Showing papers on "Shared resource published in 1986"

Method for multiple programs management within a network having a server computer and a plurality of remote computers

Abstract: The procedure which allows users of a computer system comprising a plurality of computers connected in a local area network to share both file resources and application programs on the local area network without modification to existing programs which were designed to run in a non-network environment is disclosed. The local area network comprises a server computer and at least one remote computer. Starting the network comprises an initial program load of the operating system for each of the computers, loading the local area network control program, and then loading a hypervisor or "node enabler" program. At each of the remote computers, a request to load a program or access a data file is converted by the "node enabler" to a file sharing and record locking protocol message with is transmitted to the server computer. The server computer stores a program matrix with entries indicating which programs can be run on the network without conflicts with other systems including the server computer. The server computer also maintains a list of currently running programs and accessed data files. By comparing the remote computer request with the program matrix and the list of currently running programs and accessed data files, a decision to grant a remote computer's request is made. In addition, by recording a unique identification number for each remote computer signed on to the network at the server computer, control of access to licensed programs is maintained.

On maintaining dynamic information in a concurrent environment

Abstract: This paper considers the amount of cooperation required for independent asynchronous processes to share a simple dynamic data structure. We present a scheme for designing efficient concurrent algorithms to add and remove elements from a shared pool of elements. The efficiency is measured mainly by the number of non-local operations that a process may have to make. Non-local operations may involve writing into a shared variable, locking, or sending a message, hence they introduce interference (or require cooperation). We derive upper and lower bounds on the interference in the worst case. Applications to distributed computation are also discussed.

A model of file server performance for a heterogeneous distributed system

Abstract: In this paper, we study the performance characteristics of a client-server style distributed system by a queueing network model. The system being modeled was based on an experimental distributed system currently being prototyped. As a specific detailed case study, we have evaluated the performance of a file server. A file server is a key component to achieve the data sharing necessary in a distributed system. The file server is probably the most heavily used resource of the distributed system and as a result, its performance is critical to the success of the system. The primary goal of the study was to identify and quantify the effects of the bottlenecks in the server and assess design alternatives to enhance performance. In particular, the improvements due to file caching, outboarding transport functions and the effect of multiple network interfaces were studied. We model in detail the various stages of processing of a request to a file server. Parameterization of the model was achieved by initial measurements made on the prototype system. An important characteristic modeled was the buffer overflow from the network interfaces. This required the approximate solution of a non-product-form queueing network.

Resource Sharing Interconnection Networks in Multiprocessors.

Abstract: Circuit-switched interconnection networks for resource sharing in multiprocessors are studied. Resource scheduling in systems with such an interconnection network entails the efficient search for a mapping from requesting processors to free resources such that circuit blockages in the network are minimized and resources are maximally used. The optimal mapping is obtained by transforming the scheduling problems into various network flow problems to which existing algorithms can be applied. A distributed architecture to realize a maximum flow algorithm using token propagations is described. The method is applicable to any general loop-free network configuration in which the requesting processors and free resources can be partitioned into two disjoint subsets. >

Exclusive control system for multiprocessor system

Abstract: PURPOSE:To prevent the lock-out of an overall system due to the system down of an arithmetic processor during access to a shared resource by resetting a flip-flop for exclusive control when it is set over the prescribed time CONSTITUTION:When the arithmetic processor 2-1 issues the private request to a shared resource control unit 3 in the state where the shared resource 4 is not monopolized by any arithmetic processor, the exclusive control flip-flop 31 in the shared resource control unit 3 is set, and simultaneously a timer counter 32 is activated and starts counting the time while the flip-flop 31 is set Accordingly the arithmetic processor 2-1 uses exclusively the shared resource 4 When it is broken in this state, the set state of the flip-flop 31 continues, and the timer counter 32 is timer-over When the counted time of the timer counter 32 reaches the prescribed value, an exclusive control flip-flop setting circuit 33 resets the flip-flop 31, and other arithmetic processors are caused to use the shared resource 4

Method for distributed processing in a computer network

Abstract: A METHOD FOR DISTRIBUTED PROCESSING IN A COMPUTER NETWORK Abstract A method of performing remote process execution in a computer network is disclosed. A remote process execution request, including an identification of a process to be executed, is transmitted from a requesting one of the computers (CLIENT) to a serving one of the computers. At both the requesting and the serving computers, a file addressing structure is established, so that a file reference by the remote process at the serving computer refers to a file located at the requesting computer. The remote process is activated at the serving computer in response to the remote process execution request. In response to a file reference by the remote process, the file is automatically retrieved from the requesting computer in accordance with the addressing structure. The method includes a technique by means of which files may be selectively and automatically accessed from either the requesting computer or the serving computer, as desired by a user. A computer administrator may selectively override the file accessing instructions, if desired, thereby providing the ability to maintain file security for certain files.

Method of arbitrating access to a shared resource among a plurality of uniquely identified processors

Abstract: A data communication method is disclosed which can be used to implement an unbiased arbitration system. This method can be used to identify one of N<=2**m processors (U-0,U-1,U-15) requesting access to a shared resource (102). Each of the processors is designated by a unique m-bit identifier. This method can also be used to enable the host (101) to detect the processing states of said N processors in no more than m interchanges between the host and the processors.

