Normalized compute hours overview

Normalized compute hours are the approximate hours of compute time that an instance has used. This value is an indicator of how much compute resource is being used in your infrastructure. You can use this value along with the number of instances value to estimate the cost of your public cloud or on-premises infrastructure. Also, you can determine the compute time that is used for the cost that you are paying for your infrastructure.  

Only compute-related services and resources can have the normalized compute hours value. (For example, cloud services such as Amazon EC2)

For more information, see the following sections:



Normalized compute hours calculation

A cloud provider's environment has different types of instances. The compute hours between instances can be compared based on a normalizing instance that is predefined for the cloud providers. 

The normalized compute hours value is based on the normalization factor of this predefined instance type. Normalization factor is a number that is assigned to an instance type and is used to compare the relative values of the instance types. The larger the instance size, the higher is the normalization factor, and vice versa. Every existing instance type has a normalization factor.


Normalized compute hours for an instance = (Hours of compute time) x (normalization factor of the associated instance type)

If the normalization factor for m1.small = 1, then, the normalized compute hours value for 2 hours of compute time of:

One instance of m1.small = 1 x 2 x 1 = 2

10 m1.small instances = 10 x 2 x 1 = 20


The cost of each hour of computation for an instance of a bigger instance type will be more than that for an instance of a smaller instance type. For example, the same number of instances of m1.large instance type will cost you more than m1.small instance type. 

Normalization factors

The normalization factor calculation for an instance type is based on the size of a standard instance type. The configuration of the standard instance type is the same for both the public cloud providers and on-premises infrastructure.

Standard instance type (refvm) configuration:

refvm_cpu = 1

refvm_mem = 2 GB

Normalization factor = max(vCPU/refvm_cpu, vMem/refvm_mem) = max(vCPU/1, vMem/2)

The following tables provide the normalization factors per existing instance type per provider:

AWS instances

Instance TypevCPUMemory (GB)Normalization factor

General Purpose - Current Generation

T2
t2.nano10.51
t2.micro111
t2.small121
t2.medium242
t2.large284
t2.xlarge4168
t2.2xlarge83216
M4 
m4.large284
m4.xlarge4168
m4.2xlarge83216
m4.4xlarge166432
m4.10xlarge4016080
m4.16xlarge64256128
M3
m3.medium13.751.88
m3.large27.53.75
m3.xlarge4157.5
m3.2xlarge83015

Compute Optimized - Current Generation

C4

c4.large23.752
c4.xlarge47.54
c4.2xlarge8158
c4.4xlarge163016
c4.8xlarge366036
C3
c3.large23.752
c3.xlarge47.54
c3.2xlarge8158
c3.4xlarge163016
c3.8xlarge326032

Memory Optimized - Current Generation

X1

x1.16xlarge64976488
x1.32xlarge1281952976
R4
r4.large215.257.63
r4.xlarge430.515.25
r4.2xlarge86130.5
r4.4xlarge16122

61

r4.8xlarge32244122
r4.16xlarge64488244
R3
r3.large215.25 7.63
r3.xlarge430.515.25
r3.2xlarge86130.5
r3.4xlarge1612261
r3.8xlarge32244122

Accelerated Computing Instances - Current Generation

P2

p2.xlarge46130.5
p2.8xlarge32488244
p2.16xlarge64732366
G2
g2.2xlarge8158
g2.8xlarge326032
F1
f1.2xlarge8122

61

f1.16xlarge64976

488

Storage Optimized - Current Generation

D2 - Dense Storage Instances

d2.xlarge430.5

15.25

d2.2xlarge86130.5
d2.4xlarge1612261
d2.8xlarge32244122
I3 - High I/O Instances
i3.large215.25

7.63

i3.xlarge430.515.25
i3.2xlarge86130.5
i3.4xlarge16122

61

i3.8xlarge32244122
i3.16xlarge64488244

Cluster Networking - Current Generation

I2
i2.xlarge430.5

15.25

i2.2xlarge86130.5
i2.4xlarge1612261
i2.8xlarge32244122
Previous Generation
t1.micro10.6
m1.small11.71
m1.medium13.71.85
m1.large27.5

3.75

m1.xlarge415

7.5

m2.xlarge217.18.55
m2.2xlarge434.217.1
m2.4xlarge868.434.2
c1.medium21.72
c1.xlarge87

8

cc2.8xlarge3260.532
cg1.4xlarge1622.516
cr1.8xlarge32224112
hi1.4xlarge1660.530.25
hs1.8xlarge1611758.5

 

Azure instances

Instance TypevCPUMemory GBNormalization factor

Basic Instances   

A010.751
A111.751
A223.52
A3474
A48148

Standard Instances   

A1 v2121
A2 v2242
A4 v2484
A8 v28168
A2m v22168
A4m v243216
A8m v286432
H8m811256
H16mr16224112
H16r1611256
H16m16224112
H161611256
H885628
A010.751
A111.751
A223.52
A3474
A48148
A1085628
A111611256
A52147
A642814
A785628
A885628
A91611256
D113.51.75
D112147
D1242814
D1385628
D141611256
D2273.5
D34147
D482814
D11 v22147
D12 V242814
D13 V285628
D14 v21611256
D15 v22014070
D1 V213.51.75
D2 V2273.5
D3 V24147
D4 V282814
D5 V2165628
F1121
F16163216
F2242
F4484
F88168
NC121211256
NC2424224112
NC24r24224112
NC665628
NV121211256
NV2424224112
NV665628

Additional VMs for Windows OS in East US region (Standard Tier)

G122814
G245628
G3811256
G416224112
G532448224
L161612864
L3232256128
L443216
L886432

Google Cloud Platform instances

Information

For a custom instance type (or machine type) in Google Cloud Platform, the normalization factor cannot be predefined. It is calculated dynamically (during runtime) by using the standard instance type (refvm) configuration.

Instance Type

vCPUMemory (GB)Normalization Factor
Shared Core - Bursting
f1-micro10.61
g1-small11.71
Standard
n1-standard-113.751.88
n1-standard-227.53.75
n1-standard-44157.5
n1-standard-883015
n1-standard-16166030
n1-standard-323212060
n1-standard-6464240120
High Memory
n1-highmem-221365
n1-highmem-442613
n1-highmem-885226
n1-highmem-161610452
n1-highmem-3232208104
n1-highmem-6464416208
High CPU
n1-highcpu-221.82
n1-highcpu-443.64
n1-highcpu-887.28
n1-highcpu-161614.416
n1-highcpu-323228.832
n1-highcpu-646457.664

On-premises instances

The normalization factor calculation for an on-premises virtual machine (VM) or instance is based on the configuration of the VM.

Was this page helpful? Yes No Submitting... Thank you

Comments