We review the growing power and capability of commodity computing and communication technologies largely driven by commercial distributed information systems. these systems are built from CORBA, Microsoft9;s COM, J...
详细信息
ISBN:
(纸本)3540649522
We review the growing power and capability of commodity computing and communication technologies largely driven by commercial distributed information systems. these systems are built from CORBA, Microsoft's COM, JavaBeans, and rapidly advancing Web approaches. One can abstract these to a three-tier model with largely independent clients connected to a distributed network of servers. the latter host various services including object and relational databases and of course parallel and sequential computing. High performance can be obtained by combining concurrency at the middle server tier with optimized parallel back end services. the resultant system combines the needed performance for large-scale HPCC applications withthe rich functionality of commodity systems. Further the architecture with distinct interface, server and specialized service implementation layers, naturally allows advances in each area to be easily incorporated. We illustrate how performance can be obtained within a commodity architecture and we propose a middleware integration approach based on JWORB (Java Web Object Broker) multi-protocol server technology. Examples are given from collaborative systems, support of multidisciplinary interactions, proposed visual HPCC ComponentWare, quantum Monte Carlo and distributed interactive simulations.
Java has potential to support the development of distributed and autonomous though collaborating components in a network computing environment. To provide for the flow of objects between these components some kind of ...
详细信息
NT systems based on Ethernet networks are growing. they are finding their level within the Enterprise, where each organization has to find the right balance between solutions based on mainframe, mid-range and NT and t...
详细信息
NT systems based on Ethernet networks are growing. they are finding their level within the Enterprise, where each organization has to find the right balance between solutions based on mainframe, mid-range and NT and the communications between them. the question is whether the performance of the resultant distributed systems can be managed effectively or not. Withthe arrival of Windows NT Enterprise Edition and its support for clustering technology, NT servers now appear capable of acting as larger mid-range systems for specific dedicated applications exploiting Client/Server technology. Any approach to the performance management for such client server systems has to cope withthese new technologies as well as existing applications on mainframes and UNIX-based machines. this paper discusses the performance management implications of the growth in NT & Ethernet, and the reported improvements in NT v5 alongside the parallel growth in LAN Ethernets. It describes how analytical modelling has been extended to incorporate networks and so allow meta-modelling of the total distributed client/server system to provide Service Level Management in meaningful terms of end-to-end Response Times. It concludes with a Case Study demonstrating how proactive performance management requires meta-models to predict potential bottlenecks.
Gigabit network technologies have made it possible to combine workstations into a distributed, massively-parallel computer system. Middleware, such as distributed shared objects (DSO), attempts to improve programmabil...
详细信息
ISBN:
(纸本)0818678135
Gigabit network technologies have made it possible to combine workstations into a distributed, massively-parallel computer system. Middleware, such as distributed shared objects (DSO), attempts to improve programmability of such systems, by providing globally accessible 'object' abstractions, Researchers have developed consistency protocols for replicated 'memory' objects. these protocols are well suited to scientific applications but less suited to multimedia or groupware applications. this paper addresses the state sharing needs of complex distributedapplications with (1) high-frequency symmetric accesses to shared objects, (2) unpredictable and limited locality of accesses, (3) dynamically changing sharing behavior, and (4) potential data races. We show that a DSO system exploiting application-level temporal and spatial constraints an shared objects can outperform shared object protocols which do not exploit application-level constraints. We compare our S(emantic) DSO against entry consistency using a sample application having the four properties mentioned above.
the paper presents a survey of recently developed parallel simulation technologies for stochastic differential systems with lumped and distributed parameters. Some applications of experimental parallel software to sta...
详细信息
the proceedings contain 44 papers. the special focus in this conference is on Automatic Data Distribution and Locality Enhancement. the topics include: Cross-loop reuse analysis and its application to cache optimizati...
