Abstract: Following a companion paper on analytical methods, this paper presents simulation as a complementary method for analyzing the reliability of water distribution networks. For this simulation, the distribution system is modeled as a network whose pipes and pumps are subject to failure. Nodes are targeted to receive a given supply at a given head. If this head is not attainable, supply at the node is reduced. Pumps and pipes fail randomly, according to probability distributions with user‐specified parameters. Several reliability measures are estimated with this simulation. Confidence intervals are also supplied for some of these reliability measures. Simulation results are presented for a small network (ten nodes) and a larger network (sixteen nodes). Simulation enables computation of a much broader class of reliability measures than do analytical methods, but it requires considerably more computer time and its results are less easy to generalize. It is therefore recommended that analytical and simulation me...

Abstract: In a distributed heterogeneous computer system having a plurality of computer nodes each operatively connected through a network interface to a network to provide for communications and transfers of data between the nodes and wherein the nodes each have a queue for containing jobs to be performed, an improvement for dynamically reallocating the system's resources for optimized job performance. There is first logic at each node for dynamically and periodically calculating and saving a workload value as a function of the number of jobs on the node's queue. Second logic is provided at each node for transfering the node's workload value to other nodes on the network at the request of the other nodes. Finally, there is third logic at each node operable at the completion of each job. The third logic includes, logic for checking the node's own workload value, logic for polling all the other nodes for their workload value if the checking node's workload value is below a preestablished value indicating the node as being underutilized and available to do more jobs, logic for checking the workload values of the other nodes as received, and logic for transfering a job from the queue of the other of the nodes having the highest workload value over a preestablished value indicating the other of the nodes as being overburdened and requiring job relief to the que of the checking node. The third logic is also operable periodically when the node is idle.

Abstract: When two or more computing agents work on interacting tasks, their activities should be coordinated so that they cooperate coherently. Coherence is particularly problematic in domains where each agent has only a limited view of the overall task, where communication between agents is limited, and where there is no ``controller'' to coordinate the agents. Our approach to coherent cooperation in such domains is developed in the context of a distributed problem-solving network where agents cooperate to solve a single problem. The approach stresses the importance of sophisticated local control by which each problem-solving node integrates knowledge of the problem domain with (meta-level) knowledge about network coordination. This allows nodes to make rapid, intelligent local decisions based on changing problem characteristics with only a limited amount of intercommunication to coordinate these decisions. We describe three mechanisms that improve network coherence: 1) an organizational structure that provides a long-term framework for network coordination to guide each node's local control decisions; 2) a planner at each node that develops sequences of problem-solving activities based on the current situation; and 3) meta-level communication about the current state of local problem solving that enables nodes to dynamically refine the organization. We present a variety of problem-solving situations to show the benefits and limitations of these mechanisms, and we provide simulation results showing the mechanisms to be particularly cost effective in more complex problem-solving situations. We also discuss how these mechanisms might be of more general use in other distributed computing applications.

Abstract: A method is used for collecting information at the session level for a multiple node distributed processing system. A first component software program monitors all packets passing through a node and extracts packet headers having a predetermined format from all nodes. A second component software program identifies a session in which each of the extracted packet headers was transmitted and, for each session, accumulates characterizing information about that session and calculates statistical data concerning all the sessions.

Abstract: We describe a new approach to programming distributed-memory computers. Rather than having each node in the system explicitly programmed, we derive an efficient message-passing program from a sequential shared-memory program annotated with directions on how elements of shared arrays are distributed to processors. This article describes one possible input language for describing distributions and then details the compilation process and the optimization necessary to generate an efficient program.

