You are on page 1of 29

Cloud Migration Analysis & Insights

2016 Industry Report on Infrastructures


Moving to AWS and Azure

Table of Contents
Introduction ..............................................................................................

Definitions ................................................................................................

How to Read this Report .......................................................................

Distribution of Operating Systems Migrating to AWS ..................

Instance Types, Provisioning Distribution, and Cost


Savings by Operating System on AWS ..................................... 8-15
Distribution of Operating Systems Migrating to Azure ...............

17

Instance Types, Provisioning Distribution, and Cost


Savings by Operating System on Azure ................................ 18-25
State of Provisioning on AWS vs. Azure ........................................

26

Takeaways .............................................................................................

27

www.cloudamize.com

Introduction
Gartner predicts that by 2020, more compute power will be sold via the cloud than what
is deployed in customers on-premises data centers. Gartner says that this will be a difficult
environment to manage. With most compute power moving to IaaS providers, enterprises
and vendors need to focus on managing and leveraging the hybrid combination of onpremises, off-premises, cloud and non-cloud architectures, with a focus on managing clouddelivered capacity efficiently and effectively, said Thomas Bittman, Vice President and
Distinguished analyst at Gartner.
Companies are increasingly moving to the cloud because it provides significant benefits,
including agility, efficiency, performance improvements, and cost savings. But the cloud is
complex. If you combine all of the instance types, storage options, availability zones, and
pricing plans, there are over 25 million cloud configurations available. And on top of this,
these configurations are constantly changing. Because of this complexity, we wanted to
provide real data that yields helpful insights for successfully migrating to the cloud, and what
opportunities exist for improving performance and cost savings once in the cloud. To achieve
this, we analyzed 10,000 on-premise nodes globally between April and July 2016 that were
migrating to either Azure or AWS.

This report provides key


insights such as:
The most common Operating Systems migrating from on-premise to AWS and Azure
The top instances mapped to each OS when planning for migration to AWS and Azure
Cost savings realized for mapping each OS on a specific instance on AWS and Azure
The provisioning distribution of instances by OS and in aggregate on both AWS and
Azure and what this means for opportunities to improve
Windows vs. Linux cost savings when migrating to the cloud
Average cost savings when migrating to AWS vs. Azure

www.cloudamize.com

Definitions
OS Types
The OS types are the most common operating systems identified migrating to AWS and
Azure, and are the operating systems running on the AWS or Azure instance.
Instance Types
The instance type is a type of virtual machine that is running on AWS or Azure. It is
characterized by number of cores, amount of RAM, available disk, and network IO. In this
report, we show the top instance types for each OS and combine the rest of the instance
types in Others.
Cost Savings
Cost savings are the savings realized by mapping an on-premise server or virtual machine to
a specific instance type in the cloud compared to the cost of maintaining it on-premise.
Provisioning
In the cloud, you can provision your compute, storage and network resources, giving you
more flexibility based on what your current demands are. When you optimally manage your
provisioning, you right-size your infrastructure for the cloud, allowing you to reduce capital
expenses, minimize costs, and improve performance. For this report, provisioning is defined as:
Over-Provisioned: The purchased capacity of the instance is greater than the capacity
required by the workload.
Under-Provisioned: The purchased capacity of the instance is less than the capacity
required by the workload.
Optimal Provisioning: The purchased capacity of the instance meets the capacity
required by the workload, and there is no cheaper option available that also meets that
requirements.
Idle: The purchased capacity of the instance is optimal; however, the instance can be
turned off when it is not being utilized, and you can therefore realize additional cost
savings. These types of instances are good candidates for auto-scaling when moved to
the cloud.

www.cloudamize.com

How to Read this Report


On each of the following pages, you will see the most popular
instance types for each OS when mapped to AWS and Azure, and
the cost savings realized by mapping the OS to a specific instance.
Identifying the optimal instance, however, is only the first step. Once
you migrate to the cloud, you have to actively provision the capacity
you purchase on the instance. Therefore, also on each page the
report shows the provisioning distribution for the OS, which is the
percentage of the instances that would be over-provisioned, underprovisioned, optimal, and idle if they were not actively provisioned.

www.cloudamize.com

AWS

www.cloudamize.com

Distribution of Operating Systems


Migrating to AWS
55%

53%

50%

45%

40%

Percentage

35%

30%

25%
19%
20%

15%
10%
8%

10%

3%

5%

2008

2003

2012

RHEL

3%

2%

Ubuntu CentOS Linux

1%
SUSE

Microsoft Windows
Operating System Type

www.cloudamize.com

AWS OS Type:
Microsoft Windows 2008
INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

