last year`s survey

THE CURREN T S TAT E OF
CONT AINE R US AGE
IDENTIFYING AND ELIMINATING BARRIERS TO ADOPTION
W H AT ’ S I N S I D E
Section 1: Executive S ummary
3
4.3 A re you running a database or other stateful
service in a container environment in
Section 2: User Pro les
2 .1
Us er pro
4
s
production today?
4.4 Which database or other stateful service are
2 .2 Company s ize
you using or do you plan to use in your
container environment?
Section 3: Container Usage
6
3 .1 Contai ner u sage
Section 5: Conclusion
3 .2 Contai ner of cho ice
3 .3 Wh ere are they using c ontainers
3 .4
R unn ing containers in a VM environm ent
3 .5 Clo ud or on-prem
3 .6 Wh at container orchestration tools do you use
(if a ny)?
3 .7 Wh at are the biggest barrie rs to putti ng
co nta iners in a production environm ent?
Section 4: Data Management & Stateful Services
4 .1
H ow importa nt is data mana gement
to you r container strategy?
4 .2 Wh at are the most important features of data
ma n agement to your container plans?
15
19
1
Some high -level
dings of the survey in clude:
EXECUTIVE SUMMARY
O ver 94% of respond ents had eit h er inves tig ate d o r a ctu a l l y us ed
some container technology over l a s t 1 2 mo nth s .
This report is based on the
current and planned container
usage patterns of 285
respondents. The survey was
conducted over the latter half of
May 2015.
Docker was the overwhelming ch o ice of res pon de nts ( 9 0 %+ ) a s
their container of choice.
While the sampling contained
responses from organizations
of all sizes, the largest segment
of responses were from
organizations with 1 to 500
employees. Large organizations
with over 2500 employees
represented 18% of respondents.
The respondents were primarily
from development, operations or
DevO ps teams. Both security and
Q A teams were also represented.
The survey measured respondents’
current level of container usage,
the biggest obstacles to wider
container adoption, which
container platforms they use,
and what stack they are building
on. A nother section of the
survey queried data management
priorities, usage and importance.
O nly 38% of respond ents said th ey were us in g co nt ai n ers in
actual producti on environments to day (th is is s til l h ig h er th an
other data points have reporte d ) . H owever, 6 5 % of res pon d ents
stated that they are planning on us in g co ntai n ers in pro d ucti o n in
the next 12 months.
When it came to concerns and barriers to adoption:
Among the findings uncovered by the survey were insights into what respondents
perceived to be the primary barriers to container adoption. When asked to rank
several categories on a scale of “major barrier” to “no barrier,” respondents
identified the following as major to moderate: Security 61%; Data Management
53%; Networking 51%; Skills and Knowledge 48%; Persistent Storage 48%.
I n the area of data management:
Two-thirds of respond ents said data management was critica l to
success or an import ant gating facto r for their overall container
strategy.
O ver 5 3 % of respond ents said cur rent data management
cap ab ilitie s are a mod erate to major barrie r to contai ner usage.
Over 70% would like to run a database or othe r stateful service
in their container environments, with MySQL and Redis the two
leading choices (several others close behind).
The two most important features for data management in
container soluti ons were: “integratio n of data ma nagement
cap ab ilitie s into existing container wo
ow and tools” and
“seamless movement of data b etween dev, test and producti on
environm ents”.
3
2.1
U SE R PROFILES
Given the choice to identify with
more than one team function,
development was the largest
segment identif i ed. Interesting to
note, DevOps was the second
choice, even slightly higher than
operations.
Q1 Whi ch of the fo l lowing tea ms are a you a
membe r of? Ch oose all that apply.
A ns wered: 2 7 8
Development
Whether or not DevO ps is a job
title is still a hot button topic,
but people identifying as
working on DevO ps teams
probably refer more to working on
cross-functional teams.
Security and QA were also
represented in a minority of
responses. A mong the “other”
responses, there were executives
represented as well.
O ther (please specify)
Security
O perations
QA
O perations
Development
Dev O ps
Security
O ther (please specify)
5.76%
15.83%
Dev O ps
QA
Skipped: 7
50.36%
12.59%
46.04%
61.15%
4
2.2
COMPANY SIZE
The overwhelming majority of
respondents came from smaller
organizations with under 500
employees. In fact the largest
segment of responses were from
employees at organizations with
1-100.
