been coded for a digital computer and demonstrated using a representation of Minneapolis-St. Paul having 68 nodes and 476 potential oneway links.

Computer-communication networks

Abstract: In 1968 the Advanced Research Projects Agency (ARPA) of the U.S. Department of Defense began implementation of a computer communication network which permits the interconnection of heter ogeneous computers at geographically distributed centres through out the United States. This network has come to be known as the ARPANET and has grown from the initial four node configuration in 1969 to almost forty nodes (including satellite nodes in Hawaii, Norway, and London) in late 1973. The major goal of ARPANET is to achieve resource sharing among the network users. The resources to be shared include not only programs, but also unique facilities such as the powerful ILLIAC IV computer and large global weather data bases that are economically feasible when widely shared. The ARPANEr employs a distributed store-and-forward packet switching approach that is much better suited for computer communications networks than the more conventional circuit-switch ing approach. Reasons favouring packet switching include lower cost, higher capacity, greater reliability and minimal delay. All of these factors are discussed in these Proceedings."

Optimal Locations on a Network

Abstract: The paper considers the location of centers (e.g., plants, warehouses) on a network. Necessary conditions and sufficient conditions for optimal solutions to occur at nodes in the network are determined. The conditions include consideration of capacity constraints and the possibility that not all shipments are made via the “shortest route.” The cost functions may differ from arc to arc. A proof is given for the sufficiency of node locations in situations of multistage processing when transport costs are concave over each arc. Three sets of conditions under which it is necessary for an optimal solution to occur at a node are investigated. One set of conditions involves a production function characterization of the centers being located.

Letter to the Editor—Extensions of the Augmented Predecessor Index Method to Generalized Network Problems

Abstract: The augmented predecessor indexing method is a procedure for efficiently updating the basis representation, flows, and node potentials in an adjacent extreme point (or “simplex” type) method for network problems, utilizing ideas due to Ellis Johnson in his proposed application of a triple-label representation to networks. The procedure is extended here to accommodate the more complex basis structures and updating processes of the generalized network problem, specifying rules for expediting the calculations.

Extensions of the Augmented Predecessor Index Method to Generalized Network Problems

Abstract: : The augmented predecessor indexing method is a procedure for efficiently updating the basis representation, flows and node potentials in an adjacent extreme point (or 'simplex' type) method for network problems, using ideas due to Ellis Johnson in his proposed application of a triple-label representation to networks. The procedure is extended here to accommodate the more complex basis structures and updating processes of the generalized network problem, specifying rules for expediting the calculations.

The expected node-independence number of random trees

Abstract: We shall derive a formula for the expected value μ(n) of the node-independence number of a random tree with n labelled nodes and we shall determine the asymptotic behaviour of μ(n) as n tends to infinity.

Apparatus and method for connecting between series and parallel data streams

Abstract: Apparatus and method for converting between series and parallel data streams. In a charge transfer device (CTD) shift register having 2M (M is an integer greater than one) parallel registers each having an input node and an output node, a CTD input logic tree connects all the input nodes to a single data input node, and a CTD output logic tree connects all the output nodes to a single data output node. The input logic tree successively splits up an input data stream into two data streams until the number of data streams is 2M and the output logic tree successively combines data streams from each output node two at a time until a single output data stream is achieved.

Asynchronous data transmission arrangement

Abstract: A data transmission arrangement is disclosed wherein messages consisting of a synchronization code field, a destination terminal code field, a source terminal code field with sufficient bit capacity for both a source terminal code and a synchronization code, and a data field are transmitted between a plurality of transmission node terminals which are serially connected in a closed, asynchronous data transmission loop. Each node terminal contains a control circuit and a plurality of shift registers which the control circuit selectively connects in series with the data transmission loop to receive or to transmit the various fields of a message. In receiving a message only those shift registers of the receiving node terminal which are necessary to receive the synchronization code and the destination terminal code for the message are initially connected in series with the transmission path. If the message is not intended for the particular node terminal, no additional shift registers are employed and the message passes through the node terminal with only the message delay inherent in the use of the initially connected shift registers. If the message is intended for the node terminal, additional registers are employed to receive the remaining message fields. A message may be transmitted from a node terminal whenever the terminal determines that no message is currently being received. As a message is transmitted the control circuit removes shift registers from series connection in the data path as soon as possible to minimize possible transmission delay to messages received by the node terminal during the remainder of the message transmission. If during the transmission of a message from a node terminal, a message is received by that node terminal, it is shifted into whatever shift registers are currently connected in series with the data path and, thereafter, treated as any other received message.

