A novel approach to allocating and sharing communication and computational resources in a distributed system is described. the approach, which is based on concepts drawn from microeconomics, uses algorithms that are c...
详细信息
ISBN:
(纸本)081860865X
A novel approach to allocating and sharing communication and computational resources in a distributed system is described. the approach, which is based on concepts drawn from microeconomics, uses algorithms that are competitive rather than cooperative. the effectiveness of these concepts is demonstrated by describing an economy that improves the performance of a distributed system by implementing load balancing. In this economy, competition sets prices for the resources in the system. Jobs compete for these resources by issuing bids, and the resource allocation decisions are made through auctions held by the processors. the benefits of the method include limited complexity and algorithms that are intrinsically decentralized and modular. Simulation studies show that these economies achieve substantial performance benefits.
A description is given of DAPHNE, a system of tools and run-time support routines that allow programs to be broken into parts for distributed execution on different nodes of a heterogeneous computer network. this appr...
详细信息
ISBN:
(纸本)081860865X
A description is given of DAPHNE, a system of tools and run-time support routines that allow programs to be broken into parts for distributed execution on different nodes of a heterogeneous computer network. this approach serves as a natural basis for classical network services such as remote file access or remote login while at the same time allowing arbitrary distributed applications to be written in a standard programming language. the pivot of DAPHNE is a remote procedure call mechanism that is specifically adapted to a heterogeneous environment, notably heterogeneous systems software. the current language context of DAPHNE is Modula-2. Prototype software exists for a local area network including Unix and MS-DOS systems.
Various communication aspects of locally distributed database processing are studied, using some distributed sorting and distributed hand-based join algorithms as examples. the algorithms are implemented on diskless w...
详细信息
ISBN:
(纸本)081860865X
Various communication aspects of locally distributed database processing are studied, using some distributed sorting and distributed hand-based join algorithms as examples. the algorithms are implemented on diskless workstations connected by an Ethernet network to simulate a distributed main memory system environment. this experimental testbed is described. Raw communication performance data (i.e., memory-to-memory data transfer timing) are presented. the effects of the underlying distributed operating system and the speed of the processor on the communication performance are shown. Two distributed sorting algorithms are used as examples to study the issue of concurrent transmissions of messages. distributed hash join is used as a case study for communication/local-processing tradeoff. the idea of load sharing among a number of sites to speed up the join operation is introduced.
Multiprocessor architectures using point-to-point interconnects and executing parallel algorithms require careful partitioning and allocation of data and processes. For a given interconnection, optimal allocation may ...
详细信息
ISBN:
(纸本)081860865X
Multiprocessor architectures using point-to-point interconnects and executing parallel algorithms require careful partitioning and allocation of data and processes. For a given interconnection, optimal allocation may be difficult to achieve in some cases because the data-dependent behavior of some algorithms may be impossible to predict. A mechanism called the pivot mechanism is introduced and described that controls the dynamic migration of data pages between neighboring memory modules during program execution. Flexible data migration serves the dual purpose of making algorithms less machine-specific and making possible efficient algorithm execution that is impossible to achieve by using static data allocation.
Procedure calls provide a synchronous interface to call downward through successive layers of abstraction, and remote procedure calls allow the layers to reside in different address spaces. A design is given for distr...
详细信息
ISBN:
(纸本)081860865X
Procedure calls provide a synchronous interface to call downward through successive layers of abstraction, and remote procedure calls allow the layers to reside in different address spaces. A design is given for distributed upcalls, a mechanism for propagating upcalls across address space boundaries. distributed upcalls provide a natural complement to remote procedure calls, allowing the user to arbitrarily place abstractions in the server or in the client. A server structuring system called CLAM, which is currently being used to support an extensible window management system, is presented. the CLAM system, including distributed upcalls, remote procedure call extensions to C++, dynamic loading, and basic window classes, is currently running under 4.3BSD Unix on Microvax workstations.
A distributed algorithm that uses a node-join-tree approach for the minimum-spanning-tree problem in a communication network is developed. the algorithm needs at most (2e + n(n-1)/4) messages in O(n2) time on a genera...
