PANORAMA: An approach to performance modeling and diagnosis of extreme-scale workflows

Abstract

Computational science is well established as the third pillar of scientific discovery and is on par with experimentation and theory. However, as we move closer toward the ability to execute exascale calculations and process the ensuing extreme-scale amounts of data produced by both experiments and computations alike, the complexity of managing the compute and data analysis tasks has grown beyond the capabilities of domain scientists. Thus, workflow management systems are absolutely necessary to ensure current and future scientific discoveries. A key research question for these workflow management systems concerns the performance optimization of complex calculation and data analysis tasks. The central contribution of this article is a description of the PANORAMA approach for modeling and diagnosing the run-time performance of complex scientific workflows. This approach integrates extreme-scale systems testbed experimentation, structured analytical modeling, and parallel systems simulation into a comprehensive workflow framework called Pegasus for understanding and improving the overall performance of complex scientific workflows.

An automated analysis workflow for optimization of force-field parameters using neutron scattering data

“…The optimization technique uses a set of codes that includes Pegasus [13], Nanoscale formed outside the workflow using python tools [18] and is incorporated into Mantid [16]. In the following we detail the procedures used in the global optimization technique.…”

“…In a 'manual mode', the user can prevent the waste of resources only by managing the set of batch jobs so that every job runs once the input data becomes available. daxgen [13,17] program generates the workflow object into a set of template files using a python string formatter. Thereafter, the workflow is run using Pegasus commands.…”

A Manifesto for Future Generation Cloud Computing

“…As the amount of data processed increases and extreme-scale workflows begin to emerge, it is important to consider key concerns such as fault tolerance, performance modelling, efficient data management, and efficient resource usage. For this purpose, Big Data analytics will become a crucial tool [62]. For instance, monitoring and analysing resource consumption data may enable workflow management systems to detect performance anomalies and potentially predict failures, leveraging technologies such as serverless computing to manage the execution of complex workflows that are reusable and can be shared across multiple stakeholders.…”

Task Runtime Prediction in Scientific Workflows Using an Online Incremental Learning Approach

“…To execute the workflows and collect the monitoring data as outlined in Section III, we use the Pegasus [22] WMS that is equipped with a monitoring plugin as part of the Panorama [23] project. The monitoring is done at a task level.…”