arXiv subjects

Shantenu Jha

Publications and source records attributed to Shantenu Jha.

104 records · Page 6Linked to original sources

arXiv2016

Introducing Distributed Dynamic Data-intensive (D3) Science: Understanding Applications and Infrastructure

A common feature across many science and engineering applications is the amount and diversity of data and computation that must be integrated to yield insights. Data sets are growing larger and becoming distributed; and their location, availability and properties are often time-dependent. Collectively, these characteristics give rise to dynamic distributed data-intensive applications. While "static" data applications have received significant attention, the characteristics, requirements, and software systems for the analysis of large volumes of dynamic, distributed data, and data-intensive applications have received relatively less attention. This paper surveys several representative dynamic distributed data-intensive application scenarios, provides a common conceptual framework to understand them, and examines the infrastructure used in support of applications.

cs.DC↗

arXiv2016

Ensemble Toolkit: Scalable and Flexible Execution of Ensembles of Tasks

There are many science applications that require scalable task-level parallelism and support for flexible execution and coupling of ensembles of simulations. Most high-performance system software and middleware, however, are designed to support the execution and optimization of single tasks. Motivated by the missing capabilities of these computing systems and the increasing importance of task-level parallelism, we introduce the Ensemble toolkit which has the following application development features: (i) abstractions that enable the expression of ensembles as primary entities, and (ii) support for ensemble-based execution patterns that capture the majority of application scenarios. Ensemble toolkit uses a scalable pilot-based runtime system that decouples workload execution and resource management details from the expression of the application, and enables the efficient and dynamic execution of ensembles on heterogeneous computing resources. We investigate three execution patterns and characterize the scalability and overhead of Ensemble toolkit for these patterns. We investigate scaling properties for up to O(1000) concurrent ensembles and O(1000) cores and find linear weak and strong scaling behaviour.

cs.DC↗

arXiv2016

ExTASY: Scalable and Flexible Coupling of MD Simulations and Advanced Sampling Techniques

For many macromolecular systems the accurate sampling of the relevant regions on the potential energy surface cannot be obtained by a single, long Molecular Dynamics (MD) trajectory. New approaches are required to promote more efficient sampling. We present the design and implementation of the Extensible Toolkit for Advanced Sampling and analYsis (ExTASY) for building and executing advanced sampling workflows on HPC systems. ExTASY provides Python based "templated scripts" that interface to an interoperable and high-performance pilot-based run time system, which abstracts the complexity of managing multiple simulations. ExTASY supports the use of existing highly-optimised parallel MD code and their coupling to analysis tools based upon collective coordinates which do not require a priori knowledge of the system to bias. We describe two workflows which both couple large "ensembles" of relatively short MD simulations with analysis tools to automatically analyse the generated trajectories and identify molecular conformational structures that will be used on-the-fly as new starting points for further "simulation-analysis" iterations. One of the workflows leverages the Locally Scaled Diffusion Maps technique; the other makes use of Complementary Coordinates techniques to enhance sampling and generate start-points for the next generation of MD simulations. We show that the ExTASY tools have been deployed on a range of HPC systems including ARCHER (Cray CX30), Blue Waters (Cray XE6/XK7), and Stampede (Linux cluster), and that good strong scaling can be obtained up to 1000s of MD simulations, independent of the size of each simulation. We discuss how ExTASY can be easily extended or modified by end-users to build their own workflows, and ongoing work to improve the usability and robustness of ExTASY.

cs.CE↗

arXiv2016

A Comprehensive Perspective on Pilot-Job Systems

Pilot-Job systems play an important role in supporting distributed scientific computing. They are used to consume more than 700 million CPU hours a year by the Open Science Grid communities, and by processing up to 1 million jobs a day for the ATLAS experiment on the Worldwide LHC Computing Grid. With the increasing importance of task-level parallelism in high-performance computing, Pilot-Job systems are also witnessing an adoption beyond traditional domains. Notwithstanding the growing impact on scientific research, there is no agreement upon a definition of Pilot-Job system and no clear understanding of the underlying abstraction and paradigm. Pilot-Job implementations have proliferated with no shared best practices or open interfaces and little interoperability. Ultimately, this is hindering the realization of the full impact of Pilot-Jobs by limiting their robustness, portability, and maintainability. This paper offers a comprehensive analysis of Pilot-Job systems critically assessing their motivations, evolution, properties, and implementation. The three main contributions of this paper are: (i) an analysis of the motivations and evolution of Pilot-Job systems; (ii) an outline of the Pilot abstraction, its distinguishing logical components and functionalities, its terminology, and its architecture pattern; and (iii) the description of core and auxiliary properties of Pilot-Jobs systems and the analysis of seven exemplar Pilot-Job implementations. Together, these contributions illustrate the Pilot paradigm, its generality, and how it helps to address some challenges in distributed scientific computing.