Abstract: In maintaining a communication network of processing units distributed in multiple nodes linked by communication channels, system software in a plurality of data processing units is updated by first installing the updated software in a first node. The updated software is transmitted through the network to other nodes. A trial use of the updated software is initiated in the nodes. If failures of the updated software are detected in a node, that node will be restored to the original software version. If the trial use of the updated software is completed successfully in a node, the updated version will be installed as a preferred operational version in the node. a

Abstract: A fault tolerator for an operations controller of a multiple node fault tolerant processing system having a data memory for storing the content of all received error free messages, an error file for storing the content of all received inner node error reports, an error handler for generating a base penalty count for each node based on the content of the errors recorded in the error file and for excluding each node from the operation of the multiple node processing system whose base penalty count exceeds an exclusion threshold. The fault tolerator also includes a synchronizer interface for passing the selected fields of the received messages to a synchronizer, a scheduler interface for passing selected information to a scheduler, and a message interface which stores the error free messages in the data memory and passes the selected fields of the messages to the synchronizer.

Abstract: A system and method of down loading, over a network, operating systems or other executable programs to a computer which does not have a boot device or other device containing the executable program. Down loading is accomplished without modification of the loadable image. The computer has a network interface which requests a minimum-boot program be transferred from a host computer on the network. The minimum-boot program, when executed, establishes a logical connection to a disk server on the network and allows the requesting computer to treat the disk server as a local boot device.

Abstract: A transmission system comprises nodes coupled in a ring via two multiplexed transmission paths having opposite transmission directions. The two paths can be time, space, or wavelength division multiplexed. Bidirectional communications between any two nodes are provided in normal operation using both paths around a first, most direct, part of the ring. In the event of a fault affecting such communications, a protection switch is made to maintain bidirectional communications between the two nodes using both paths around a second, remaining, part of the ring. In this manner, each path can simultaneously carry both normal and protection switched signals. In an embodiment of the invention, only the two nodes adjacent a fault need be aware of the fault, each of these nodes effecting a protection switch for signals terminated at the node and effecting a protective loopback for other signals needing protection. The invention enhances the total capacity of the paths for carrying normal and protection switched signals. A network may comprise a plurality of such ring transmission systems intersecting at at least one network node, the network node comprising a node of each ring and a cross connect switch for switching signals therebetween.

Abstract: In a distributed processing system having a least an originating node and a responding node connected through a communication path, the responding node comprising a plurality of processors and a memory device and where each of these processors is capable of accessing information from a corresponding database residing within the memory device, the inventive method involves: storing context information for an associated conversational transaction using a first processor situated within the responding node wherein the context information is stored at a pre-defined address in a first database residing within the memory device and associated with the first processor; producing a first message using the first processor for transmission from the responding node over the communication path to the originating node wherein the first message contains a first transaction identifier field having a value that corresponds to the pre-defined address; generating within the originating node a second message for transmission from the originating node over the communication path to the responding node wherein the second message contains a second transaction identifier field having substantially the same value as the first transaction identifier filed; and accessing the record using a second processor located within the responding node in response to reception of the second message at the second processor by generating the pre-defined address within the second processor using the value of the second transaction identifier field so as to permit the second processor to incrementally process said conversational transaction within the responding node.

Abstract: A distributed digital data processing system includes a plurality of nodes which communicate over a network. A node maintains one or more objects, each of which may be a file, that is, an addressable unit in the system, such as a program, database, text file, or the like, or a directory which may contain one or more files or other directories. One node maintains a naming service which associates each object in the system with one or more protocol towers. Each protocol tower identifies the object name and a series of entries each identifying a name for each of the protocol layers, along with the communications parameters and address information, to be used in communicating with the object. When a node requires access to an object maintained by another node, it first retrieves from the naming service the protocol towers for the object. The node also maintains a tower identifying the names of each of the protocols over which it can communicate. The node then compares the protocol names in the retrieved protocol towers with the protocol names over which it can communicate. If the protocol names in a retrieved tower match the protocol names in the node's tower, the node uses the communications parameters and address information in future communications with the object. If the node is unable to identify a retrieved protocol tower which matches its supported tower or towers, it is unable to communicate with the object.

