Advances, Systems and Applications
Table 2 Host underload detection algorithms
From: Proactive dynamic virtual-machine consolidation for energy conservation in cloud data centres
Algorithm | Policies | Characteristic | |||
|---|---|---|---|---|---|
Available Capacity | Migration Delay | Number of VM | Host power | ||
Least utilized host [21] | √ |  |  |  | Base on a host with minimum resources Not cover number of VMs on the host |
Static Threshold [21] | √ |  |  |  | Depend on the mean of the last CPU used Difficult to find the optimal value of the threshold |
Available capacity [14] | √ |  |  |  | Base on available host capacity compared to others; Not necessary PMs has less power than the other |
Migration delay [14] |  | √ |  |  | Base on minimum time to complete VMs migration process; Need a lot of predication and estimation |
Hybrid [14] | √ | √ | √ |  | based on MCDM More complicated and difficult for practical |
Weighted CPU utilization [124] | √ |  | √ |  | Combines host utilization and number of VMs Need less computation then hybrid |