cs.DC↗

arXiv2016

Integrating Abstractions to Enhance the Execution of Distributed Applications

One of the factors that limits the scale, performance, and sophistication of distributed applications is the difficulty of concurrently executing them on multiple distributed computing resources. In part, this is due to a poor understanding of the general properties and performance of the coupling between applications and dynamic resources. This paper addresses this issue by integrating abstractions representing distributed applications, resources, and execution processes into a pilot-based middleware. The middleware provides a platform that can specify distributed applications, execute them on multiple resource and for different configurations, and is instrumented to support investigative analysis. We analyzed the execution of distributed applications using experiments that measure the benefits of using multiple resources, the late-binding of scheduling decisions, and the use of backfill scheduling.

cs.DC↗

arXiv2016

Synapse: Synthetic Application Profiler and Emulator

We introduce Synapse motivated by the needs to estimate and emulate workload execution characteristics on high-performance and distributed heterogeneous resources. Synapse has a platform independent application profiler, and the ability to emulate profiled workloads on a variety of heterogeneous resources. Synapse is used as a proxy application (or "representative application") for real workloads, with the added advantage that it can be tuned at arbitrary levels of granularity in ways that are simply not possible using real applications. Experiments show that automated profiling using Synapse represents application characteristics with high fidelity. Emulation using Synapse can reproduce the application behavior in the original runtime environment, as well as reproducing properties when used in a different run-time environments.

cs.DC↗

arXiv2016

Hadoop on HPC: Integrating Hadoop and Pilot-based Dynamic Resource Management

High-performance computing platforms such as supercomputers have traditionally been designed to meet the compute demands of scientific applications. Consequently, they have been architected as producers and not consumers of data. The Apache Hadoop ecosystem has evolved to meet the requirements of data processing applications and has addressed many of the limitations of HPC platforms. There exist a class of scientific applications however, that need the collective capabilities of traditional high-performance computing environments and the Apache Hadoop ecosystem. For example, the scientific domains of bio-molecular dynamics, genomics and network science need to couple traditional computing with Hadoop/Spark based analysis. We investigate the critical question of how to present the capabilities of both computing environments to such scientific applications. Whereas this questions needs answers at multiple levels, we focus on the design of resource management middleware that might support the needs of both. We propose extensions to the Pilot-Abstraction to provide a unifying resource management layer. This is an important step that allows applications to integrate HPC stages (e.g. simulations) to data analytics. Many supercomputing centers have started to officially support Hadoop environments, either in a dedicated environment or in hybrid deployments using tools such as myHadoop. This typically involves many intrinsic, environment-specific details that need to be mastered, and often swamp conceptual issues like: How best to couple HPC and Hadoop application stages? How to explore runtime trade-offs (data localities vs. data movement)? This paper provides both conceptual understanding and practical solutions to the integrated use of HPC and Hadoop environments.

cs.DC↗

arXiv2016

RepEx: A Flexible Framework for Scalable Replica Exchange Molecular Dynamics Simulations

Replica Exchange (RE) simulations have emerged as an important algorithmic tool for the molecular sciences. RE simulations involve the concurrent execution of independent simulations which infrequently interact and exchange information. The next set of simulation parameters are based upon the outcome of the exchanges. Typically RE functionality is integrated into the molecular simulation software package. A primary motivation of the tight integration of RE functionality with simulation codes has been performance. This is limiting at multiple levels. First, advances in the RE methodology are tied to the molecular simulation code. Consequently these advances remain confined to the molecular simulation code for which they were developed. Second, it is difficult to extend or experiment with novel RE algorithms, since expertise in the molecular simulation code is typically required. In this paper, we propose the RepEx framework which address these aforementioned shortcomings of existing approaches, while striking the balance between flexibility (any RE scheme) and scalability (tens of thousands of replicas) over a diverse range of platforms. RepEx is designed to use a pilot-job based runtime system and support diverse RE Patterns and Execution Modes. RE Patterns are concerned with synchronization mechanisms in RE simulation, and Execution Modes with spatial and temporal mapping of workload to the CPU cores. We discuss how the design and implementation yield the following primary contributions of the RepEx framework: (i) its ability to support different RE schemes independent of molecular simulation codes, (ii) provide the ability to execute different exchange schemes and replica counts independent of the specific availability of resources, (iii) provide a runtime system that has first-class support for task-level parallelism, and (iv) required scalability along multiple dimensions.

