RESUMO
Cloud computing is a computational model in which providers offer
on-demand services to clients in a transparent way. In this context,
the efficient provisioning of resources is extremely important, since
it is directly related to Quality of Service (QoS), with the Service Level
Agreement (SLA) compliance and the cost of the service. In this
way, virtualization becomes a crucial factor in resource provisioning.
In this paper, performance evaluations were performed considering
environments configured with different virtualization strategies: by
virtual machines and by containers. In the first set of experiments,
we evaluated the boot time of the system with the KVM, Xen and
Docker tools and with variations in the number of instances and
virtual cores. In the second, the same mechanisms were analyzed
during the execution of different benchmarks (Apache and Smallpt)
and with variations in the number of virtual resources provisioned.
In the results, Docker proved to be more efficient when compared
to other tools.