34
Survey data gathered by T. Bell, T. Fifield, R. Lane, J.C. Martin 13 th May 2014 OpenStack User Survey and Feedback

OpenStack Atlanta User Survey

Embed Size (px)

DESCRIPTION

User survey results for the Atlanta OpenStack Summit.

Citation preview

Page 1: OpenStack Atlanta User Survey

!

Survey data gathered by T. Bell, T. Fifield, R. Lane, J.C. Martin !

13th May 2014

OpenStack User Survey and Feedback

Page 2: OpenStack Atlanta User Survey

• Survey • Feedback and Comments • Ongoing Activities • How to Get Involved

AGENDA

Page 3: OpenStack Atlanta User Survey

• What was new ? • How do you consume OpenStack ? • Support multiple entries for different

clouds (dev/QA/prod) • Additional input for comments

SURVEY SPRING 2014

Page 4: OpenStack Atlanta User Survey

1780506

Top 10 Countries

(deployments)

107

13

266

195

1921

512 Companies293 UG Members

DEPLOYMENT SURVEY

USCNINFRGBCAAUCHDEBR 9

14

14

18

21

22

31

35

36

171

20

31

255143

37

460

DeploymentsSurvey Responses

Top 10 Countries

(all)USCNINFRGBCAAUDEJP

BR 19

25

30

30

36

32

53

78

83

418

Page 5: OpenStack Atlanta User Survey

DEPLOYMENT SURVEY

10,000+ 16%

5,001-10,000 8%

1,001-5,000 10%

501-1,000 6% 101-500

13%

21-100 17%

1-20 30%

Organization Size (employees)Avoiding vendor lock-in

Cost savingsOperational efficiency

Open technologyFlexibility of underlying technology choices

Ability to innovate/competeTime to market

Attracting talentControl

Other 1084

96119

157160166170170174

Business Drivers

Dev/QA

Proof of Concept

Production 209

218

210

Stage of Deployment

Page 6: OpenStack Atlanta User Survey

DEPLOYMENT SURVEY

Other 20%

Telecommunications 6%

Academic / Research 16%

Information Technology 58%

Other 30%

unspecified 10%Film/Media

18%

Government / Defense 12%

Retail 8%

Manufacturing/Industrial 7%

Healthcare 6%

Finance 5% Energy

2%

Transportation/Shipping 1%

Consumer Goods 1%

Industry

Page 7: OpenStack Atlanta User Survey

DEPLOYMENT SURVEY

docs.openstack.org

Blogs

OpenStack Mailing List

The OpenStack Operations Guide

Launchpad Answers

IRC

OpenStack Dev Mailing List

ask.openstack.org

Other Online Forums

forums.openstack.org

OpenStack Planet

Source Code

Other 73

81

98

106

248

264

340

344

400

485

512

516

759

Information sources User Types

Cloud Consumer

Ecosystem Vendor

Service Provider

Cloud operator 522

436

369

273

On-Premise Private Cloud

Hosted Private Cloud

Public Cloud

Hybrid Cloud

Community Cloud 6

22

57

65

224

Deployment Types

Page 8: OpenStack Atlanta User Survey

DEV/QA

unspecified1-100

101-500501-1,000

1,001-5,0005,001-10,000

10,000+50,000+ 0

141110

42107

19

# of coresunspecified

1-10,00010,001-100,000

100,001 to 1 million1-100 million

100-500 million500+ million 1

04819

7191

# of objects

unspecified

1-100

101 to 1,000

1,001 to 10,000

10,000+ 4

15

53

80

42

# of IPs

Unspecified0-10 TB

11-100 TB100-500 TB

500+ TB1,000+ TB 2

9

8

46

81

48

Block Storage Size

unspecified1-100

101-500501-1,000

1,001-5,0005,000-10,000

10,000+ 43151628

10919

# of instancesunspecified

1-5051-100

101-500501-1,000

1,000+ 2

4

9

14

147

18

# of compute nodes

Page 9: OpenStack Atlanta User Survey

DEV/QA

Austin (2010.1)

Bexar (2011.1)

Cactus (2011.2)

Diablo (2011.3)

Essex (2012.1)

Folsom (2012.2)

Grizzly

Havana (2013.2)

Icehouse (2014.1)

Trunk 28

14

47

91

74

43

7

3

2

2

Release

Openstack Compute (Nova)

Openstack Image Service (Glance)

Openstack Identity Service (Keystone)

Openstack Dashboard (Horizon)

Openstack Network (Neutron)