ISBN:
(纸本)3540630910
the proceedings contain 44 papers. the special focus in this conference is on Automatic Data Distribution and Locality Enhancement. the topics include: Cross-loop reuse analysis and its application to cache optimizations;locality analysis for distributed shared-memory multiprocessors;data distribution and loop parallelization for shared-memory multiprocessors;data localization using loop aligned decomposition for macro-dataflow processing;exploiting monotone convergence functions in parallel programs;exact versus approximate array region analyses;context-sensitive interprocedural analysis in the presence of dynamic aliasing;initial results for glacial variable analysis;compiler algorithms on if-conversion, speculative predicate assignment and predicated code optimizations;determining asynchronous pipeline execution times;compiler techniques for concurrent multithreading with hardware speculation support;resource-directed loop pipelining;integrating program optimizations and transformations withthe scheduling of instruction level parallelism;parametric computation of margins and of minimum cumulative register lifetime dates;global register allocation based on graph fusion;automatic parallelization for non-cache coherent multiprocessors;eliminating lock overhead in automatically parallelized object-based programs;optimal reordering and mapping of a class of nested-loops for parallel execution;communication-minimal tiling of uniform dependence loops;communication-minimal partitioning of parallel loops and data arrays for cache-coherent distributed-memory multiprocessors and resource-based communication placement analysis.
the most successful compilation techniques for distributed memory architectures are based on static analysis of the memory accesses. Loop iterations with similar comportment on the parallel memories are combined in or...
详细信息
Panda is a high-performance library for accessing large multidimensional array data on secondary storage of parallel platforms and networks of workstations. When using Panda as the I/O component of a scientific applic...
详细信息
ISBN:
(纸本)0818679530
Panda is a high-performance library for accessing large multidimensional array data on secondary storage of parallel platforms and networks of workstations. When using Panda as the I/O component of a scientific application, H3 expresso, on the IBM SP2 at Cornell theory Center, we found that some nodes are more powerful with respect to I/O than others, requiring the introduction of load balancing techniques to maintain high performance. We expect that heterogeneity will also be a big issue for DBMSs or parallel NO libraries designed for scientific applications running on networks of workstations, and the methods of allocating data to servers in these environments will need to be upgraded to fake heterogeneity into account, while still allowing users to exert control over data layout. We propose such an approach to load balancing, under which we respect the user's choice of high-level disk layout, bur introduce automatic subchunking. the use of subchunks allows us to divide the very large chunks typically specified by the user's disk layout into more manageable-size units that can be allocated to I/O nodes in a manner that fairly distributes the load. We also present two techniques for allocating subchunks to nodes, static and dynamic, and evaluate their performance on the SP2.
the proceedings contain 18 papers. the special focus in this conference is on Modeling Techniques and Tools for Computer Performance Evaluation. the topics include: A performability modeling environment tool;dependabi...
ISBN:
(纸本)9783540631019
the proceedings contain 18 papers. the special focus in this conference is on Modeling Techniques and Tools for Computer Performance Evaluation. the topics include: A performability modeling environment tool;dependability evaluation and the optimization of performability;design and implementation of a network computing platform using JAVA;storage alternatives for large structured state spaces;an efficient disk-based tool for solving very large markov models;efficient transient overload tests for real-time systems;towards an analytical tool for performance modeling of ATM networks by decomposition;an embedded network simulator to support network protocols’ development;synchronized two-way voice simulation for internet phone performance analysis and evaluation;processes as language-oriented building blocks of stochastic petri nets;measurement tools and modeling techniques for evaluating WEB server performance;workload characterization of input/output intensive parallelapplications;interval based workload characterization for distributed systems;bounding the loss rates in a multistage ATM switch;simple bounds for queues fed by markovian sources and on queue length moments in fork and join queuing networks with general service times.
this paper describes the static and dynamic task allocation tools in PVM environment for distributed memory parallel systems. For the static mapping the objective function is used to evaluate the optimality of the all...
详细信息
暂无评论