Showing papers on "Network topology published in 1985"

Network Observability: Identification of Observable Islands and Measurement Placement

Abstract: Two algorithms are presented; one for (i) testing the observability of a network and (ii) identifying the observable islands when the network is unobservable, and the other for selecting a minimal set of additional measurements to make the network observable. The two algorithms are based on triangular factorization of the gain matrix and are characterized by (i) being extremely simple, (ii) using, subroutines already in a state estimation program, and (iii) incurring very little extra computation. The design and testing of the algorithms are presented in this paper.

A Distributed Drafting Algorithm for Load Balancing

Abstract: It is desirable for the load in a distributed system to be balanced evenly. A dynamic process migration protocol is needed in order to achieve load balancing in a user transparent manner. A distributed algorthim for load balancing which is network topology independent is proposed in this paper. Different network topologies and low-level communications protocols affect the choice of only some system design parameters. The "drafting" algorithm attempts to compromise two contradictory goals: maximize the processor utilization and minimize the communication overhead. The main objective of this paper is to describe the dynamic process migration protocol based on the proposed drafting algorithm. A sample distributed system is used to further illustrate the drafting algorithm and to show how to define system design parameters. The system performance is measured by simulation experiments based on the sample system.

Regular mesh topologies in local and metropolitan area networks

Abstract: The throughput per user in loop and bus configured local area networks decreases linearly with the number of users. These networks cannot be extended to a metropolitan area with many users. A class of mesh networks is described that increases the throughput of conventional local area networks by decreasing the fraction of the network capacity needed to transmit information between a source and a destination. These networks have multiple paths that increase the reliability of the networks, and have point-to-point links that can cover a metropolitan area. In general, mesh networks require complex store-and-forward nodes that also route messages, control the flow of data entering the network, resequence packets at the destination, and recover packets with errors. However, there are characteristics of the local or metropolitan area that allow these functions to be simplified. As a result of these simplifications, loop access protocols are extended to mesh networks and the need to store and forward data is eliminated. A file transfer protocol that does not require packet resequencing is described. Three mesh networks are studied, and the desirable characteristics of networks are determined. One network, the Manhattan street network, has many of the desirable characteristics.

Reliable Loop Topologies for Large Local Computer Networks

Abstract: Single-loop networks tend to become unreliable when the number of nodes in the network becomes large. Reliability can be improved using double loops. In this paper a highly reliable and efficient double-loop network architecture is proposed and analyzed. This network is based on forward loop backward hop topology, with a loop in the forward direction connecting all the neighboring nodes, and a backward loop connecting nodes that are separated by a distance ⌊√N⌋where N is the number of nodes in the network. It is shown that this topology is optimal, among this class of double-loop networks, in terms of diameter, average hop distance, processing overhead, delay, throughput, and reliability. The paper includes derivation of closed form expressions for diameter and average hop distance, throughput, and number of distinct routes between two farthest nodes. For fault-tolerance study, the effect of node and link failures on the performance of the network is analyzed. A simple distributed routing algorithm for reliable loop network operation is also presented.

Fiber Optic Configurations for Local Area Networks

Abstract: A number of fiber optic configurations for a new class of demand assignment multiple-access local area networks requiring a physical ordering among stations are proposed. In such networks, the data transmission and linear-ordering functions may be distinguished and be provided by separate data and control subnetworks. The configurations proposed for the data subnetwork are based on the linear, star, and tree topologies. To provide the linear-ordering function, the control subnetwork must always have a linear unidirectional bus structure. Due to the reciprocity and excess loss of optical couplers, the number of stations that can be accommodated on a linear fiber optic bus is severely limited. Two techniques are proposed to overcome this limitation. For each of the data and control subnetwork configurations, the maximum number of stations as a function of the power margin, for both reciprocal and nonreciprocal couplers, is computed.

Computer-Aided Noise Analysis of Linear Multiport Networks of Arbitrary Topology

Abstract: The paper discusses a new algorithm for the noise analysis of a linear multiport network. The circuit may include any kind of passive components introducing thermal noise only, and any number of two-port devices described by the usual four noise parameters. On output, the algorithm produces the correlation matrix of the Norton equivalent noise current sources at the network ports. The approach is suitable for implementation into any general-purpose microwave circuit design program.