Openstack Block Storage (Cinder)

Openstack Object Storage (Swift)

Ceilometer

Heat

OpenStack Database as a Service (Trove)

OpenStack Bare Metal (Ironic) 6

9

43

51

113

159

169

180

188

195

204

Services

Page 10: OpenStack Atlanta User Survey

DEV/QA

Puppet

DevStack

Chef

Other 36

22

26

42

PackStackSaltStackCrowbar

OtherNone

CFEngineForeman

FuelStackOps

SubstratumFAI

CustomAnvil

Ansible 0000111112355

16

Deployment Tool

kvm

xen

esx

Other 82

22

30

175

QEMUlxc

Bare Metalhyperv

xenserverDocker

OpenVZOther

PowerVM 3

3

5

7

9

9

11

16

19

Hypervisor

SQLLDAPPAMKVS

ADTemplated 1

31011

5892

Identity Driver

Open vSwitch

Linux Bridge

Cisco

Other 92

24

59

97

NiciraBrocadeHyper-V

nova-networkModular Layer 2 Plugin

Big SwitchRyu

NECMidoNet

OtherPLUMgrid

MellanoxJuniperCustom

AristaEmbrane

Extreme NetworksRuijie Networks

ML2 0111123334456889111012

Network Driver

JSON

XML 36

66

API Format

Page 11: OpenStack Atlanta User Survey

DEV/QA

UbuntuCentOS

Red Hat Enterprise LinuxWindows

DebianFedora

SUSE Linux EnterpriseFreeBSD

Scientific LinuxOther

openSUSE 0011234921

3666

Operating SystemPrefer not to say

1-100101-1,000

1,001-5,0005,001-10,000

10,001-50,000100,000+ 2

00717

9688

# of users

EC2 compatibility API

S3 compatibility API

OCCI compatibility API

GCE compatibility API 1

4

54

71

Compatibility APIs

LVMCeph RBD

NFSGlusterFS

NetAppNexenta

WindowsSAN/HP

EMCXIV

SAN/SolarisStorwizeSolidFire

CoraidXenapi

SheepdogZadara

HP 3PARHuawei

HDSHP LeftHand

ScalityDell EqualLogic

EqualLogicMellanox

OtheriSCSI 1

1111222334444355667881723

3851

89

Block Storage Drivers

Page 12: OpenStack Atlanta User Survey

PROOF OF CONCEPT

unspecified1-100

101-500501-1,000

1,001-5,0005,001-10,000

10,000+50,000+ 0

4165

34128

25

# of cores

unspecified

1-100

101 to 1,000

1,001 to 10,000

10,000+ 1

11

51

93

47

# of IPs

unspecified1-10,000

10,001-100,000100,001 to 1 million

1-100 million100-500 million

500+ million 0025

208294

# of objectsunspecified

0-10 TB11-100 TB

100-500 TB500+ TB

1,000+ TB 0

8

6

37

98

54

Block Storage Size

unspecified1-100

101-500501-1,000

1,001-5,0005,000-10,000

10,000+ 2171228

12825

# of instancesunspecified

1-5051-100

101-500501-1,000

1,000+ 2

5

6

7

159

24

# of compute nodes

Page 13: OpenStack Atlanta User Survey

PROOF OF CONCEPT

Austin (2010.1)

Bexar (2011.1)

Cactus (2011.2)

Diablo (2011.3)

Essex (2012.1)

Folsom (2012.2)

Grizzly

Havana (2013.2)

Icehouse (2014.1)

Trunk 31

10

37

102

87

38

10

3

4

2

Release

Openstack Compute (Nova)

Openstack Image Service (Glance)

Openstack Identity Service (Keystone)

Openstack Dashboard (Horizon)

Openstack Block Storage (Cinder)

Openstack Network (Neutron)

Openstack Object Storage (Swift)

Ceilometer

Heat

OpenStack Database as a Service (Trove)