36%
Average Cost
Savings

45%

40%

22%
35%

32%

44%

Percentage

30%

33%
13%

Average Cost
Savings

25%

6%

6%

Over-provisioned

Average Cost
Savings

Average Cost
Savings

Under-provisioned

20%

Idle

2%

Average Cost
Savings

15%

%
Optimal

Average Cost
Savings

10%

5%

42%
rge

m3.la

15%
edium

m3.m

1O%
ge

r3.lar

8%
rge

c3.la

5%
rge

c3.xla

20%
rs

othe

Instance Types

www.cloudamize.com

AWS OS Type:
Microsoft Windows 2003
INSTANCE TYPE
DISTRIBUTION
45%

PROVISIONING
DISTRIBUTION

15%
Average Cost
Savings

40%

41%
Average Cost
Savings

35%

20%
29%

30%

9%

11%
Percentage

36%

Average Cost
Savings

25%

18%
Average Cost
Savings

12

20%

Over-provisioned
Under-provisioned

Average Cost
Savings

Idle

15%

Optimal

4%
Average Cost
Savings

10%

5%

32%
edium

m3.m

31%
rge

m3.la

15%
rge

c3.la

6%
rge

c3.xa

4%
ge

r3.lar

12%
rs

othe

Instance Types

www.cloudamize.com

AWS OS Type:
Microsoft Windows 2012
INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

