Towards Elastic Virtual Machine Placement in Overbooked OpenStack Clouds under Uncertainty

Cloud computing datacenters currently provide millions of virtual machines in highly dynamic Infrastructure as a Service (IaaS) markets. As a first step on implementing algorithms previously proposed by the authors for Virtual Machine Placement (VMP) in a real- world IaaS middleware, this work prese...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: López-Pires, Fabio, Barán, Benjamín, Pereira, Carolina, Velázquez, Marcelo, González, Osvaldo
Formato: Objeto de conferencia
Lenguaje:Inglés
Publicado: 2018
Materias:
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/70003
Aporte de:
Descripción
Sumario:Cloud computing datacenters currently provide millions of virtual machines in highly dynamic Infrastructure as a Service (IaaS) markets. As a first step on implementing algorithms previously proposed by the authors for Virtual Machine Placement (VMP) in a real- world IaaS middleware, this work presents an experimental comparison of these algorithms against current algorithms considered for solving VMP problems in OpenStack. Several experiments considering scenario- based simulations for uncertainty modelling demonstrate that the proposed algorithms present promising results for its implementation towards real-world operations. Next research steps are also summarized.