OpenStack Bare Metal (Ironic) 8

10

35

46

116

156

158

171

176

182

202

Services

Page 14: OpenStack Atlanta User Survey

PROOF OF CONCEPT

Puppet

DevStack

Chef

Other 45

23

28

37

PackStackCrowbar

SaltStackFuel

NoneOther

ForemanSubstratum

AnsibleStackOps

AnvilCFEngine

CustomFAI 0

0001122222

81114

Deployment Tool

kvm

esx

xen

Other 83

37

38

169

lxc

xenserver

QEMU

hyperv

Bare Metal

Docker

OpenVZ

Other

PowerVM 2

2

3

4

10

13

15

17

17

Hypervisor

SQLLDAPPAMKVS

ADTemplated 2

41117

6783

Identity Driver

Open vSwitch

Linux Bridge

Cisco

Other 80

30

58

95

Hyper-VBrocade

NiciraBig Switch

nova-networkRyu

JuniperMidoNet

Modular Layer 2 PluginNEC

PLUMgridOther

CustomMellanox

AristaEmbrane

Extreme NetworksML2

Ruijie Networks 0000011244456688101011

Network Driver

JSON

XML 56

132

API Format

Page 15: OpenStack Atlanta User Survey

PROOF OF CONCEPT

Prefer not to say1-100

101-1,0001,001-5,000

5,001-10,00010,001-50,000

100,000+ 221515

93100

# of usersUbuntuCentOS

Red Hat Enterprise LinuxDebian

WindowsSUSE Linux Enterprise

openSUSEFedora

FreeBSDOther

Scientific Linux 11124788

2242

66

Operating System

EC2 compatibility API

S3 compatibility API

GCE compatibility API

OCCI compatibility API 3

3

50

68

Compatibility APIs

LVMNFS

Ceph RBDNetApp

GlusterFSEMC

WindowsSAN/HP

XenapiNexenta

SAN/SolarisXIV

SolidFireStorwize

CoraidHP LeftHand

HDSHuawei

HP 3PARZadaraScality

SheepdogZFS

CustomDell EqualLogic

EqualLogicIBM GPFS

MellanoxOther

XenAPI Storage Manager 1111111223333334555678911132327

4249

83

Block Storage Drivers

Page 16: OpenStack Atlanta User Survey

PRODUCTION

unspecified1-100

101-500501-1,000

1,001-5,0005,001-10,000

10,000+50,000+ 2

1371814

4879

22

# of cores

unspecified

1-100

101-1,000

1,001-10,000

10,000+ 9

28

59

52

55

# of IPs

unspecified1-10,000

10,001-100,000100,001 to 1 million

1-100 million100-500 million

500+ million 2341223

49110

# of objectsunspecified

0-10 TB11-100 TB

100-500 TB500+ TB

1,000+ TB 2

13

13

57

59

59

Block Storage Size

unspecified1-100

101-500501-1,000

1,001-5,0005,000-10,000

10,000+ 96

2420

4477

23

# of instances

unspecified1-50

51-100101-500

501-1,0001,000+ 12

71721

12521

# of compute nodes

Page 17: OpenStack Atlanta User Survey

PRODUCTION

Austin (2010.1)

Bexar (2011.1)

Cactus (2011.2)

Diablo (2011.3)

Essex (2012.1)

Folsom (2012.2)

Grizzly

Havana (2013.2)

Icehouse (2014.1)

Trunk 26

5

42

84

64

61

10

2

2

1

Release

Openstack Compute (Nova)

Openstack Image Service (Glance)

Openstack Identity Service (Keystone)

Openstack Dashboard (Horizon)

Openstack Block Storage (Cinder)

Openstack Network

Openstack Object Storage (Swift)

Ceilometer

Heat

OpenStack Database as a Service (Trove)