Abstract: In a distributed system comprising a plurality of nodes, each node is provided with the same set of generic configuration rules which configures the resources for the node according to the application requirements and the hardware configuration of the node. The resource server modules are configurable at run-time by node-based configuration management processes in accordance with information contained in resource definition messages. A resource definition message is derived from a resource template message in accordance with the information contained in the node definition message. Accordingly, adding or modifying resources at a given node can be accomplished at start-up or at run time without affecting the remainder of the system.

Abstract: A security technique for use in an intelligent network. The security technique provides a method for authorizing access by a process located in an invocation node to an object, or a network resource, located in an execution node. The method includes the steps of granting permission to the invocation node to access the object by transmitting a capability and a signature from the execution node to the invocation node. The capability includes a unique indentifier of the object and access rights to the object. The signature is formed at the execution node by encryption of the capability with an encryption key that is unique to the invocation node and is stored only in the execution node. A request for access to the object is transmitted with the capability and the signature from the invocation node to the execution node. At the execution node, the request is authenticated by encryption of the capability with the encryption key that is associated with the invocation node to form a test signature. Access to the object is authorized only when the test signature matches the signature received from the invocation node.

Abstract: With traditional event-list techniques, evaluating a detailed discrete event simulation-model can often require hours or even days of computation time. By eliminating the event list and maintaining only sufficient synchronization to ensure causality, parallel simulation can potentially provide speedups that are linear in the numbers of processors. A set of shared-memory experiments using the Chandy-Misra distributed simulation algorithm, to simulate networks of queues is presented. Parameters of the study include queueing network topology and routing probabilities, number of processors, and assignment of network nodes to processors. These experiments show that Chandy-Misra distributed simulation is a questionable alternative to sequential simulation of most queuing network models. >

Abstract: NX/2 is the operating system which runs on the nodes of the Intel iPSC®/2 concurrent supercomputer. NX/2 provides all of the standard system services found in the original iPSC node operating system, such as memory management, multiple process control, message passing services, and intertask protection.This paper focuses on the major node operating system enhancements brought about by two different requirements. First, NX/2 had to support very high speed and high throughput message passing. In this regard, we show how NX/2 was tuned to the 32-bit architecture of the iPSC/2 nodes and the Direct Connect technology used to implement the communication subsystem.Second, NX/2 had to support a more streamlined and flexible set of message passing service calls. In this regard, we describe the new set of message passing system calls and discuss how NX/2 implements them. The calls range from a simple, effective set of synchronous calls to advanced asynchronous calls which allow overlap of message passing and processing as well as interrupt-driven message handling.

Abstract: When computing a least weight path from an origin node to a destination node in a data communication network, a route-computing network node uses information provided by the origin and destination nodes to compute least weight routes from those nodes to adjacent network nodes. The route-computing network node uses information in the topology database to compute least weight routes from network nodes adjacent the origin node to network nodes adjacent the destination node. The results of the computations are concatenated to determine an optimum route from the origin node to the destination node. The topology database need not include information about the transmission groups connecting the origin/destination nodes to network nodes.

Abstract: Data communications neworks may be composed of many different physical devices linked together in various layers and using various protocols for communication. The multiple physical elements that make up such communication networks or links are often supplied by diverse vendors. As a result, isolation and diagnosis of communication failures becomes an extremely difficult and cumbersome process. The present system and method simplify the process greatly by premitting a system control operator or control application program to issue generic, non-device-specific problem isolation, control or diagnostic commands to the communication neetwork by identifying the link and target node for which problem isolation and diagnosis is required. The non-device-specific commands are received first at an intermediate translation facility which retrieves communication link physical configuration data and identifies the physical components and characteristics thereof that make up that communication link to the identified target node. The translation facility then issues one or more device-specific problem determination commands on the communication link directed toward said node. It receives one or more device-specific responses from one or more physcial devices in the communication link and, responsive thereto, either issues further device-specific commands on the communication link to complete the diagnosis or issues generic, non-device-specific problem identification results as responses to the requesting control nodes' initial request.

