Simulating High Availability Scenarios in Cloud Data Centers: A Closer Look

Abstract

Migrating to the cloud is becoming a necessity for the majority of businesses. Cloud tenants require certain levels of performance in aspects like high availability and service rate and deployment options. On the other hand, Cloud providers are in constant pursuit of a system that satisfies client demands for resources, maximizes availability, minimizes power consumption and, in turn, minimizes the cloud providers' cost. A main challenge cloud providers face here is ensuring high availability (HA). High availability includes the combined reliability of components of all categories including network, computational, hardware and software components of all layers. In this work, we first address the need for a cloud simulator that enables HA algorithm testing in cloud environments and observe its impact on energy efficiency. We introduce a framework to amend cloud simulators with critical HA features. We take GreenCloud, a major simulator with a direct focus on green computing, and implement these features as an additional measurement layer. We demonstrate these added features by simulating their impact on a phased communication application (PCA).