OpenStack Bare Metal (Ironic) 5

7

26

53

101

135

151

160

185

187

196

Services

Page 18: OpenStack Atlanta User Survey

PRODUCTION

Puppet

Chef

DevStack

Other 31

18

32

65

PackStackSaltStack

OtherCrowbarCustom

SubstratumAnsible

AnvilCFEngine

FAIStackOpsForeman

FuelNone 0

00111112223

611

Deployment Tool

kvm

esx

xen

QEMU

Other 40

18

18

19

173

xenserver

lxc

hyperv

Bare Metal

OpenVZ

Docker

Other

PowerVM 1

1

2

5

5

6

10

10

Hypervisor

SQLLDAP

KVSPAM

ADTemplated 0

368

55112

Identity Driver

Open vSwitch

Linux Bridge

nova-network

Other 74

23

41

84

Network Driver

CiscoNicira

Hyper-VBrocadeCustom

NECModular Layer 2 Plugin

Big SwitchMidoNet

RyuJuniper

AristaPLUMgrid

Extreme NetworksOtherML2

MellanoxRuijie Networks

Embrane 00011122233334455

1520

JSON

XML 47

135

API Format

Page 19: OpenStack Atlanta User Survey

PRODUCTION

Prefer not to say1-100

101-1,0001,001-5,000

5,001-10,00010,001-50,000

100,000+ 4

3

4

14

38

72

74

# of usersUbuntuCentOS

Red Hat Enterprise LinuxWindows

DebianSUSE Linux Enterprise

openSUSEFedora

Scientific LinuxFreeBSD

Other 112233782031

89

Operating System

EC2 compatibility API

S3 compatibility API

OCCI compatibility API

GCE compatibility API 0

4

54

87

Compatibility APIs

LVMCeph RBD

NFSNetApp

GlusterFSEMC

WindowsSAN/Solaris

SAN/HPNexentaStorwizeSolidFire

SheepdogHP LeftHand

ZadaraXIV

CoraidCustomScalityViolin

XenapiHuawei

Dell EqualLogicEqualLogic

HDSHP 3PAR

IBM GPFSXenAPI Storage Manager

ZFS 1111111222223444455678912172830

5863

Block Storage Drivers

Page 20: OpenStack Atlanta User Survey

ICEHOUSE/JUNO COMPARISON

User Types:

Iceh

ouse

Juno

Clo

ud C

onsu

mer

Ecos

yste

m V

endo

rSe

rvic

e Pr

ovid

erC

loud

ope

rato

rC

loud

Con

sum

erEc

osys

tem

Ven

dor

Serv

ice

Prov

ider

Clo

ud o

pera

tor

522

377436

343 369290 273

194

Dev/QA Proof of Concept ProductionIcehouse Juno Icehouse Juno Icehouse Juno

209218210165178173

Stage of Deployment

Survey Responses Deployments Companies UG MembersIcehouse Juno Icehouse Juno Icehouse Juno Icehouse Juno

293

512506

629

216

539

387

822General Stats

Page 21: OpenStack Atlanta User Survey

APP SURVEY

OpenStack clients

jclouds

None

FOG

Deltacloud

php-opencloud

pkgcloud

OpenStack.net

Other 35

7

17

17

29

33

34

35

139

226 Survey Responses71 Companies81 UG Members

Application Libraries

CustomLibcloud

Euca/BotoCloudforms

JujuOCCIMoxieBosh

SlipStreamHybridfox

RallyVagrant

Other 15111111222224

Python

Java

Ruby

PHP

Perl

Node.js

C/C++

C#

Other 26

8

15

16

17

29

34

45

123

Languages

BashGo

ScalaClojure

JuliaOther 14

11235

VimEclipse

SublimeOtherIntelliJEmacs

Visual StudioAtom

PycharmOther 5

33911172033

69121

Developer EnvsLinux

Windows

Mac OS X

Solaris

BSD 2

4

40

40

184

OS

PuppetHeat

DockerChef

AnsibleOpenShiftSaltStack

Cloud FoundryOther 23

14233137

6468

85111

Config Tools

JSON

XML 33

136

API FormatOpen technology

Avoiding vendor lock-inAbility to innovate/compete

Cost savingsOperational efficiency

Flexibility of underlying technology choicesTime to market

Attracting talentControl

Other 843

5375

101104105106

110114

Business Drivers

Page 22: OpenStack Atlanta User Survey

APP SURVEY

Information TechnologyAcademic / ResearchTelecommunications

Film/MediaGovernment / Defense

Manufacturing/IndustrialFinance

HealthcareInsurance

Consumer GoodsRetail

EnergyOther 8

1111334461931

150

Industry

Top 10 Countries

USFRJP

CNGBAUINDECANL 5

577810121315

87

10,000+ 17%

5,001-10,000 14%1,001-5,000

10%501-1,000

6%

101-500 13%

21-100 11%

1-20 28%

Organization Size

Cloud operator

Service Provider

Ecosystem Vendor

Cloud Consumer 67

76

82

125

User Types

docs.openstack.orgBlogs

OpenStack Mailing ListThe OpenStack Operations Guide

ask.openstack.orgIRC

Launchpad AnswersOpenStack Dev Mailing List

Source CodeOther Online Forums

OpenStack PlanetOther 33

43677178808190

111112117

170

Info Sources

Page 23: OpenStack Atlanta User Survey

• Users asked for free-form feedback (including some requests from the projects) • Foundation Priorities • What they like about OpenStack • What they feel should be the emphasis for 2014 • Neutron compared with nova network • What do we do with the comments ? • Anonymise and classify • Distribute summaries and details to the PTLs prior to the summit • Provide input to design summit discussions • Grouped into • What people like about OpenStack – 202 comments • User committee and Foundation priorities for 2014 – 252 comments • Further enhancements – 431 comments • Why nova-network – 38

COMMENTS

Page 24: OpenStack Atlanta User Survey

Over the last 6 months I think the documentation has improved significantly. I have seen others recently hold up OpenStack docs as a model to strive for.