Distributed Resource Sharing in Computer Networks

Abstract: In this paper we address a resource sharing problem that arises in distributed systems. The problem is to provide a mechanism whereby a node can obtain, in a distributed network, control of a set of resources that are necessary in performing a certain task. The mechanism has to provide a "reasonable guarantee" that if sufficient resources are available, the node will be able to obtain them. We define what such a "reasonable guarantee" means in a dynamic environment with resources constantly being added or removed from the pool of available resources. We then propose a distributed algorithm and show that it provides such a guarantee. Finally, we suggest extensions to the basic mechanism to improve performance.

Shared resource updating system on storage device

Abstract: PURPOSE:To execute quickly updating processing of shared resources by processing the second updating processing which has been generated in the course of the first updating processing, in parallel to the first updating on a copy of the shared resources, and validating it after confirming that there is no overlap of the updating. CONSTITUTION:When the second updating request is generated from a terminal equipment 2 in the course of updating processing of a file 4 on a magnetic disk 4 of shared resources by the first updating processing request from a terminal equipment 1, a central processor 3 makes a copy of the file 5 as a file 7 on a main storage device 6. Thereafter, the central processor 3 executes the second updating processing to the file 7 in parallel to the first updating processing to the file 5. After the updating has been ended, updating map tables 8, 9 which have recorded the updating part are checked and collated, and unless the updating parts are overlapped, the file 5 is updated by the updating data of the file 7.

Automation in the university of london

Abstract: The University of London's far‐reaching plan for a new generation of library automation has become a casualty of the UGC's latest round of cuts in funding for universities. A strategy had been developed over recent years to replace the four existing shared systems and to introduce a five to seven year phase of expansion. The integration of all library functions, the provision of online access, and emphasis on resource sharing and a rational approach to library provision in London for the remainder of the century were the main objectives. A distributed network was planned to link all the diverse libraries of the University in a common system, affording online access to shared bibliographic and, where appropriate, shared borrower data. Considerable interest in the scheme was being shown by libraries both at home and abroad.

A simple file server for a terminal support network environment

Abstract: Terminal Support Networks (TSNs) are becoming popular, particularly in the university environment, to interconnect microcomputer systems and also to use them as terminals to access mainframe computers. In such environments it seems attractive to share the disk space available on a large time sharing system. This paper describes the design and implementation of a file server which is simple to implement and user friendly. A broadband Local Area Network (LAN) provides the TSN environment. The main emphasis is on the communication aspects between the client and the file server, and the automatic recovery provided by the system after a system start up from a failure.

Processor failure recovery for a resource sharing algorithm

Abstract: With the increase in popularity of distributed computer systems, the reliability of the system as a whole is becoming more important. A recently published combined resource sharing algorithm showed how the atomic operations required for resource management in a closely coupled multiprocessor system could be provided. The paper describes a recovery system that may be incorporated within the earlier algorithm to enable continued and correct operation of the system despite the failure of one or more component processors. A distributed simulation of the recovery mechanism is described and results from simulation runs are presented.

Process-control task scheduling: a deterministic approach with shared resources

Abstract: A strong need has been observed for task sheduling models and algorithms, to be used in process-control operating systems, that would be characterized by: (1) Explicit use of trade-off criteria combining technological performance of the plant with critical time and implementation cost aspects of the control actions, that allow designers to straightforwardly consider both sides of computer control system performance; (2) Extensive resource sharing which allows the computer control system to operate at lower resource costs, with specific constraints corresponding with real technological and operational restrictions. This paper presents a new flexible approach to process-control task scheduling that meets the above requirements. It proved to be particularily useful in memory and re-entrant code management. A general model of the allocation problem with examples is presented. The corresponding optimization problem is formulated and heuristics useful in the construction of scheduling algorithms are given.

An Integrated Bibliographic Information System: Concept and Application for Resource Sharing in Special Libraries.

Abstract: : The Defense Dep. Scientific and Technical Information (STI) network is composed of over 200 technical libraries and information centers tied together by the Defense Technical Information Center (DTIC). As the DOD clearinghouse for STI, DTIC seeks to improve the flow of information throughout the STI network by promoting shared cataloging and integrated retrieval systems. Through sponsorship of the Local Automation Model project, DTIC will offer libraries and information centers a fully resident computer system supporting local collection cataloging, retrieval, and circulation management and control. What sets the system apart from others is the additional capability offered by an intelligent gateway: the ability to interconnect and share information with geographically remote, heterogenous computers and databases. By combining local collection management capabilities with an intelligent gateway, users simultaneously access and search diverse bibliographic resources - a local catalog, closed community resources, and commercial databases. Libraries and information centers can readily share bibliographic citations, thus reducing duplication of intellectual and manual effort associated with acquiring new STI holdings. The prototype system will be operational during 1986, providing the opportunity to demonstrate library automation concepts that will shorten the time required to acquire and disseminate STI.