Toan Nguyen

Dynamic resilient workflows for collaborative design

Large-scale simulation and optimization are demanding applications that require high-performance computing platforms. Because their economic impact is fundamental to the industry, they also require robust, seamless and effective mechanisms to support dynamic user interactions, as well as faulttolerance and resiliency for multidiscipline applications. Distributed workflows are considered here as a means to support large-scale dynamic and resilient multiphysics simulation and optimization applications, such as multiphysics aircraft simulation.

Resilient workflows for computational mechanics platforms

Workflow management systems have recently been the focus of much interest and many research and deployment for scientific applications worldwide [26, 27]. Their ability to abstract the applications by wrapping application codes have also stressed the usefulness of such systems for multidiscipline applications [23, 24]. When complex applications need to provide seamless interfaces hiding the technicalities of the computing infrastructures, their high-level modeling, monitoring and execution functionalities help giving production teams seamless and effective facilities [25, 31, 33]. Software integration infrastructures based on programming paradigms such as Python, Mathlab and Scilab have also provided evidence of the usefulness of such approaches for the tight coupling of multidisciplne application codes [22, 24]. Also high-performance computing based on multi-core multi-cluster infrastructures open new opportunities for more accurate, more extensive and effective robust multi-discipline simulations for the decades to come [28]. This supports the goal of full flight dynamics simulation for 3D aircraft models within the next decade, opening the way to virtual flight-tests and certification of aircraft in the future [23, 24, 29].

A Hadoop Use Case for Engineering Data

This paper presents the VELaSSCo project (Visualization for Extremely LArge-Scale Scientific Computing). It aims to develop a platform to manipulate scientific data used by FEM (Finite Element Method) and DEM (Discrete Element Method) simulations. The project focuses on the development of a distributed, heterogeneous and high-performance platform, enabling the scientific communities to store, process and visualize huge amounts of data. The platform is compatible with current hardware capabilities, as well as future hardware.

Resilience for collaborative applications on clouds: fault-tolerance for distributed HPC applications

Because e-Science applications are data intensive and require long execution runs, it is important that they feature fault-tolerance mechanisms. Cloud and grid computing infrastructures often support system and network fault-tolerance. They repair and prevent communication and software errors. They allow also checkpointing of applications, duplication of jobs and data to prevent catastrophic hardware failures. However, only preliminary work has been done so far on application resilience, i.e., the ability to resume normal execution following application errors and abnormal executions. This paper is an overview of open issues and solutions for such errors detection and management. It also overviews the implementation of a workflow management system to design, deploy, execute, monitor, restart and resume distributed HPC applications on cloud infrastructures in cases of failures.

A dynamic workflow simulation platform

In numeric optimization algorithms errors at application level considerably affect the performance of their execution on distributed infrastructures. Hours of execution can be lost only due to bad parameter configurations. Though current grid workflow systems have facilitated the deployment of complex scientific applications on distributed environments, the error handling mechanisms remain mostly those provided by the middleware. In this paper, we propose a collaborative platform for the execution of scientific experiments in which we integrate a new approach for treating application errors, using the dynamicity and exception handling mechanisms of the YAWL workflow management system. Thus, application errors are correctly detected and appropriate handling procedures are triggered in order to save as much as possible of the work already executed.

Resilient workflows for cooperative design

This paper describes an approach to extend process modeling for engineering design applications with fault-tolerance and resilience capabilities. It is based on the requirements for application-level error handling, which is a requirement for petascale and exascale scientific computing. This complements the traditional fault-tolerance management features provided by the existing hardware and distributed systems. These are often based on data and operations duplication and migration, and on checkpoint-restart procedures. We show how they can be optimized for high-performance infrastructures. This approach is applied on a prototype tested against industrial testcases for optimization of engineering design artifacts.his electronic document is a “live” template. The various components of your paper [title, text, heads, etc.] are already defined on the style sheet, as illustrated by the portions given in this document.

Resilient workflows for high-performance simulation platforms

Workflows systems are considered here to support large-scale multiphysics simulations. Because the use of large distributed and parallel multi-core infrastructures is prone to software and hardware failures, the paper addresses the need for error recovery procedures. A new mechanism based on asymmetric checkpointing is presented. A rule-based implementation for a distributed workflow platform is detailed.

Applications resilience on clouds

Cloud computing infrastructures support system and network fault-tolerance. They transparently repair and prevent communication and software errors. They also allow duplication and migration of jobs and data to prevent hardware failures. However, only limited work has been done so far on application resilience, i.e., the ability to resume normal execution after errors and abnormal executions in distributed environments and clouds. This paper addresses open issues and solutions for application errors detection and management. It also overviews a testbed used to to design, deploy, execute, monitor, restart and resume distributed applications on cloud infrastructures in cases of failures.

Workflow Resiliency for Large-Scale Distributed Applications

Large-scale simulation and optimization are demanding applications that require high-performance computing platforms. Because their economic impact is fundamental to the industry, they also require robust, seamless and effective mechanisms to support dynamic user interactions, as well as fault-tolerance and resiliency on parallel computing platforms. Distributed workflows are considered here as a means to support large-scale dynamic and resilient multiphysics simulation and optimization applications, such as multiphysics aircraft simulation.