Showing papers on "Server published in 1985"

Distributed process groups in the V Kernel

Abstract: The V kernel supports an abstraction of processes, with operations for interprocess communication, process management, and memory management. This abstraction is used as a software base for constructing distributed systems. As a distributed kernel, the V kernel makes intermachine boundaries largely transparent.In this environment of many cooperating processes on different machines, there are many logical groups of processes. Examples include the group of tile servers, a group of processes executing a particular job, and a group of processes executing a distributed parallel computation.In this paper we describe the extension of the V kernel to support process groups. Operations on groups include group interprocess communication, which provides an application-level abstraction of network multicast. Aspects of the implementation and performance, and initial experience with applications are discussed.

Blocking when service is required from several facilities simultaneously

Abstract: This paper analyzes a mathematical model of a blocking system with simultaneous resource possession. There are several multiserver service facilities without extra waiting space at which several classes of customers arrive in independent Poisson processes. Each customer requests service from one server in each facility in a subset of the service facilities, with the subset depending on the customer class. If service can be provided immediately upon arrival at all required facilities, then service begins and all servers assigned to the customer start and finish together. Otherwise, the attempt is blocked (lost without generating retrials). The problem is to determine the blocking probability for each customer class. An exact expression is available, but it is complicated. Hence, this paper investigates approximation schemes.

Blocking when service is required from several facilities simultaneously

Abstract: This paper analyzes a mathematical model of a blocking system with simultaneous resource possession. There are several multiserver service facilities without extra waiting space at which several classes of customers arrive in independent Poisson processes. Each customer requests service from one server in each facility in a subset of the service facilities, with the subset depending on the customer class. If service can be provided immediately upon arrival at all required facilities, then service begins and all servers assigned to the customer start and finish together. Otherwise, the attempt is blocked (lost without generating retrials). The problem is to determine the blocking probability for each customer class. An exact expression is available, but it is complicated. Hence, this paper investigates approximation schemes.

The Optimality of Unbalancing Both Workloads and Machine Group Sizes in Closed Queueing Networks of Multiserver Queues

Abstract: Using a closed queueing network model, we explore the consequences of varying workloads among multiserver queues that may be of unequal size. In addition, we solve the problem of assigning servers of similar types to the queues in the network to maximize expected throughput. We show that 1 unbalanced configurations of assigned servers are superior to balanced ones, and 2 unbalanced workloads are better than balanced ones. We find that there can be significant differences in system throughput from balanced versus unbalanced configurations/workloads. Finally, we discuss applications to planning problems of flexible manufacturing systems.

Method of propagating resource information in a computer network

Abstract: METHOD OF PROPAGATING RESOURCE INFORMATION IN A COMPUTER NETWORK Abstract A method of propagating resource information among computers of a computer network in a fully distributed (or decentralized) fashion. A solicit message from a client one of the computers is transmitted to one or more prescribed server ones of the computers each time the client computer is made operative in the network. In response to the solicit message, each of the prescribed server computers determines if it is available as a resource to the client computer. The server then transmits a positive response message or a negative response message to the client computer if the server computer is available or unavailable, respectively. In addition, when a server computer becomes available as a resource to one or more client computers, it transmits an advertisement message to the prospective client or clients.

Approximating the Equilibrium Behavior of Multi-Server Loss Systems

Abstract: A procedure is given for approximating the equilibrium behavior of multi-server loss systems having distinguishable servers and multiple customer types under light to moderate traffic intensity. Calls are assumed to arrive according to independent Poisson processes for each customer type. General service time distributions which may depend on both server and customer type are allowed. For N servers, the approximation routine consists of a simple iterative procedure in N nonlinear equations. Computational experience with respect to the convergence properties of the iteration and the accuracy of the approximation is given.

Industrial processing and manufacturing systems

Abstract: An automated industrial processing or manufacturing installation including a series of computer controlled facilities such as machine tools, workpiece and tool conveyors, and storage devices is controlled by a computer system wherein the control software is structured as a series of elemental function modules which intercommunicate by way of a high level communication interface. The functions are structured so as to fall into one of the categories job-route sequencer, activity managers, controllers and servers with a route-sequencing module controlling the sequence in which various activities are performed in relation to a processing or manufacturing job being carried out by the installation, each activity manager distributing activities to a group of controllers for which it is responsible, and the controllers being responsible for executing activities as and when they are received from the associated manager, the servers meanwhile providing general services for the other elements of the system.

Prefix Tables: A Simple Mechanism for Locating Files in a Distributed System.

Abstract: : Prefix tables provide a mechanism for locating files in a system whose storage is distributed among many servers. The result is a single file system hierarchy visible uniformly and transparently to all clients. Each client of the filesystem maintains a local prefix table that identifies the server for a file based on the initial part of the file name. Prefix tables are built and modified using a simple broadcast protocol that is flexible enough to allow dynamic server reconfiguration and a simple form of replication.