45%

33%
Average Cost
Savings

40%

35%

20%

30%

22%

9%

Percentage

Average Cost
Savings

55

25%

Average Cost
Savings

8%
Average Cost
Savings

20%

1%

15%

Average Cost
Savings

9%
Over-provisioned
Under-provisioned

Average Cost
Savings

5%

49%

Idle
Optimal

9%
Average Cost
Savings

10%

5%

43%
rge

m3.la

20%
ge

r3.lar

9%

6%

4%

4%

14%

large r3.xlarge
large 3.medium others
m3.x
m

m2.x

Instance Types

www.cloudamize.com

10

AWS OS Type: RHEL

INSTANCE TYPE
DISTRIBUTION
45%

PROVISIONING
DISTRIBUTION

22%
Average Cost
Savings

40%

14%
Average Cost
Savings

35%

11%
%

42

4%

30%

40

Average Cost
Savings
Percentage

44%

Average Cost
Savings

9%

25%

16%

Over-provisioned

Average Cost
Savings

20%

Under-provisioned
Idle

4%

15%

Optimal

Average Cost
Savings

10%

5%

31%
rge

m3.la

21%
ge

r3.lar

13%
edium

m3.m

10%
large

m2.x

8%
rge

c3.la

17%
rs

othe

Instance Types

www.cloudamize.com

11

AWS OS Type: Ubuntu

INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

45%

40%

12%
Average Cost
Savings

7%

35%

29%
30%

62%

6%

2%

Percentage

Average Cost
Savings

25%

13

Average Cost
Savings

Average Cost
Savings

Average Cost
Savings

20%

5%

Over-provisioned
Under-provisioned
Idle

23

15%

Optimal

Average Cost
Savings

5%
Average Cost
Savings

10%

5%

24%
rge

m3.la

20%
edium

m3.m

11%

9%

large

m2.x

8%
e

xlarg

m2.2

rge

c3.la

7%
arge

c3.xl

7%
ge

r3.lar

Instance Types

www.cloudamize.com

12

AWS OS Type: CentOS

INSTANCE TYPE
DISTRIBUTION

50%

PROVISIONING
DISTRIBUTION

45%

45%

Average Cost
Savings

3%

40%

35%

38%

Percentage

30%

57%

5%
Average Cost
Savings

5%

25%

2%

10%

Average Cost
Savings

Average Cost
Savings

Average Cost
Savings

20%

15%

Under-provisioned
Idle
Optimal

4%

15%

Over-provisioned

Average Cost
Savings

19%

10%

Average Cost
Savings

5%

47%
rge

m3.la

16%
edium

m3.m

7%

6%

6%

3%

e
large c3.large
edium r3.4xlarg
ml.m

m2.x

15%
rs

othe

Instance Types

www.cloudamize.com

13

AWS OS Type: Linux

INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

40%

46%

35%

Average Cost
Savings

13%
30%

Average Cost
Savings

6%

Percentage

9%

17%

Average Cost
Savings

20%

30%
58%

Average Cost
Savings

25%

9%

Over-provisioned

Average Cost
Savings

Under-provisioned
Idle

11%

15%

3%

Average Cost
Savings

Optimal

10%

5%

20%
ge

r3.lar

16%
rge

m3.la

13%
edium

m3.m

13%
rge

r3.xla

9%
large

m2.x

29%
rs

othe

Instance Types

www.cloudamize.com

14

AWS OS Type: SUSE

INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

40%

35%

23%

19%
30%

Average Cost
Savings

11%

46

Percentage

Average Cost
Savings

25%

12%

9%

Average Cost
Savings

20%

Average Cost
Savings

60%

Over-provisioned

Average Cost
Savings

Under-provisioned

3%

Idle

Average Cost
Savings

15%

9%
8%

Optimal

10%

5%

24%
ge

r3.lar

23%
rge

m3.la

11%
arge

c3.xl

11%
rge

r3.xla

10%
edium

m2.m

21%
rs

othe

Instance Types

www.cloudamize.com

15

Azure

www.cloudamize.com

16

Distribution of Operating Systems


Migrating to Azure
55%

53%

50%

45%

40%

Percentage

35%

30%

25%
21%
20%

17%

15%

10%

5%

4%

2008

2003

2012

RHEL

2%

1%

1%

1%

Ubuntu SUSE CentOS Linux

Microsoft Windows
Operating System Type

www.cloudamize.com

17

Azure OS Type:
Microsoft Windows 2008
INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

26%
Average Cost
Savings

40%

35%

25%

24%
30%

Average Cost
Savings

1%

10%

Percentage

Average Cost
Savings

25%

9%

4%

Average Cost
Savings

20%

11

63%

Over-provisioned

Average Cost
Savings

Under-provisioned

3%

Idle

Average Cost
Savings

15%

Optimal

10%

5%

16%
.a3

basic

12%
.a1

basic

11%
.a2

basic

8%

9%
1

std.d

s2
std.d

44%
rs

othe

Instance Types

www.cloudamize.com

18

Azure OS Type:
Microsoft Windows 2012
INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

22%
Average Cost
Savings

40%

35%

30%

19%

1%

6%
Average Cost
Savings

7%

73%

Percentage

Average Cost
Savings

25%

3%

5%

Average Cost
Savings

20%

Over-provisioned

Average Cost
Savings

Under-provisioned
Idle

15%

Optimal

Average Cost
Savings

10%

5%

17%
.a1

basic

14%
1

std.d

10%
.a2

basic

9%
.a2

basic

7%
5

std.a

43%
rs

othe

Instance Types

www.cloudamize.com

19

Azure OS Type:
Microsoft Windows 2003
INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