Does this mean containers are a
startup technology? N ot from the
interest that it is garnering from
the likes of IBM, Microsoft, H P,
E MC, etc. Large organizations did
represent approximately 20% of
responses.
We expect this segment
to continue to grow as the
technology matures.
Q2 How ma ny e mployees does the
o rg an i zati o n you wo rk for have?
A ns wered: 2 8 5
1-100
101-500
Skipped : 0
501-2500
2500+
18.60%
39.65%
12.28%
29.47%
5
3.1
C O N TA I N E R U S A G E
Perhaps one of the most powerful
findings was that nearly 95%
of respondents have used or
investigated container technology
over the last 12 months.
Q3 Have yo u used or investigated any
co ntai ne r te ch n o l o gy in the las t 1 2 months?
A ns wered: 2 8 3
YES
Skipped: 2
NO
5.65%
94.35%
6
3.2
CONTAINER OF CHOICE
Bearing out the market buzz, over
92% of those surveyed have used
or investigated Docker.
LXC, the foundational container
technology in the Linux kernel
itself that Docker popularized by
exposing a simple REST API, was a
distant second, but still a solid
choice over Rocket, which was
third with only 21% of the
sample having either used or
investigated it.
Q4 Whi ch conta i n er technolog y have yo u
us e d o r investig ated?
A ns wered: 2 54
O ther
FreeBSD Jails
Skippe d: 31
Solaris Zones
LX D
Rocket
LX C
100.00%
Docker
92.13%
90.00%
80.00%
70.00%
60.00%
50.00%
40.00%
31.89%
30.00%
20.87%
20.00%
10.00%
4.33%
7.09%
10.24%
12.20%
0.00%
7
3.3
W HE R E A R E THE Y
U S I N G CO N TA I NE R S
As many have postulated, not
everyone using container
technology is deploying them
in production. In fact less than
40% of respondents are using
them in production.
Q8 Whe re are yo u using container
tech no lo g y? Choose all that a pply.
A ns wered: 2 6 6
N o plans at this time
Skipped : 1 9
5.64%
Production environment
38.72%
Q A environment
39.85%
Dev/ Test environment
72.93%
In a lab for experimental or
proof-of-concept purposes
62.78%
0%
10%
20%
30%
40%
50%
60%
70%
80%
8
3.3
W HE R E A R E THE Y
U S I N G C ON T A I N E R S,
CON T IN U E D...
A s should equally be expected,
the optimistic view of moving
more containers into production
environments (as well as into all
environments), in the next 12
months was also represented by
the responses.
Q9 Whe re are yo u planning to use container
tech no lo g y i n the nex t 1 2 months? Choose al l that apply.
A ns wered: 2 4 6
N o plans at this time
Production environment
Q A environment
Skipped: 3 9
6.91%
65.45%
63.41%
Dev/ Test environment
In a lab for experimental or
proof-of-concept purposes
76.42%
57.32%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
9
3.4
R U N NI
N IN
NG
G C O N TA I N E R S
C
I NO AN T
VAM
IN E R
N SVIIN
RO
A NVM
ME N T
E N V I R O NM E N T
While some view containers as the
death some
knell of
hypervisors,
While
view
containersthe
as the
responses
the
death
knellindicate
of hypervisors,
the
responses
indicate
the of
overwhelming
majority
overwhelming
ajority
of them
container usersmare
running
container
u sers are runn ing them
in a hypervisor.
in a h ypervisor.
Q19 A re yo u runni ng con tainers i n a
hyperv iso r v irtualized e nv ironment?
A nswered: 2 2 9
Skipped : 5 6
5.68%
21.40%
73%
Yes
No
N ot running containers at all
10
3 .. 44
RU
R U N NI
N INNGG C O N T A I N E R S
C
ET
I NONTA
A V IMNEERNSV II N
R OAN V
MM
EN
C
O
N
T
I
N
U
E
D
.
.
.
N VI RON ME N T
CON T IN U ED. . .