The shortest and the K‐shortest routes as assignment problems

Abstract: The problem of finding a shortest route in a network with unrestricted costs is approached through solving an assignment problem associated to the network. The upper bound on the number of elementary calculations required for the solution is 0(m3). However, in most cases, the actual number of computations is considerably less and depends on different network characteristics than Dynamic Programming algorithms do. In examples of networks generated stochastically, this number was below 0(m2.5). A parametric analysis is presented. It is shown that if after a shortest route is determined, the costs on all arcs incident into or out of a node are modified in any form, at most 0(m2) elementary calculations will determine a new optimal solution. This feature, shared by Dynamic Programming algorithms only for cases where all cost decrease, can be applied to problems such as the determination of the K-shortest routes and the K-smallest assignments, leading to upper bounds of 0(Km3) in both cases.

A Note on Yen's Algorithm for Finding the Length of All Shortest Paths in N-Node Nonnegative-Distance Networks

Abstract: An error in Yen's algorithm is pointed out, and an alternative is offered which will produce correct results.

A loop network for general purpose data communications in a heterogeneous world

Abstract: How do you interconnect a large, heterogeneous group of computers, batch terminals and conversational terminals to form a general purpose network of computing resources? What techniques and technologies can be applied to maximize throughput and minimize response times? Is major surgery on host processor operating systems really necessary to interface to a communications network?This paper will describe a packet switching data communications network under development at the National Security Agency for resource sharing and the future development of distributed processing and filing systems. Starting with an experimental three node loop, the network will consist of a hierarchy of interconnected loops or rings probably based on Bell System T carrier digital transmission technology.Also discussed will be the means of nodal connection to the loops, the configuration of a node, network protocol, security considerations and implications for the future.

Data transmission system with multiple access for the connected users

Abstract: In a communication network having a plurality of branches with associated user stations, a data transmission system is utilized in which interconnections are provided by a time-division multiplex multiple access method between the user stations, each station being connected in a branch and the branches being interconnected through a plurality of nodes. Each branch includes a transmission line and a receiving line, which are connected at one point in the network. A synchronizing word generator, which is centrally located within the network, supplies synchronizing words to the system. At each of the user stations, the interval between successive synchronizing words is divided into n equal time slots. When a user station desired to establish a connection with the system, the user transmits a test block signal into a free time slot. If this test block signal simultaneously arrives at a node simultaneously with a signal from another station, the test block from the user is shifted into a subsequent unoccupied time slot and then transmits its data in that subsequent slot during each successive interval.

Network statistics for computer-aided network analysis

Abstract: Comparative studies of computer-aided network analysis (CANA) program performance have been handicapped by a lack of statistics about networks. This paper derives bounds and presents empirical values for: 1) the number of elements per node; 2) the sparsity of the cutset matrix; 3) the density of nonlinear nodes; 4) the relative percentage of excess capacitors and inductors; and 5) the relationship between the number of elements of each element type and the total number of elements.

Optimal source location to feed a lossy distribution tree

Abstract: Given a large but finite tree network designed to distribute a commodity to some or all of its nodes, the lossy model takes into account the attenuation of the flow through the branches of the network. The capacity of the source needed to feed the network depends on the node to which it is attached. In a finite network, there must be at least one node where the required source capacity is not higher than that at any other node. This short paper proves that there can be only one such node or, at most, two adjacent nodes with the same value for the required source capacity. Regarding the required source capacity as a function over the set of nodes, it is shown that this function can have only one local minimum (a node where the value is not higher than at any adjacent node), and no local maximum. Based on these properties, search algorithms are outlined to locate the node with the optimal source capacity.

Technical Note-Maximal Funnel-Node Flow in an Undirected Network

Abstract: This note formulates, examines, and solves a funnel-node maximal-flow problem for an undirected network. The solution procedure requires only applications of the single-commodity flow algorithm, and is therefore extremely efficient. Several applications are presented.

A time reference distribution concept for a time division communication network

Abstract: Starting with an assumed ideal network having perfect clocks at every node and known fixed transmission delays between nodes, the effects of adding tolerances to both transmission delays and nodal clocks is described. The advantages of controlling tolerances on time rather than frequency are discussed. Then a concept is presented for maintaining these tolerances on time throughout the network. This concept, called time reference distribution, is a systematic technique for distributing time reference to all nodes of the network. It is reliable, survivable and possesses many other desirable characteristics. Some of its features such as an excellent self monitoring capability will be pointed out. Some preliminary estimates of the accuracy that might be expected are developed and there is a brief discussion of the impact upon communication system costs. Time reference distribution is a concept that appears very attractive. It has not had experimental evaluation and has not yet been endorsed for use in any communication network.