Fast networks have made it possible to aggregate distributed CPU, memory, storage, and data to provide the potential for application performance superior to that attainable on any single system. However, achieving such performance on these metacomputing systems has proved to be difficult. Experience with the I-WAY [DFP + ss] and other metacomputing platforms demonstrates that effective application scheduling is critical to the achievement of performance for metacomputing applications. Currently, application developers develop customized application schedules to achieve performance on a metacomputer. Such application-centric schedules promote the performance of the application by evaluating system performance in terms of application resource requirements. To formalize and generalize the, as yet, ad hoc notion of application-centric scheduling emerging from the practices of metacomputing application developers [EMRP, SAR, GWP93], we are developing metacomputing scheduling agents calle...

There is a current need for scheduling policies that can leverage the performance variability of resources on multiuser clusters. We develop one solution to this problem called stochastic scheduling that utilizes a distribution of application execution performance on the target resources to determine a performance-efficient schedule. In this paper, we define a stochastic scheduling policy based on time-balancing for data parallel applications whose execution behavior can be represented as a normal distribution. Using three distributed applications on two contended platforms, we demonstrate that a stochastic scheduling policy can achieve good and predictable performance for the application as evaluated by several performance measures.

As Data Grids become more commonplace, large data sets are being replicated and distributed to multiple sites, leading to the problem of determining which replica can be accessed most efficiently. The answer to this question can depend on many factors, including physical characteristics of the resources and the load behavior on the CPUs, networks, and storage devices that are part of the end-to-end path linking possible sources and sinks.

Abstract not found

Accurate performance predictions are difficult to achieve for parallel applications executing on production distributed systems. Conventional point-valued performance parameters and prediction models are often inaccurate since they can only represent one point in a range of possible behaviors. We address this problem by allowing characteristic application and system data to be represented by a set of possible values and their probabilities, which we call stochastic values. In this paper, we give a practical methodology for using stochastic values as parameters to adaptable performance prediction models. We demonstrate their usefulness for a distributed SOR application, showing stochastic values to be more effective than single (point) values in predicting the range of application behavior that can occur during execution in production environments. 1 Introduction Parallel and distributed production platforms provide a challenging environment in which to achieve performance. The impa...

ASSIST (A Software development System based upon Integrated Skeleton Technology) is a programming environment oriented to the development of parallel and distributed high-performance applications according to a unified approach. The language and implementation features of ASSIST are a result of our long-term research in parallel programming models and tools. ASSIST is evolving towards programming environments for high-performance complex enabling platforms, especially Grids. In this paper, we show how ASSIST can act as a valid research vehicle to study, experiment and realize Grid-aware programming environments for high-performance applications. Special emphasis is put on the innovative methodologies, strategies and tools for dynamically adaptive applications, that represent the necessary step for the success of Grid platforms. First we discuss the conceptual framework for Grid-aware programming environments, based upon structured parallel programming and components technology, anticipating how ASSIST possesses the essential features required by

{vazhkuda,

This paper presents a simulation-based performance prediction framework for large scale data-intensive applications on large scale machines. Our framework consists of two components: application emulators and a suite of simulators. Application emulators provide a parameterized model of data access and computation patterns of the applications and enable changing of critical application components #input data partitioning, data declustering, processing structure, etc.# easily and #exibly. Our suite of simulators model the I#O and communication subsystems with good accuracy and execute quickly on a high-performance workstation to allow performance prediction of large scale parallel machine con#gurations. The key to e#cient simulation of very large scale con#gurations is a technique called loosely-coupled simulation where the processing structure of the application is embedded in the simulator, while preserving data dependencies and data distributions. Weevaluate our performance ...

The increasingly common practice of replicating datasets and using resources as distributed data stores in Grid environments has led to the problem of determining which replica can be accessed most efficiently. Because of diverse performance characteristics and load variations of several components in the end-to-end path linking these various locations, selecting a replica from among many requires accurate prediction information of the data transfer times between the sources and sinks. In this paper we present a prediction system that is based on combining end-to-end application throughput observations and network load variations, capturing wholesystem performance and variations in load patterns, respectively. We develop a set of regression models to derive predictions that characterize the effect of network load variations on file transfer times. We apply these techniques to the GridFTP data movement tool, part of the Globus Toolkit™, and observe performance gains of up to 10 % in prediction accuracy when compared with approaches based on past system behavior in isolation.

There is a wide range of performance models being developed for the performance evaluation of parallel and distributed systems. A performance modelling approach described in this paper is based on a layered framework of the PACE methodology. With an initial implementation system, the model described by a performance specification language, CHIPS, can provide a capability for rapid calculation of relevant performance information without sacrificing accuracy of predictions. An example of the performance evaluation of an ASCI kernel application, Sweep3D, is used to illustrate the approach. The validation results on different parallel and distributed architectures with different problem sizes show a reasonable accuracy (approximately 10% error at most) can be obtained, allows cross-platform comparisons to be easily undertaken, and has a rapid evaluation time (typically less than 2s).