Abstract: In a communications network, each network node can maintain its own list of network resources in a topology database. When the state of a resource "owned" by a particular node changes, that node broadcasts a topology database update (TDU) message to adjacent nodes. Each adjacent node updates its own topology database and rebroadcasts the message. To minimize the amount of information that must be included in TDU messages when two nodes are reconnected after an outage, each node assigns flow reduction sequence numbers (FRSNs) to TDU meassages and keeps a record of the FRSN for the last TDU message sent to an adjacent node. The node also records, for each resource in its database, the FRSN of the last TDU message including that resource. When two nodes are reconnected, the sending node includes in the TDU message only those resources having a FRSN greater than the FRSN assigned to the last TDU sent to the adjacent node to which the TDU message is directed.

Abstract: Disclosed is a system for preparing a route end node series which is equipped with a route search means 7 for preparing a node series, an intersection series or a combination of both as the data series for navigation, and a route end node series preparing means 3 for preparing the node series from the starting point and the destination to the respective nearest particular points. The route end node series preparing means 3 prepares node series by successively performing node search until the node nearest to the point indicated by the positional information on the starting point or the destination is attained. This construction makes it possible to perform course guidance through navigation based upon the node series between the starting point or the destination and the nearest intersection thereto even if the starting point or the destination is set at a point between intersections when particular points are set as intersections.

Abstract: A multiprocessing system is presented having a plurality of processing nodes interconnected together by a communication network, each processing node including a processor, responsive to user software running on the system, and an associated memory module, and capable under user control of dynamically partitioning each memory module into a global storage efficiently accessible by a number of processors connected to the network, and local storage efficiently accessible by its associated processor.

Abstract: Multiple-access algorithms that handle erasures as well as captures are introduced. The algorithms are evaluated according to the maximal throughput that they can support for a Poisson arrival process. An example is given which shows that, in practice, the positive effect of captures compensates the negative effect of erasures. An approach that effectively utilizes the capture phenomena is introduced. This approach incorporates a random power-level-selection scheme that allows each node to choose randomly to transmit in one of several allowable levels of power. Design issues such as number of levels, selection schemes, etc. are discussed. >

Abstract: An adaptive link assignment scheme for dynamically changing communication node topologies such as satellite networks, fleets of ships or aircraft, etc. Periodically each node in the network transmits topology information to all the nodes in the network. Then each node determines the degree of connectivity of the network. It is preferred that the network be at least triconnected, and if the network is less than that each node determines what connections it can make to improve the network connectivity. If more than one alternative is available, a choice is made based on line of sight endurance and then on traffic delay. The identification of the selected connection is then broadcast to all the nodes in the network. Each node thus receives the proposed changes from all the network nodes, and each node then resolves conflicts between the broadcast selections and determines what change it should make. Finally, the changes are implemented. The scheme emphasizes network connectivity to bring the network to a triconnected state and then emphasizes line of sight endurance and reduction of traffic delay.

Abstract: A controlled flow parallel computer incorporating a node drive register which designates the location of executable instructions. The node drive register allows instructions to be executed concurrently and non-deterministically without a complex control. Control token locations in the node drive register associated with a given instruction are filled upon the completion of a prior instruction required to be executed before executing that instruction. The instruction is ready for execution once the control token locations are filled.

Abstract: A method and apparatus is described for enabling efficient, bilateral communications in a network connecting a plurality of target address modules and a plurality of source address generators. The source address generators are enabled to generate requests to target addresses within the target address modules. The function of the network is to forward the requests to the target address modules holding the respective target addresses and to return the replies generated by the target address module to the respective source address generators. The network interconnects the source generators to each target address module and vice versa. The interconnection network includes a plurality of interconnected nodes, each node having M inputs and N outputs (where M may or may not be equal to N) and a processor for carrying out a communication protocol. The protocol comprises each node first placing incoming target addresses from messages appearing on the node's inputs into a queue associated with each respective input. The target addresses of the messages at the head of each queue are examined, and the message is transmitted whose target address bears a predetermined relationship to the other message's target address. The messages are combined in case their target addresses are found to be equal. Additionally, a direction bit queue is provided in which is stored a sequence of bits indicating from which of the M inputs to the node, the transmitted message was received. A "ghost" message is transmitted (which includes the target address of the message transmitted) on the node's other outputs, to notify interconnected nodes of the transmitted target address. In addition, each source address generator periodically generates an "end of stream" message.