Concurrent Broadcast for Information Dissemination

Abstract: Concurrent broadcast involves the dissemination of a database, consisting of messages initially distributed among the nodes of a network, so that a copy of the entire database eventually resides at each node. One application is the dissemination of network status information for adaptive routing in a communications network. This paper examines the time complexity and communication complexity of several distributed procedures for concurrent broadcast. The procedures do not use information depending on the network topology. The worst-case time complexity of a flooding procedure for concurrent broadcast is shown to be linear in the number of nodes plus the number of messages, and no other procedure for concurrent broadcast has a better worst-case time complexity. A variant of flooding is proposed to eliminate redundant message receipts from the flooding process by real-time signaling between neighbors concerning messages residing at each. This variant can reduce communication complexity, while having a worst-case time complexity similar in form to that of the flooding procedure. Special properties of concurrent broadcast in a tree are also given. The present time complexity results can be used to bound the time during which inconsistent databases may reside at different nodes, to evaluate and compare procedures for (or including) concurrent broadcast, and to schedule a sequence of instances of concurrent broadcast so that the instances do not overlap and there is no need for sequence numbers.

Reconfiguration Strategies for Parallel Architectures

Abstract: Communication paths between parallelarchitecture resources can be reconfigured to suit new computational structures; however, this capability places new demands on efficient architecture use. In the past decade a significant amount of research focused on the development of highly parallel architectures.' Progress in research on Very-Large-Scale-Integrated-Circuit (VLSI) and Very-High-Speed-Integrated-Circuit (VHSIC) technologies is making it feasible to consider the construction ofcomplex parallel architectures that comprise a number of processors communicating by means of a high-bandwidth interconnection network. (Two often-proposed organizations, the processor-memory and the processor-processor network, are shown in Figure 1.) By utilizing a large number of such processors, these architectures have the potential to provide enormous throughputs leading to the computation of processes previously considered impractical due to their complexity. However, the availability of multiple processors alone is not sufficient to ensure practical solutions to computationally intensive tasks. One of several modes of operation may be de-

Rude-CSMA: A Multihop Channel Access Protocol

Abstract: In this paper, we define a two-parameter family of protocols designed for multihop packet radio networks. We call these protocols rude-CSMA because under certain circumstances, maximum throughput is obtained when nodes, even after sensing a busy channel, transmit packets anyway with a nonzero rate. The performance of these protocols is analyzed for various special and random topologies.

A Distributed Evolutionary Algorithm for Reorganizing Network Communications

Abstract: The Distributed Evolutionary Algorithm (DEA) presented in this paper produces efficient TDMA schedules for communication in an environment of changing network topology by passing topological and traffic information among nodes communicating on the schedule. As information is gained in this way more nodes are added to the communicating group until the entire network is communicating on a common schedule. The key feature of the DEA presented is that the reorganization phase is fast, requiring a number of TDMA slots only on the order of the number of nodes. A secondary feature is that some network communication can take place, and thus some of the traffic requirements can be satisfied, even while the network is being reorganized. Numerical results are presented for a number of randomly generated networks which show how the algorithm performs.

Basis exchange characterizations for the simplex son algorithm for LP/embedded networks

Abstract: The simplex special ordered network (SON) algorithm is a partitioning method for solving LP problems with embedded network structure. The algorithm derives from a theoretical characterization of the network topology of the basis embodied in a specially constructed master basis tree. In this paper we show that the topology of the master basis tree and the rules by which it can admissibly be restructured can be characterized by seven mutually exclusive and collectively exhaustive basis exchange cases. Further, these seven cases will always keep the network portion of the basis at its maximum dimension.

An Analysis of the Effect of Network Parameters on the Performance of Distributed Database Systems

Abstract: Performance analysis studies of distributed database systems in the past have assumed that the message transmission time between any two nodes of a network is constant. They disregard the effect of communication network parameters such as network traffic, network topology, and capacity of transmission channels. In this paper, an analytical model is used to estimate the delays in transmission channels of the long haul network supporting the distributed database system. The analysis shows that the constant transmission time assumption cannot be justified in many cases, and that the response time is sensitive to the parameters mentioned above. Extensions and performance analysis in the context of interconnection networks are also discussed.

The fibernet II ethernet-compatible fiber optic LAN

Abstract: Fibernet II is a fiber-optic local area network (LAN) having an active-star configuration. It is plug compatible with the 10- Mbit/s coaxial-cable Ethernet LAN at its transceiver cable interface. Ethernet requires the detection of packet collisions; this function is implemented in Fibernet II at the active star node. Collision presence is signaled to the host transceivers with a unique in-band optical signal. Fibernet II features improved electromagnetic immunity, absence of signal radiation from cables, freedom from ground-loop currents due to ground potential differences between remote sites, new network topology options, and a growth path to future broad-band services. This paper describes the Fibernet II experimental system as implemented at the Xerox Palo Alto Research Center. A space-time analysis is presented which confirms that Fibernet II can support a network diameter of over 4.0 km, in contrast to the coax Ethernet which is limited to 2.5 km.