40%

35%

30%
Average Cost
Savings

4%

12%

Average Cost
Savings

29%

Average Cost
Savings

60%

30%

Percentage

2%

Average Cost
Savings

25%

9%

5%
Average Cost
Savings

Over-provisioned

20%

Under-provisioned

6%

Idle

Average Cost
Savings

15%

Optimal

10%

5%

22%

17%

.a1

std.d

basic

14%
.a3

basic

11%
.a2

basic

9%
s2
std.d

27%
rs

othe

Instance Types

www.cloudamize.com 20

Azure OS Type: RHEL

INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

27%
45%

Average Cost
Savings

2%

40%

10%

18%

35%

Average Cost
Savings

88%

30%

Percentage

9%
Average Cost
Savings

9%

25%

Over-provisioned

Average Cost
Savings

Under-provisioned

20%
Idle

5%
Average Cost
Savings

15%

Optimal

5%

10%

Average Cost
Savings

5%

39%
.a1

basic

24%
.a0

basic

8%
.a3

basic

9%
.a2

basic

3%
.a4

basic

17%
rs

othe

Instance Types

www.cloudamize.com

21

Azure OS Type: Ubuntu

INSTANCE TYPE
DISTRIBUTION

45%

40%

PROVISIONING
DISTRIBUTION

18%
Average Cost
Savings

35%

3%
Average Cost
Savings
Percentage

13%

Average Cost
Savings

5%

30%

6%
3%

25%

78%

8%
Average Cost
Savings

20%

Over-provisioned
Under-provisioned

15%

Idle

6%

8%

Average Cost
Savings

Average Cost
Savings

Optimal

10%

5%

28%
.a1

basic

21%
.a0

basic

9%
.a4

basic

9%
.a5

basic

9%
sl2

std.d

24%
rs

othe

Instance Types

www.cloudamize.com 22

Azure OS Type: SUSE

INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

45%

40%

2.6%

1%

Average Cost
Savings

Average Cost
Savings

35%

6%
%

2
2%

90%

4%

30%

Percentage

Average Cost
Savings

25%
Over-provisioned

35

20%

2%

Average Cost
Savings

15%

Idle
Optimal

0.3%

0.4%

Average Cost
Savings

10%

Under-provisioned

Average Cost
Savings

Average Cost
Savings

5%

22%
.a1

basic

20%
.a3

basic

13%

7%

s4

std.g

.a4

basic

7%
.a2

basic

4%
.a0

basic

27%
rs

othe

Instance Types

www.cloudamize.com 23

Azure OS Type: CentOS

INSTANCE TYPE
DISTRIBUTION

PROVISIONING
DISTRIBUTION

45%

3%

40%

Average Cost
Savings

34

35%

Average Cost
Savings

5%

12%

1%

30%

Percentage

2%

81%

Average Cost
Savings

25%
Over-provisioned

1%
1%

Average Cost
Savings

20%

15%

2%

Idle
Optimal

9%

Average Cost
Savings

10%

Under-provisioned

Average Cost
Savings

Average Cost
Savings

5%

42%
.a1

basic

12%
.a0

basic

7%

5%

s2

std.g

.a3

basic

5%
s1

std.d

5%
s4

std.g

24%
rs

othe

Instance Types

www.cloudamize.com 24

Azure OS Type: Linux

INSTANCE TYPE
DISTRIBUTION

45%

PROVISIONING
DISTRIBUTION

18%
40%

Average Cost
Savings

12%

35%

8%

1%

Percentage

Average Cost
Savings

Average Cost
Savings

30%

18%
70%

7%

25%

Average Cost
Savings

20%

50%

Over-provisioned

Average Cost
Savings

Under-provisioned
Idle

19%

15%

Optimal

Average Cost
Savings

10%

5%

36%
.a1

basic

19%
.a0

basic

12%
.a2

basic

7%

7%
1

std.d

1v2

std.d

19%
rs

othe

Instance Types

www.cloudamize.com 25

State of Provisioning on AWS vs. Azure

19%

Over-provisioned

33%

Under-provisioned

AWS

38%

Over-provisioned

Idle

9%

Optimal

22%

Under-provisioned
Idle
Optimal

19%

68%

AZURE

9%

www.cloudamize.com 26

Takeaways
Companies migrating to the cloud can realize significant cost savings
annually:

Mapping these OS types to these instances results in an average


of 18% cost savings on AWS vs. 14% cost savings on Azure.

Average cost-savings on AWS for Windows is 22% and for Linux


is 15%.

Average cost savings on Azure for Windows is 17% and for Linux
is 10%.

However, there are significant opportunities to improve cost savings


(and performance) even more in the cloud. Companies can save more
by actively provisioning the capacity they purchase on each instance
on an ongoing basis. In this report:

Only 19% and 22% of the instances on AWS and Azure


respectively are optimally provisioned just as they are, without
right-sizing for capacity.

33% of the instances on AWS are over-provisioned, while 33%


and 19% of the instances on AWS and Azure respectively are idle.
Companies are paying for more than they need, and paying for
what theyre not even using.

9% of the instances on both AWS and Azure are underprovisioned, meaning they will exceed their capacity and suffer
performance issues.

www.cloudamize.com

27

About Cloudamize
Mapping to the right instances in the cloud and right-sizing them on
an ongoing basis for the best performance and cost benefits requires
in-depth infrastructure analytics. For more information on the analytics
you need and which analytics were used to map instances in this
report, see our webinar:
How to Plan a Successful Migration to the Cloud.
Cloudamize is a cloud infrastructure analytics platform that helps
companies maximize the value of their cloud investments. Our SaaSbased platform provides deep visibility into your infrastructure and
actionable analysis that empowers you to make data-driven decisions
with confidence at each stage of your cloud journey from assessing
options and planning migrations to managing cloud deployments.
With Cloudamize, companies have experienced significant monthly
savings on their cloud deployments while simultaneously increasing
performance.

www.cloudamize.com 28

1735 Market St. Suite 2502


Philadelphia, PA 19103
215-557-3735
info@cloudamize.com

You might also like