In terms of their hypervisor
of terms
choice,
picture
is a bitIn
of the
their
hypervisor
more
muddled.
About
of choice, the p icture isaa third
bitofore
respondents
they
use
m
m ud dled. say
About
a third
VMware.
A further
20%use
use
of
respo ndents
say they
whateverAhypervisor
VMware.
further 20%their
u se cloud
or
service
provider
uses.
For
whatever hyp ervisor their cloud
questions
containers
or
service pof
rovider
uses. vs.
For
hypervisors,
which hypervisor
questions
of containers
vs.
they use with
containers
is less
hypervisors,
which
hypervisor
important
than
the
fact
that
they use with containers is lessthey
doprun
them
in the
a hypervisor.
im
ortant
than
fact that they
do run them in a hypervisor.
Q2 0 I f yo u answered yes, which hyperv i sor
are you usi ng ? C h o ose all that apply.
A nswered : 1 7 3
Skipped : 112
2.60%
V MWare
19.70%
30.90%
Xen
Citrix XenServer
Hyper V
KVM
15.60%
I use the hypervisor provided
by my cloud provider
O ther
13.40%
8.60%
9.30%
11
3.5
C lo u d o r o n - p r e m
When asked where they were
running their container
environments, the somewhat
surprising winner was in “my own
data center.” This was followed by
Amazon Web Services (AWS) as a
close second.
Q1 1 Wh at i nfras tr ucture are you running or
do yo u pl an to r un conta iners on?
Cho o se all that apply.
A ns wered: 2 4 7
Skipped: 3 8
57.09%
60.00%
52.23%
50.00%
40.00%
30.00%
21.86%
20.00%
17.00%
17.81%
13.77%
8.10%
10.00%
0.40%
0.00%
ean
c
al O
it
Dig
0.40% 0.00% 0.40%
12
3.6
W H A T CO N T A I NER
ORC H E ST R A T I O N TO O LS
ARE YOU CONSIDERING IF ANY?
While this is still a nascent
market, some early leaders did
emerge from the responses.
Docker Swarm, the orchestration
technology from Docker itself,
was the clear winner, with nearly
50% of respondents indicating
that they planned to investigate
Swarm. Close behind were
Kubernetes and Mesos.
Q1 2 Wh i ch co ntai ner orchestration tools
a re yo u co nsi de ring? (choose all that apply)
Ans wered: 2 4 2
Skipped: 4 3
60.00%
50.00%
40.00%
49.59%
37.60%
34.71%
31.82%
30.00%
19.42%
20.00%
19.83%
16.94%
12.81%
10.00%
6.61%
0.00%
13
3.7
W H AT A R E T H E
B I G G E S T B A R R I E R S TO
P U T T I N G C O N TA I N E R S
IN A PRODUCTION
ENVIRONMENT?
In this question respondents
had the option of rating certain
categories as a major barrier,
moderate barrier, minor barrier
or no barrier at all.
Security was the highest rated
barrier to increased adoption.
The second biggest barrier was
data management.
Q10 P le a s e ra t e t h e f o ll o w in g based o n h ow
mu ch o f a b a r r ie r to ado pti o n th ey are f o r
p u t t in g co nt a in ers in a pro ducti o n
e nv iro n me nt.
A nsw e re d : 2 4 9
Sk ip p e d : 3 6
70%
60%
50%
Major Barrier
Moderate Barrier
40%
30%
Note: we combined the major and
moderate barrier responses and
grouped them to weigh biggest
barriers.
20%
10%
0%
14
4.1
HOW IMPORTANT IS
DATA MANAGEMENT TO
YOUR CONTAINER
STATEGY?
In concert with the previous
question, respondents are clear in
their belief that data management
is of critical importance to their
container strategy.
2/3 of respondents said it was
either critical or an important
gating factor to their container
adoption.
Q1 3 How i mpo rta nt is data mana geme nt to
your contai ner strategy?
A ns wered: 2 4 7
Skipped: 3 8
Critical to success
Important gating factor
Moderately important
N ot important to us
4.86%
35.22%
28.74%
31.17%
15
4.2
WHAT ARE THE MOST IMPORTANT FEATURES OF DATA MANAGEMENT TO YOUR
CONTAINER PLANS?