Distributed match-making for processes in computer networks (preliminary version)

Abstract: In the very large multiprocessor systems and, on a gander scale, computer networks now emerging, processes are not tied to fixed processors but run on processors taken from a pool of processors. Processors are released when a process dies, migrates or when the process crashes. In distributed operating systems using the service concept, processes can be clients asking for a service, servers giving a service or both. Establishing communication between a process asking for a service and a process giving that service, without centralized control in a distributed environment with mobile processes, constitutes the problem of 1 distributed matchmaking. Logically, such a match-making phase precedes routing in store-and-forward d6mputer networks of this type. Algorithms for distributed match-making are developed and their complexity is investigated in terms of message passes and in terms of storage needed. The heoretical limitations of distributed matchmaking are established, and the techniques are applied to several network topologies.

Hybrinet: a hybrid bus and token ring network

Abstract: We propose a hybrid local area network, called the Hybrinet, which combines all the good qualities of the CSMA/CD network and the token ring network. The performance of this network is shown to be superior to that of the token ring network for all offered loads where delay is finite. Furthermore, over the range of offered loads where the CSMA/CD network performs better than the token ring network, it is shown that the Hybrinet's performance is very close to that of the CSMA/CD network.

Conditions for Product Form Solutions in Multihop Packet Radio Network Models

Abstract: : Consider multihop packet radio networks operating under a general class of channel access protocols. For the purpose of throughput analysis, analytical models are considered which describe the joint activity of the transmitters in the network, under the assumptions of heavy traffic and zero propagation and processing delays. The problem addressed in this report is that of finding conditions for the existence of product form solutions for the steady-state probabilities of these models. The main result states that a necessary and sufficient conditions for a given network topology, channel access protocol, and traffic pattern, to lead to a product form solution is that the blocking between each pair of used links, as specified by the access protocol, by symmetric. This result assumes Poisson scheduling point processes associated with the links of the network. The proof is given in two steps: first, for systems where all packet length distributions are exponential, giving rise to Markovian processes; and second, for general packet length distributions (subject to the restriction of possessing a positive density almost everywhere), giving rise to Generalized Semi-Markov Processes. It is also shown that a product form solution does not exist whenever any of the scheduling point process in the network is not Poisson. In addition, it is proven that the computation of the normalization factor appearing in the expression of the product form solution is an NP-hard problem. (rh)

The number of ₂-precompact group topologies on free groups

Abstract: It is shown that every free group F admits exactly 221FI T2precompact group topologies. Those topologies can even be chosen to be finer than the finite-index topology on F.

L-Expressnet: The Communication Subnetwork for the C-NET Project

Abstract: L-Expressnet is the communication suxbnetwork for the CNET local area network project of the CNR, Italy. Based on a single bus topology, it utilizes a simple and efficient virtual-token access protocol which provides ordered and collision-free transmission. The paper discusses the needs which led to L-Expressnet, describes the protocol, proves its correct operation, and compares its throughput-delay performance with those of other typical bus Protocols. A major advantage of L-Expressnet is that its implementation may be based on Ethernet standard transceivers and network interfaces.

A Comparison of the Performance of Protocols in Packet Radio Networks

Abstract: The relative performance and merits of a number of multiple access protocols for multihop packet radio networks are compared. The effects of network size, topology, connectivity, and noise immunity are also studied.

Automatic Update of Replicated Topology Databases

Abstract: In computer communication networks, routing is often accomplished by maintaining copies of the network topology and dynamic performance characteristics in various network nodes. The present paper describes an algorithm that allows complete flexibility in the placement of the topology information. In particular, we assume that an arbitrary subset of network nodes are capable of maintaining the topology. In this environment, protocols are defined to allow automatic updates to flow between these more capable nodes. In addition, protocols are defined to allow less capable nodes to report their topology data to the major nodes, and acquire route information from them.

A Queueing Model for Token-Passing Computer Networks

Abstract: Token-passing computer networks operating on ring and bus topologies can be modelled as multiqueucing systems. The free token behaves as a server (or key to channel access) that provides each smion with a chance to use the chaIUlel for a finite time. The strictly cyclic service pattern and non-exhaustive service causes interference bctween queues, and consequently, dependencies between the differem queueing processes. In the past, an assumption of independence between queues (station independence) has commonly been used for analyzing such systems. In this paper an exact method is presented, based on a Markov chain embedded in a certain semiMarkov process. The result yields a computational fonn for the steady-state distribution of the token's random cycle-time on th network. This result confinns the inadequacy of the s[:J.tion independence assumption at all but very extreme system loads. In addition to the exact res an assumption of packet independence in token-passing networks is introduced, in order to simplify computational effort. It is shown that unlike station independence, the packet independence assumption works remarkably well. Using asymmetric Poisson arrivals and otherwise general distributions, the cycle-time distribution of the token is computed exactly, as well as approximately under ooth kinds of independence assumption The method we imroduce leads to a vacationing-server based queueing model for general, asymmetric loken-passing systems. A few applications of cyle-times are included, for mean channel utilization, distribution of channel utilization, and busy and vacation periods of the token with respect to each station.

