Using Agent-Based VM Placement Policy

The huge expansion in infrastructure and services in recent years to cover the high demand on processing big data has created a mega Cloud Data enter of high complexity with increasing difficulties to identify and allocate efficiently an appropriate host for a requested virtual machine (VM). Thus, it is vital to establish a good awareness of all descanter's resources in order to enable allocation "placement" policies to make the best decision in reducing the required time for the creation and allocation of a VM at a proper host. Most of current placement "allocation" algorithms have a leakage in the broad awareness of datacenter's resources with adverse impactions on the allocation progress of their policies. This paper presents a new Agent-based placement policy that employs some multi-agent system's features to achieve a good awareness of Cloud Datacenter's resources and also provide an efficient allocation decision for the requested VMs. Consequently, it reduces the response time of VM allocation and usage of occupied resources. The agent-based policy is implemented by using the Cloud Sim toolkit [10, 11] and is favourably compared against the toolkit's own default policy. The comparative study is based on typical numerical experiments, focusing on the response time of VM allocation and other aspects such as the number of available VM types and the amount of occupied resources.