Dynamic Resource Management in Clouds: A Probabilistic Approach

TL;DR

This paper introduces a probabilistic model inspired by epidemiology to predict workload surges in cloud resource management, enabling more effective and adaptive resource provisioning policies for Video on Demand services.

Contribution

It proposes a novel analytical model using Large Deviation Principles to characterize rare workload spikes, informing dynamic resource management policies in cloud computing.

Findings

Model verifies a Large Deviation Principle for workload extremes

Provides insights into abnormal system behaviors during workload surges

Supports development of SLA policies for elastic resource provisioning

Abstract

Dynamic resource management has become an active area of research in the Cloud Computing paradigm. Cost of resources varies significantly depending on configuration for using them. Hence efficient management of resources is of prime interest to both Cloud Providers and Cloud Users. In this work we suggest a probabilistic resource provisioning approach that can be exploited as the input of a dynamic resource management scheme. Using a Video on Demand use case to justify our claims, we propose an analytical model inspired from standard models developed for epidemiology spreading, to represent sudden and intense workload variations. We show that the resulting model verifies a Large Deviation Principle that statistically characterizes extreme rare events, such as the ones produced by "buzz/flash crowd effects" that may cause workload overflow in the VoD context. This analysis provides valuable insight on expectable abnormal behaviors of systems. We exploit the information obtained using the Large Deviation Principle for the proposed Video on Demand use-case for defining policies (Service Level Agreements). We believe these policies for elastic resource provisioning and usage may be of some interest to all stakeholders in the emerging context of cloud networking