cs.DC↗

arXiv2015

Pilot-Abstraction: A Valid Abstraction for Data-Intensive Applications on HPC, Hadoop and Cloud Infrastructures?

HPC environments have traditionally been designed to meet the compute demand of scientific applications and data has only been a second order concern. With science moving toward data-driven discoveries relying more on correlations in data to form scientific hypotheses, the limitations of HPC approaches become apparent: Architectural paradigms such as the separation of storage and compute are not optimal for I/O intensive workloads (e.g. for data preparation, transformation and SQL). While there are many powerful computational and analytical libraries available on HPC (e.g. for scalable linear algebra), they generally lack the usability and variety of analytical libraries found in other environments (e.g. the Apache Hadoop ecosystem). Further, there is a lack of abstractions that unify access to increasingly heterogeneous infrastructure (HPC, Hadoop, clouds) and allow reasoning about performance trade-offs in this complex environment. At the same time, the Hadoop ecosystem is evolving rapidly and has established itself as de-facto standard for data-intensive workloads in industry and is increasingly used to tackle scientific problems. In this paper, we explore paths to interoperability between Hadoop and HPC, examine the differences and challenges, such as the different architectural paradigms and abstractions, and investigate ways to address them. We propose the extension of the Pilot-Abstraction to Hadoop to serve as interoperability layer for allocating and managing resources across different infrastructures. Further, in-memory capabilities have been deployed to enhance the performance of large-scale data analytics (e.g. iterative algorithms) for which the ability to re-use data across iterations is critical. As memory naturally fits in with the Pilot concept of retaining resources for a set of tasks, we propose the extension of the Pilot-Abstraction to in-memory resources.

cs.DC↗

arXiv2014

A Tale of Two Data-Intensive Paradigms: Applications, Abstractions, and Architectures

Scientific problems that depend on processing large amounts of data require overcoming challenges in multiple areas: managing large-scale data distribution, co-placement and scheduling of data with compute resources, and storing and transferring large volumes of data. We analyze the ecosystems of the two prominent paradigms for data-intensive applications, hereafter referred to as the high-performance computing and the Apache-Hadoop paradigm. We propose a basis, common terminology and functional factors upon which to analyze the two approaches of both paradigms. We discuss the concept of "Big Data Ogres" and their facets as means of understanding and characterizing the most common application workloads found across the two paradigms. We then discuss the salient features of the two paradigms, and compare and contrast the two approaches. Specifically, we examine common implementation/approaches of these paradigms, shed light upon the reasons for their current "architecture" and discuss some typical workloads that utilize them. In spite of the significant software distinctions, we believe there is architectural similarity. We discuss the potential integration of different implementations, across the different levels and components. Our comparison progresses from a fully qualitative examination of the two paradigms, to a semi-quantitative methodology. We use a simple and broadly used Ogre (K-means clustering), characterize its performance on a range of representative platforms, covering several implementations from both paradigms. Our experiments provide an insight into the relative strengths of the two paradigms. We propose that the set of Ogres will serve as a benchmark to evaluate the two paradigms along different dimensions.