Abstract: A ring hub in a local area network. A parallel ring bus (20, 22, 24, 26) connects a plurality of ring controllers (10, 12, 14, 16) in a closed loop. At reset time one controller (12) is selected to act as a ring monitor and is not connected to a network node. Each one of the remaining controllers are connected to a single node, so that one controller is linked to a corresponding node. The ring monitor and the ring controllers communicate by inserting message packets onto the bus and stripping messages and control information off of the bus. A transfer request packet is used by a source controller (10) to signal a destination controller (14) for permission to send a data packet. A packet acknowledge signal is generated at the destination controller (14) by asserting one of the bus control lines. The packet acknowledge signal is used to signal to the source controller that the transfer request packet has been received and placed in a request queue (40) at the destination controller. A send data reply packet is used by the destination controller to signal the source controller that the queued request can now be serviced. Data packets are inserted on the ring bus by the source controller to send the actual data message.

Abstract: This paper describes an integrated tool for monitoring distributed systems continuously during operation. A hybrid monitoring approach is used. As special hardware support a test and measurement processor (TMP) was designed, which is part of each node in an experimental multicomputer system. Each TMP runs local parts of the monitoring software for its node, while all the TMPs are connected to a central test station via a separate TMP interconnection network. The monitoring system is transparent to users. It permanently observes system behavior, measures system performance and records system information. The immense amount of information is graphically displayed in easy-to-read-charts and graphs in an application-oriented manner. The tools promote an improved understanding of run time behavior and performance measurements to derive qualitative and even quantitative assessments about distributed systems. A prototype of the monitoring facility is operational and currently experiments are being conducted in our distributed system consisting of several MC68000 microcomputers.

Abstract: Studies were carried out on an ATM (asynchronous transfer mode) transport network based on the virtual path concept. An arrangement of node functions in this network is proposed in which the call set-up process is excluded from transit nodes. This leads to a reduction in node processing and simplification of equipment by specialization of transit node functions. Virtual path control is shown to bring benefits in efficiency and flexibility. Among them a transmission efficiency improvement due to bandwidth control, a simple three-value bandwidth control algorithm improves the efficiency considerably with only a slight processing increase. It is concluded that the virtual path concept can provide simple, economical, flexible, and high-performance networks. >

Abstract: A link (19) connects a first controller (6) and a second controller (10) in a local area network. The second controller sends a header to the first controller (6) containing status information as to the status of an input buffer (70) at the second controller. The flow of the message data from a first buffer (44) at the first controller is controlled depending upon the status of the input buffer (70) at the second node. This prevents overflow of the input buffer (70) by controlling the rate at which new messages are placed on the transmit data link (19).

Abstract: A computer communication system including a comunication medium, a plurality of nodes coupled to the communication medium, and a transfer format selection means for selecting a format for the transfer of data between nodes. The system includes at least one default node and at least two supplemented nodes. Each node has a set of data transfer formats. A default format is included in the format set of each node. Each supplemented node has at least one supplemental format. Transfer format selection means in the form of circuitry and software provides for the selection of a data transfer format which is included in the source node format set and the destination node format set and is compatible with the communication medium. The source node includes a cache of node format sets. The source node searches for the destination node format set in the source node cache and selects a format which is common to the format sets of the source node and destination node. Our invention also includes an attention signal, a network interface and the other devices, apparatus, methods and subject matter disclosed herein.