Deploying Cost-Efficient Virtual Elastic Clusters across Multi-Clouds

Autores UPV
Año
CONGRESO Deploying Cost-Efficient Virtual Elastic Clusters across Multi-Clouds

Abstract

Elastic Cloud Computing Cluster (EC3) [1] is a tool that creates self-managed cost-efficient virtual hybrid elastic clusters on top of Infrastructure as a Service (IaaS) Clouds. These self-managed clusters have the capability to adapt the size of the cluster to the workload, creating the illusion of a real cluster without requiring an investment beyond the actual usage. They can be fully customized and dynamically and automatically enlarged and shrunk (in terms of the number of nodes) depending on the number of jobs queued up at the Local Resource Management System (LRMS). The elasticity management is automatically carried out by the front-end node of the cluster, with the help of CLUES [2]. EC3 can also deploy hybrid clusters across on-premises and public Cloud resources, where on-premises resources are supplemented with public Cloud resources to accelerate the execution process by accessing an additional number of resources, thus introducing Cloud bursting capabilities. Different instance types and the use of spot instances [3] combined with on-demand resources are also cluster configurations supported by EC3. Moreover, using spot instances, together with checkpointing techniques, our tool can significantly reduce the total cost of executions while introducing automatic fault tolerance. EC3 eases the deployment of virtual clusters providing both a Software as a Service (SaaS) graphical web interface, where no registration is required [4], and a powerful Command Line Interface (CLI), for enhanced capabilities.