cs.DC↗

arXiv2013

Pilot-Data: An Abstraction for Distributed Data

Scientific problems that depend on processing large amounts of data require overcoming challenges in multiple areas: managing large-scale data distribution, controlling co-placement and scheduling of data with compute resources, and storing, transferring, and managing large volumes of data. Although there exist multiple approaches to addressing each of these challenges, an integrative approach is missing; furthermore, extending existing functionality or enabling interoperable capabilities remains difficult at best. We propose the concept of Pilot-Data to address the fundamental challenges of co-placement and scheduling of data and compute in heterogeneous and distributed environments with interoperability and extensibility as first-order concerns. Pilot-Data is an extension of the Pilot-Job abstraction for supporting the management of data in conjunction with compute tasks. Pilot-Data separates logical data units from physical storage, thereby providing the basis for efficient compute/data placement and scheduling. In this paper, we discuss the design and implementation of the Pilot-Data prototype, demonstrate its use by data-intensive applications on multiple production distributed cyberinfrastructure and illustrate the advantages arising from flexible execution modes enabled by Pilot-Data. Our experiments utilize an implementation of Pilot-Data in conjunction with a scalable Pilot-Job (BigJob) to establish the application performance that can be enabled by the use of Pilot-Data. We demonstrate how the concept of Pilot-Data also provides the basis upon which to build tools and support capabilities like affinity which in turn can be used for advanced data-compute co-placement and scheduling.

cs.DC↗

arXiv2012

Survey and Analysis of Production Distributed Computing Infrastructures

This report has two objectives. First, we describe a set of the production distributed infrastructures currently available, so that the reader has a basic understanding of them. This includes explaining why each infrastructure was created and made available and how it has succeeded and failed. The set is not complete, but we believe it is representative. Second, we describe the infrastructures in terms of their use, which is a combination of how they were designed to be used and how users have found ways to use them. Applications are often designed and created with specific infrastructures in mind, with both an appreciation of the existing capabilities provided by those infrastructures and an anticipation of their future capabilities. Here, the infrastructures we discuss were often designed and created with specific applications in mind, or at least specific types of applications. The reader should understand how the interplay between the infrastructure providers and the users leads to such usages, which we call usage modalities. These usage modalities are really abstractions that exist between the infrastructures and the applications; they influence the infrastructures by representing the applications, and they influence the ap- plications by representing the infrastructures.

cs.DC↗

arXiv2012

P*: A Model of Pilot-Abstractions

Pilot-Jobs support effective distributed resource utilization, and are arguably one of the most widely-used distributed computing abstractions - as measured by the number and types of applications that use them, as well as the number of production distributed cyberinfrastructures that support them. In spite of broad uptake, there does not exist a well-defined, unifying conceptual model of Pilot-Jobs which can be used to define, compare and contrast different implementations. Often Pilot-Job implementations are strongly coupled to the distributed cyber-infrastructure they were originally designed for. These factors present a barrier to extensibility and interoperability. This pa- per is an attempt to (i) provide a minimal but complete model (P*) of Pilot-Jobs, (ii) establish the generality of the P* Model by mapping various existing and well known Pilot-Job frameworks such as Condor and DIANE to P*, (iii) derive an interoperable and extensible API for the P* Model (Pilot-API), (iv) validate the implementation of the Pilot-API by concurrently using multiple distinct Pilot-Job frameworks on distinct production distributed cyberinfrastructures, and (v) apply the P* Model to Pilot-Data.

cs.DC↗

arXiv2005

Effects of Disorder on Electron Transport in Arrays of Quantum Dots

We investigate the zero-temperature transport of electrons in a model of quantum dot arrays with a disordered background potential. One effect of the disorder is that conduction through the array is possible only for voltages across the array that exceed a critical voltage $V_T$. We investigate the behavior of arrays in three voltage regimes: below, at and above the critical voltage. For voltages less than $V_T$, we find that the features of the invasion of charge onto the array depend on whether the dots have uniform or varying capacitances. We compute the first conduction path at voltages just above $V_T$ using a transfer-matrix style algorithm. It can be used to elucidate the important energy and length scales. We find that the geometrical structure of the first conducting path is essentially unaffected by the addition of capacitive or tunneling resistance disorder. We also investigate the effects of this added disorder to transport further above the threshold. We use finite size scaling analysis to explore the nonlinear current-voltage relationship near $V_T$. The scaling of the current $I$ near $V_T$, $I\sim(V-V_T)^β$, gives similar values for the effective exponent $β$ for all varieties of tunneling and capacitive disorder, when the current is computed for voltages within a few percent of threshold. We do note that the value of $β$ near the transition is not converged at this distance from threshold and difficulties in obtaining its value in the $V\searrow V_T$ limit.

cond-mat.dis-nn↗