A Queueing System with General-Use and Limited-Use Servers

Abstract: We consider a queueing system with two types of servers and two types of customers. General-use servers can provide service to either customer type while limited-use servers can be used only for one of the two. Though the apparent Markovian state space of this system is five-dimensional, we show that an aggregation results in an exact two-dimensional representation that is also Markovian. Matrix geometric theory is used to obtain approximations for the mean delay times and other measures of interest for each customer type. We illustrate the methodology by applying it to analyze a token discount policy used by the Triborough Bridge and Tunnel Authority.

Experiments in optimal control of a flexible arm

Abstract: ©1985 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.

Sequencing Problems in Two-Server Systems

Abstract: We analyze a service system in which two identical servers move one at a time along a linear array of N positions. Requests for service, each designating one of the N positions, join a first-in-first-out queue, where processing of the nth request does not begin until processing of the (n − 1)th requested is completed. Processing the nth request entails determining which server to move, moving this server to the requested position, and then performing the service. Several potential applications of the model are mentioned, the most notable being the design of computer storage systems with multiple access devices. Within a simple probability model we compare server-selection policies in terms of the equilibrium expected distance a server is moved in processing a request. Distance is measured under two regimes, both assigning a unit distance between adjacent positions. In the first, the positions are thought of as equally spaced along a line (interval), and in the second, equally spaced along a circle (positi...

Algorithms for synchronizing network clocks

Abstract: This RFC discussed clock synchronization algorithms for the ARPA- Internet community, and requests discussion and suggestions for improvements. The recent interest within the Internet community in determining accurate time from a set of mutually suspicious network clocks has been prompted by several occasions in which errors were found in usually reliable, accurate clock servers after thunderstorms which disrupted their power supply. To these sources of error should be added those due to malfunctioning hardware, defective software and operator mistakes, as well as random errors in the mechanism used to set and synchronize clocks. This report suggests a stochastic model and algorithms for computing a good estimator from time-offset samples measured between clocks connected via network links. Included in this report are descriptions of certain experiments which give an indication of the effectiveness of the algorithms.

The slow server problem

Abstract: The problem is what to do with a slow server in a service facility which has fast and slow servers. Should the slow server be used to render service, or is it better not to use it at all? Simple models for answering this question are formulated and studied.

Solving the combined zoning and location problem for several emergency units.

Abstract: This paper discusses methods for the solution of the combined zoning and location problem for several emergency units This is a stochastic optimization problem The stochastic aspects include ordered preference of servers A constraint of equal workload for the servers is also dealt with Non-linear programming techniques are adapted for its solution

Client/Server Model of Distributed Processing

Abstract: The client/server model of distributed computing is a structuring concept which can be identified in many different distributed systems. Clients and servers are active modules communicating through messages. Systems composed of server machines and personal workstations as their clients exemplify this concept. The paper focuses mainly on two issues: server design and communication protocols. File servers are used as an example in the server design overview. The communication issues are discussed in the framework of the OSI Reference Model and protocols.

Non-Discretionary Controls for Inter-Organization Networks

Abstract: This paper describes a conceptual model for implementing usage controls in Inter-Organization Networks (IONs). After describing security requirements in networks that cross organization boundries, I suggest how traditional, non-discretionary controls can be adapted to support usage control in IONs.

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.

Optimal Location-Relocation Decisions on Stochastic Networks

Abstract: This paper deals with stochastic transportation networks in which links' travel times are random variables. Transitions among states of the network (each state differs from any other state by at least one link with different travel time) are assumed Markovian. The problem is to find an optimal location-relocation decision for p mobile servers in such a network. The criterion is to minimize the long-term expected average cost in units of time per period taking into account both response times and relocation costs. The paper generalizes its predecessor (Berman, O., A. R. Odoni. 1982. Locating servers on a network with Markovian properties. Networks 12 73–86.), by permitting the chosen set of locations in a given state of the network to depend on the former state of the network. Markovian decision theory is employed with the problem's special structure used to reduce the complexity of the resulting algorithm.

Customer routing to parallel servers with different rates

Abstract: We consider the problem of routing customers to parallel servers having different rates. There are no buffers in the system. Each customer must be rooted to a server immediately upon its arrival and if the server to which it is routed is occupied, then the customer is aborted. The aim is to maximize throughput (the proportion of customers which are successfully routed to a free server), when the routing must be done without knowing which servers are occupied and which are free. An upper bound on the throughput is found for a general renewal arrival process and geometric service times. Furthermore, a new routing policy, the golden ratio policy, is suggested and shown to approach a limit which is within at least 98.4 percent of the upper bound. The golden ratio policy is a generalization of the round robin policy, when the service rates of the servers are different.

Efficient Computing of Relational Algebraic Primitives in a Database Machine Architecture

Abstract: Algorithms are described and analyzed for the efficient evaluation of the primitive operators of a relational algebra on a database machine architecture. The architecture contains a RAM divided up into partitions, each partition having a separate server. Tuples from the first operand relation are stored in linked lists in the partitioned RAM via bit and pointer arrays based on hashed column values. Each tuple in the second operand relation is either deposited in the corresponding partition server queue based on the hashed value, or it is ignored. Cross-referencing involved in the operations is removed without performing a sorting operation, which significantly reduces the time complexity. A procedure is also presented for computing the optimal number of partition servers for different applications.

