Predictive Performance Modeling for Distributed Computing using Black-Box Monitoring and Machine Learning

Abstract

In many domains, the previous decade was characterized by increasing data volumes and growing complexity of computational workloads, creating new demands for highly data-parallel computing in distributed systems. Effective operation of these systems is challenging when facing uncertainties about the performance of jobs and tasks under varying resource configurations, e.g., for scheduling and resource allocation. We survey predictive performance modeling (PPM) approaches to estimate performance metrics such as execution duration, required memory or wait times of future jobs and tasks based on past performance observations. We focus on non-intrusive methods, i.e., methods that can be applied to any workload without modification, since the workload is usually a black-box from the perspective of the systems managing the computational infrastructure. We classify and compare sources of performance variation, predicted performance metrics, required training data, use cases, and the underlying prediction techniques. We conclude by identifying several open problems and pressing research needs in the field.

19 pages, 3 figures, 5 tables

References in corpus (3)

• Cloud Computing and Grid Computing 360-Degree Compared
• LIKWID: Lightweight Performance Tools
• LIKWID: A lightweight performance-oriented tool suite for x86 multicore environments

Cited by in corpus (13)

• Machine learning the computational cost of quantum chemistry
• Tarema: Adaptive Resource Allocation for Scalable Scientific Workflows in Heterogeneous Clusters
• Lotaru: Locally Estimating Runtimes of Scientific Workflow Tasks in Heterogeneous Clusters
• Lotaru: Locally Predicting Workflow Task Runtimes for Resource Management on Heterogeneous Infrastructures
• Sizey: Memory-Efficient Execution of Scientific Workflow Tasks
• Towards Advanced Monitoring for Scientific Workflows
• Portability of Scientific Workflows in NGS Data Analysis: A Case Study
• Scalability Model for the LOFAR Direction Independent Pipeline
• Collaborative Cluster Configuration for Distributed Data-Parallel Processing: A Research Overview
• Load Balancing with Job-Size Testing: Performance Improvement or Degradation?
• Optimizing Memory Allocation in Distributed Clusters with Predictive Modeling
• A Forecasting System of Computational Time of DFT/TDDFT Calculations under the Multiverse ansatz via Machine Learning and Cheminformatics
• Cost models for geo-distributed massively parallel streaming analytics