MY FAVORITE COMMENT

Page 25: OpenStack Atlanta User Survey

• Community • Open Source • Flexibility • Extendible • Code is accessible and easy to understand • ‘Not a Lot’ • Scalable • Easy automation • ‘It pays my salary’

WHAT PEOPLE LIKED ABOUT OPENSTACK

Page 26: OpenStack Atlanta User Survey

• Channel operations needs to development and companies • Arrange regular feedback from user groups and ambassadors • Close the gap for Enterprise IT • Industry specific working groups • Focus on those consuming OpenStack • Expand training and certification marketplace • Define what OpenStack is • Improve/replace Launchpad

USER COMMITTEE/FOUNDATION PRIORITIES

Page 27: OpenStack Atlanta User Survey

• Stability of core should be a priority above adding new functions • Add how-to guides, problem management documentation, expire old

documentation, end user guide (but much less than previous surveys) • Zero downtime migrations • Installation and configuration (but much less than previous surveys) • Cross Project consistency with APIs, SDKs and CLIs • High availability VMs • Neutron stability, simplification, resilience, IPv6 and scalability • Improved function and usability in Horizon • Security, auditing (but much less than previous survey) • AWS/EC2 compatibility

FOCUS FOR FURTHER ENHANCEMENTS

Page 28: OpenStack Atlanta User Survey

• Mid-cycle gathering of operators • At eBay in San Jose – Thanks Subbu • Moderated by Sean Roberts and Tom Fifield • 50+ attendees • Take-away • Very positive atmosphere – no tranquilizers needed • Operators want to get involved early – Blueprint-on-blueprints • Many volunteers to help ‘outside of the day job of running OpenStack clouds’ • Lots of ideas for the summit • Aim to repeat in August/September 2014 • ‘Brief’ summary at https://etherpad.openstack.org/p/operators-

feedback-mar14

OPERATOR MINI-SUMMIT

Page 29: OpenStack Atlanta User Survey

• Operators Meetup • Modelled on design summit – no slides! • Monday afternoon and Friday morning • http://junodesignsummit.sched.org/overview/type/ops+meetup • Dev/Ops Design Summit sessions • Each project has a Dev/Ops session for in-depth conversations on needs, usage and

priorities • http://junodesignsummit.sched.org/?s=Dev%2FOps • Volunteers for design sessions • Follow the work of blueprint-on-blueprints to give operator input at the design stage • Volunteer at https://etherpad.openstack.org/p/juno-summit-ops-volunteer

ACTIVITIES AT THE SUMMIT

Page 30: OpenStack Atlanta User Survey

• User Committee changes and scale-out • More activities, need more help • Session to discuss how to structure the feedback on Friday • http://junodesignsummit.sched.org/event/797439515b4bd045895dc8195e72d317 • There is also a conference going on ☺ • 25 sessions in the operators track • 13 talks in the community building • Lots of other user related sessions in the project specific tracks

ACTIVITIES AT THE SUMMIT (II)

Page 31: OpenStack Atlanta User Survey

• If you are interested, have the time to help and the skills to bridge from user feedback to development, please get in touch with us • Volunteer at https://docs.google.com/forms/d/

1HOwsPp44fNbWv9zgvXW8ZnCaKszg_XKu7vmLbrPFMzQ/viewform • [email protected] • Follow up discussions to refine requirements and maximise inclusion • Operations mini-summit on Friday https://etherpad.openstack.org/p/ATL-ops-

unconference-RFC • [email protected] • Get involved in blueprint-on-blueprints process to improve the design

before coding starts

CAN YOU HELP ?

Page 32: OpenStack Atlanta User Survey

• Survey remains open at http://www.openstack.org/user-survey • Survey November 2013 at http://www.openstack.org/assets/Uploads/

Deployments-IceHouse-v1.1.pptx • Survey April 2013 at http://www.openstack.org/summit/portland-2013/

session-videos/presentation/openstack-user-committee-update-and-survey-results

REFERENCES

Page 33: OpenStack Atlanta User Survey

Backup Slides

Page 34: OpenStack Atlanta User Survey

• Ensure compatibility across different implementations • Stability/Hardening of existing functionality • Establish certification program • Encourage training for different skill sets • Support user groups, especially outside U.S.

FOUNDATION PRIORITIES (FEEDBACK FROM APRIL 2013)