Skip to main content

Advances, Systems and Applications

Table 4 Experimental parameters

From: Integrated resource management pipeline for dynamic resource-effective cloud data center

RMP phase

Algorithm

Parameters

Definition

Values

Underload

THR

n

last CPU utilization values to be averaged

n=2

detection

 

THRunderload

underload CPU utilization threshold

THRoverload=0.5

Overload

THR

n

last CPU utilization values to be averaged

n=2

detection

 

THRoverload

overload CPU utilization threshold

THRoverload=0.8,0.9,1.0

 

LRR

n

last CPU utilization values to be averaged

n=2

  

α

safety parameter

α=0.9,1.0,1.1

 

DES

n

last CPU utilization values to be averaged

n=2

  

N

Input samples

N=30

  

K

Forecast sample

K=1

 

SHW

n

last CPU utilization values to be averaged

n=2

  

N

Input samples

N=30

  

K

Forecast sample

K=1

VM

MMTMC

n

last CPU utilization values to be averaged

n=2

selection

MMTMC 1

n

last CPU utilization values to be averaged

n=2

 

MMTMC 2

n

last CPU utilization values to be averaged

n=2

  

uinitial

The initial threshold value

uinitial=3

  

λ

The rate of growth

λ is an empirical value

VM

Modified best fit

THRPlacement

CPU threshold from total server’s capacity

THRPlacement=80%

placement

Medium-fit (MF)

THRPlacement

CPU threshold from total server’s capacity

THRPlacement=80%