Respondents were asked to rank various features of data management for their container plans. They had to rank
each item from 1 to 9 with 1 being most important.
The chart summarizes barriers ranked either 1, 2 or 3 most important. Integration of data management capabilities
into existing container workf l ow and tools was the most important consideration. Seamless movement of data
between dev, test and production environments was also considered very important. Other items in the top 5 most
important considerations were f l exible shared storage, ability to achieve high availability (HA) and the ability to
update database schemas easily.
Together, these items indicate that many of the data management capabilities taken for granted in VM-based
applications, are equally important for container-based apps. Given that over half of respondents in the survey stated
that data management was a major or moderate barrier to container adoption (Report Section 3.7, Question 10), on
f i rst glance, it looks like some VM-based data management solutions are not directly applicable to containers.
Q18 Pleas e ra n k the i mpo r tance of the fo l lowi n g data management features to your container
pla ns (with 1 be i n g th e mo st i mpo r tant, a nd 9 be i n g the lea st impo rtant).
A ns wered: 21 2
Sk ipped: 7 3
Chart d ata summari z ed by barri ers ranked 1, 2, or 3 most i mpo r t an t
Fast, local, persistent storage
0.94%
Cross-cloud or cross-datacenter disaster recovery
8.96%
Cross-cloud or cross-datacenter backups
9.91%
Ability to use generic local storage instead of proprietary storage solutions
16.04%
Container live migration
16.98%
Integration of containers with existing storage solutions
18.87%
Ability to update database schema easily
38.21%
Ability to achieve high availability (HA)
41.98%
Flexible shared storage
47.64%
Seamless movement of data between dev, test and production environments
48.11%
Integration of data management capabilities into existing container workflow and tools
51.89%
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
16
4.3
W
LD
A ROEU Y
O UY O
RU
UNNING A
LDI A
K TE AT
RBO
UANSRN
U
I
NR
G
AO
AT H E R
E N
O
D
AA
TA
B FAU
SL
E SOERR O
ST
TE
V ITCHEE R
IN
SAT C
AT
E
F
U
L
S
E
R
V
I
C
E
ONTAINER
IENN A
N TE ?R
V IC
RO
ON
N TMAEI N
ENVIRONMENT?
Q1 5 Would you l ike to b e able to run a
d atabase o r other stateful serv ice i n a
container as p art of a p ro ducti on
deployment?
A nswered: 2 4 0
When asked if they would like to
When
they would
like to
be ableasked
to runif stateful
services
in
be
able to run
stateful
services in
containers,
70%
of respondents
said yes.
containers,
70% of respondents
said yes.
Skippe d : 45
15.83%
70%
14.17%
Yes
No
Unsure
17
4.4
WHICH DATABASE OR
OTHER STATEFUL
SERVICE ARE YOU USING
OR DO YOU PLAN TO
USE IN YOUR
CONTAINER
ENVIRONMENT?
In a blend of old and new, MySQL
and R edis virtually tied for first
place.
Q1 7 Wh ich s tateful services are you using
o r pl an to us e? C hoose all that apply.
A ns wered: 1 1 9
60.00%
50.00%
52.94%
49.58%
Skipped : 1 6 6
52.10%
42.86%
40.00%
36.13%
30.00%
26.89%
21.01%
20.00%
10.00%
17.65% 16.81%
7.56% 8.40%
0.00%
18
CONCLUSION
Containers are being used across businesses of all sizes. While the current usage patterns have the majority of
container environments in development or testing situations, a sizeable number of production environments have
been stood up, with many more anticipated over the next 12 months.
Docker is the undisputed king of the container space today. Its container runtime is by far the most popular choice
and its orchestration framework, Docker Swarm, is leading the pack amongst people who are looking for ways to
manage containerized distributed applications.
The biggest barriers to greater container adoption are security and data management functionality. Additionally,
users want to be able to integrate container data management functionality into their existing container workflow
and tools.
Users want to be able to run their database and other stateful processes in containers. Tools that will enable this
and manage it should find a welcome reception in the market.
The container market is the subject of intense interest and hype right now. How fast and how great the impact will
be on the IT market will be more clear over the next few months as more tools that overcome container challenges
are brought to market.
19