Sequencing two servers on a sphere

Abstract: We analyze a service system in which two servers move independently on the surface of the n-dimensional sphere. Requests for service arrive independently and uniformly over the surface. The ith request is to be served completely before service of the (i+1)st request begins. In an earlier paper the authors showed that the nearer server (NS) policy is optimal among all server selection policies in the sense that it minimizes the equilibrium expected angular distance E(D) which a server moves to process a request. In the present paper we obtain for all n the integral equation satisfied by the equilibrium measure for the angular distance Φ between servers under the NS policy. The equation is solved numerically. We also show that E(D)+E(Φ) = π, and unse this to compute E(D).

Partitioning of function in a distributed graphics system (computer networks, operating, window)

Abstract: Although recent advances in graphics workstations promise much computing power for the future needs of researchers, traditional approaches to software organization waste much of this power. Most systems treat the workstation as either a fixed-function terminal or a self-contained personal computer; these roles have limitations that can be overcome by considering the workstation a multi-function component of a distributed system. Traditional standard graphics packages and object-oriented window systems offer important functionality, but a third approach, virtual terminal management systems, is more appropriate for a distributed operating system. The Stanford Distributed Systems Group has implemented such a distributed system for graphics workstations, organized as a collection of servers providing services to clients. Major issues are how to partition functions between the server and its clients, and physically partition the server. In particular, the service that displays graphical objects is called the Virtual Graphics Terminal Service (VGTS). The VGTS architecture is described, as well as a prototype implementation. This thesis discusses the trade-offs involved in partitioning of function in a distributed graphics system. Performance is one important property traded for advanced functionality or decreased cost. To provide adequate performance in a distributed system, communication costs should be kept low, as well as the frequency of the communication. By providing modeling as well as viewing facilities, the VGTS reduces the communication required between applications and the service. Measurements verify that performance is insensitive to network bandwidth, but depends heavily on CPU speed and protocol characteristics. Using structure provides important speed improvements in some cases, but other basic factors such as inner loop optimization and proper batching of requests make even larger differences. Finally, conclusions are drawn regarding the partitioning approaches taken in the VGTS. The VGTS is suitable for a large class of applications that perform graphics as an aid to user interface, and is portable to a wide range of powerful workstations. Moreover, the VGTS can be used as a basis for further research on many open questions in distributed systems.

An approximation of a loss system with two heterogeneous types of calls

Abstract: This paper presents an efficient and fairly accurate approximation method for a loss system having two types of calls and three types of server groups. In this model, two types of calls arrive at the individual server group for each call type. If an arriving call fmds all the servers of a corresponding group busy, the call overflows to a common server group. If the overflowed call again fmds all the common servers busy, the call is lost. This model is constructed around a telephone based ticket reservation system, as well as around integrated services digital net. works which will be introduced in the near future. By approximating the overflow process using an interrupted Poisson process (lPP), the main problem is reduced to solving a queueing model with a common server group having two independent lPP inputs with different mean service times. The proposed method makes it possible to compute loss probabilities in a far shorter time as compared with the exact method employing the lumping method. The accuracy of the results is shown to be sufficiently good for practical use carrying out numerical experiments.

A testbed for information retrieval research: the Utah retrieval system architecture

Abstract: The Utah Retrieval System Architecture provides an excellent testbed for the development and testing of new algorithms or techniques for information retrieval. URSA™ is a message-based structure capable of running on a variety of system configurations, ranging from a single mainframe processor to a system distributed across a number of dissimilar processors. It can readily support a variety of specialized backend processors, such as high-speed search engines.The architecture divides the components of a text retrieval system into two classes: servers and clients. A triple of servers (index, search, and document access) for each database provide the capabilities normally associated with a retrieval system. Possible clients for these servers include a window-based user interface, whose query language can be easily modified, a connection to a mainframe host processor, or Al-based query modification programs that wish to use the database.Any module in the system can be replaced by a new module using a different algorithm as long as the new module complies with the message formats for that function. In fact, with some care this module switch can occur while the system is running, without affecting the users. A monitor program collects statistics on all system messages, giving information regarding query complexity, processing time for each module, queueing times, and bandwidths between every module.This paper discusses the background of URSA and its structure, with particular emphasis on the features that make it a good testbed for information retrieval techniques.

Arca:A Local Network File Server

Abstract: Distributed systems can be built to a client-server model in which one of the important servers is the file server, consisting of at least one processor and disk drive, attached to the network with the sole purpose of storing information on behalf of its clients. In the Department of Computing at Lancaster University a file server which is connected to both of their networks has been designed and built: an Ethernet-type network (called Strathnet, of local design) and a Cambridge Ring type network. The Arca File Server is based on work started at Cambridge University. This paper describes the background of file servers for local networks and the design and implementation of the system.