An Analytical Framework for Slotted Random Access Spread Spectrum Networks

Abstract: An analytic framework is proposed for the study of Random Access, Packet-Switched, Spread Spectrum Links under various network topologies and channel conditions. The key feature of the theory is the identification of a set of probabilistic parameters which, based on a symmetry argument, serve to efficiently summarize the effect of various network considerations on the link performance such as transmitter/receiver configuration, spread spectrum code allocation, error correction and detection mechanisms, spreading format, jamming condition, etc. The concept is illustrated through examples which show how past results (a) can be included as special cases and (b) can be generalized to a multitude of scenarios.

A CAD Solution to the Generalized Problem of Noise Figure Calculation

Abstract: The paper describes a new algorithm for computing the noise figure of a linear two-port network of arbitrary topology. The discussion is oriented towards CAD applications, making the approach easy to be implemented into any existing general-purpose analysis and/or design program. A few sample results are compared with measured data drawn from the technical literature.

Distributed algorithms for election in unidirectional and complete networks (traversal, synchronous, leader, asynchronous, complexity)

Abstract: Consider a data communication network of n nodes, each of which has a unique identifier (id); otherwise the nodes are identical. The nodes are asleep and have no global information about network topology, number and ids of other nodes, etc. A distributed election algorithm is a means by which the nodes of the network distinguish one among them as the leader. The problem of distributively electing a leader in a network is viewed as a problem of synchronization among potential candidates for leadership. Each candidate tries to capture all the nodes. To guarantee that only one succeeds, all but one candidate are killed. Following this view election algorithms in a general, two component framework are designed. Component one is a capturing and termination detection mechanism, assuming only one candidate. Component two is a synchronization mechanism, to eliminate all but one candidate. In arbitrary networks the synchronization is complicated by the uncertainties of nodes about the network topology and the relative location of candidates. Two network models are considered: first, a complete network in which a bidirectional communication link connects every node with every other, thus eliminating topological uncertainties; and second, the opposite extreme in which topological uncertainties are at maximum--a strongly connected unidirectional network with some or all links transmitting messages in one direction only. The study produces an O(n(.)log n) messages O(log n) time synchronous and O(n(.)log n) messages O(n) time asynchronous election algorithm in complete networks. For unidirectional networks we derive a distributed election algorithm whose communication complexity is O(n(.)(VBAR)E(VBAR) + n('2)log n) bits, where (VBAR)E(VBAR) is the total number of links. We also establish that (OMEGA)(n(.)log n) is a lower bound on the total number of messages transmitted for achieving election in synchronous complete networks. Moreover, it is shown that the time complexity of message-optimal synchronous algorithms is (OMEGA)(log n), hence the optimality of our synchronous complete network algorithm. It remains open whether a sublinear time, message-optimal, asynchronous complete network election algorithm exists.

Determining the updating interval of a round robin sequence for token passing mobile networks

Abstract: The optimal token passing sequence for a round robin token passing protocol depends on node spatial configuration (it is the sequence that minimizes the sum of token passing times) When a network contains mobile nodes, a once optimal sequence can, after some time, yield poor performance due to changes in node configuration The sequence must then be updated Performance deterioration is a function of node speeds, type of motion, and time We evaluate this deterioration (by means of theoretical analysis and simulation experiments) for the purpose of determining an inter-update interval that allows the network to maintain good performance

Test point selection methods for the self-testing based analogue fault diagnosis system

Abstract: Two methods for selecting test points of an analogue circuit are presented. The purpose of making such a selection is to increase the testability of the self-testing algorithm which is used to allocate the defective components of a given analogue circuit. Some closed relations between network equations and network topology have been investigated so that the computational effort can be reduced to minimum. This comes up with the first test point selection method. The second method for selecting a proper set of test points for fault diagnosis is a table look-up approach by which no complex matrix computation is required. Instead, only a few searches, which are based on a simple rule, on a binary matrix are needed. The cost of applying this table look-up method is that the testability of the target circuit due to the selected test points may not be the highest. Both methods are applicable to linear and nonlinear circuits.