详细信息
ISBN:
(纸本)081860865X
A distributed algorithm that uses a node-join-tree approach for the minimum-spanning-tree problem in a communication network is developed. the algorithm needs at most (2e + n(n-1)/4) messages in O(n2) time on a general random graph. In the best case, it needs only 2e messages in O(log n) time. the parameters e and n are the number of edges and nodes, respectively. Although the worst-case performance is not better than that of tree-join-tree algorithms under some extreme cases, simulation results show that it provides better performance in most cases. the algorithm is initialized from a single node, so that there is no need to wake up all nodes at the beginning. It is less complicated than other algorithms, so that a reliable implementation is easier to achieve. the results can be used to improve the algorithms for many other problems in distributedcomputing, such as leader-election, node-counting, deadlock-resolution, and message-broadcasting.
A description is given of a data-processing structure for real-time systems with sensory feedback and a distributed computer architecture that supports it. the computer system is implemented within an autonomous mobil...
详细信息
ISBN:
(纸本)081860865X
A description is given of a data-processing structure for real-time systems with sensory feedback and a distributed computer architecture that supports it. the computer system is implemented within an autonomous mobile robot and is based on a set of independent processing modules, each of which consists of a processing core, a transport unit, and a common memory between them. Communication between modules is done in two ways. For command transportation, a direct point-to-point connection is used and transport of raw and processed sensor data is done by broadcasting. Transportation is controlled by an adjustable timetable, which is maintained by the supervisor, who slows down the driving speed and adjusts the timetable if the module does not process the given task in time. this facility allows speed of the robot to be adapted to the current processing capacity and to the complexity of the environment.
the authors present the design, implementation, and testing of the critical path analysis technique using the IPS performance measurement tool for parallel and distributed programs. they create a precedence graph of a...
详细信息
ISBN:
(纸本)081860865X
the authors present the design, implementation, and testing of the critical path analysis technique using the IPS performance measurement tool for parallel and distributed programs. they create a precedence graph of a program's activities (program activity graph) withthe data collected during the execution of a program. the critical path, the longest path in the program activity graph, represents the sequence of the program activities that take the longest time to execute. Various algorithms are developed to track the critical path from this graph. the events in this path are associated withthe entities in the source program, and the statistical results are displayed on the basis of the hierarchical structure of the IPS. the test results from the measurement of sample programs show that the knowledge of the critical path in a program's execution helps users identify performance problems and better understand the behavior of a program.
A description is given of Clouds, an operating system designed to run on a set of general-purpose computers that are connected via a medium-to-high-speed local area network. the structure of Clouds promotes transparen...
详细信息
ISBN:
(纸本)081860865X
A description is given of Clouds, an operating system designed to run on a set of general-purpose computers that are connected via a medium-to-high-speed local area network. the structure of Clouds promotes transparency, support for advanced programming paradigms, and integration of resource management, as well as a fair degree of autonomy at each site. the system structuring paradigm chosen for Clouds is an object/thread model. All instances of services, programs, and data in Clouds are encapsulated in objects. the concept of persistent objects does away withthe need for file systems, replacing it with a more powerful concept, namely, the object system. the facilities in Clouds include integration of resources by location transparency;support for various types of atomic operations, including conventional transactions;advanced support for achieving fault tolerance;and provisions for dynamic reconfiguration.
the integration of accounting functions and an account server into the distributed academic computing network operating system (DACNOS), which is an NOS built on top of a heterogeneous collection of networked computer...
详细信息
the integration of accounting functions and an account server into the distributed academic computing network operating system (DACNOS), which is an NOS built on top of a heterogeneous collection of networked computers, is described. the accounting is supported by NOS-kernel mechanisms that can be tailored to the needs of the heterogeneous servers. Accounting data is collected and stored by an account server in the NOS that manages the balances and the resource consumption data of clients. Depending on privileges, accounting data is accessible to clients, servers, and an account server administrator who may use this information for various resource and request management purposes.< >
暂无评论