Measuring the Robustness of Resource Allocations in a Stochastic Dynamic Environment

Abstract

Heterogeneous distributed computing systems often must operate in an environment where system parameters are subject to uncertainty. Robustness can be defined as the degree to which a system can function correctly in the presence of parameter values different from those assumed. We present a methodology for quantifying the robustness of resource allocations in a dynamic environment where task execution times are stochastic. The methodology is evaluated through measuring the robustness of three different resource allocation heuristics within the context of a stochastic dynamic environment. A Bayesian regression model is fit to the combined results of the three heuristics to demonstrate the correlation between the stochastic robustness metric and the presented performance metric. The correlation results demonstrated the significant potential of the stochastic robustness metric to predict the relative performance of the three heuristics given a common objective function.

Evaluating Robustness of Resource Allocation in Uniprocessor Real Time System

“…In stochastic approach [11], each perturbation parameter, or uncertainty parameter, is modeled as a random variable fully described by a probability mass function (pmf). Our second approach differs from that in [1], where a single deterministic estimated value for each of the identified perturbation parameters is used.…”

“…In stochastic model [5,6,11,12], more information regarding the variation in the perturbation parameters is assumed known. Representing the uncertainty parameters of the system as stochastic variables enables the robustness metric in the stochastic model to account for all possible outcomes for the performance of the system.…”

“…Representing the uncertainty parameters of the system as stochastic variables enables the robustness metric in the stochastic model to account for all possible outcomes for the performance of the system. The stochastic robustness metric [9,11,12] requires more information and is far more complex to calculate than its deterministic counterpart. To handle the computational complexity, [3] considered the FFT and bootstrap approximation methods that greatly simplify the required calculations.…”

Models and Heuristics for Robust Resource Allocation in Parallel and Distributed Computing Systems

“…In evaluating our results [9], we demonstrate that in a dynamic environment a small number of simulation trials are required to sufficiently indicate the performance of a resource allocation heuristic relative to our given performance objective.…”

“…The dynamic SRM values for different resource allocation heuristics can then be compared to select the approach that is more robust within the given environment [9].…”