Mirrors,
networks, and
boundaries
what technical leaders need to know
for the next 10 years of devops
Lindsay Holmwood
Slide 2
Slide 2 text
No content
Slide 3
Slide 3 text
10 years
Slide 4
Slide 4 text
Sensitive
dependence on
initial conditions
Slide 5
Slide 5 text
What were the
initial conditions?
Slide 6
Slide 6 text
Post-GFC
scarcity
Slide 7
Slide 7 text
Organisations were
requiring increased
operational tempo
Slide 8
Slide 8 text
Siloisation
was becoming
a less viable
organisational
strategy
Slide 9
Slide 9 text
Infrastructure was
becoming
commoditised &
on-demand
Slide 10
Slide 10 text
Barrier of entry
to dev tools
had been lowered
Slide 11
Slide 11 text
No content
Slide 12
Slide 12 text
The last
10 years
Slide 13
Slide 13 text
Three lenses:
Delivery
Systems
People
Slide 14
Slide 14 text
Delivery
Slide 15
Slide 15 text
Operating
models
Delivery
Before
Waterfall
Slide 16
Slide 16 text
Operating
models
Delivery
After
Agile, Lean
Slide 17
Slide 17 text
Devops started as
“agile systems
administration”
Slide 18
Slide 18 text
Deploy
frequency
Delivery
Before
Weekly-to-monthly deploys
Slide 19
Slide 19 text
Deploy
frequency
Delivery
After
Multiple deploys a day
Slide 20
Slide 20 text
Continuous Delivery
is the standard in
web industry
Slide 21
Slide 21 text
Systems
Slide 22
Slide 22 text
Infrastructure Systems
Before
Blend of:
◦ bare metal + virtualised
◦ on-prem or “hosting”
Slide 23
Slide 23 text
Infrastructure Systems
After
IaaS, PaaS
Slide 24
Slide 24 text
Commodification
of infrastructure
Slide 25
Slide 25 text
Infrastructure
management
Systems
Before
Scripting
Slide 26
Slide 26 text
Infrastructure
management
Systems
After
Powerful config management
Slide 27
Slide 27 text
Architecture Systems
Before
Monoliths
Slide 28
Slide 28 text
Architecture Systems
After
Microservices
Slide 29
Slide 29 text
Is this reality or
aspirational?
Slide 30
Slide 30 text
Deployment
unit
Systems
Before
Virtual machines
Slide 31
Slide 31 text
Deployment
unit
Systems
After
Containers
Slide 32
Slide 32 text
Deployment
unit
Systems
Future
Functions?
Slide 33
Slide 33 text
People
Slide 34
Slide 34 text
How we
respond to
failure
People
Before
Blame culture
Slide 35
Slide 35 text
How we
respond to
failure
People
After
Just culture
Slide 36
Slide 36 text
Blameless
post-mortems
Slide 37
Slide 37 text
Relationships to
our co-workers
People
Before
Silos
Slide 38
Slide 38 text
Relationships to
our co-workers
People
After
Empathy
Slide 39
Slide 39 text
Build DevOps
culture through
shared experiences
Slide 40
Slide 40 text
Devs on call
+
Ops shipping code
Slide 41
Slide 41 text
Organising around
the deployment pipeline
Slide 42
Slide 42 text
What’s next?
Slide 43
Slide 43 text
“Prediction is very
difficult, especially
about the future.”
– Niels Bohr
Slide 44
Slide 44 text
I’m not going to
make predictions
Slide 45
Slide 45 text
Yes, but it’s not necessarily a bad thing.
As long as the technology is an enabler of the
underlying principles (communication,
collaboration, a holistic view), the DevOps
movement is still sound.
Slide 46
Slide 46 text
What are the
conditions now?
Slide 47
Slide 47 text
The DevOps dream
ain’t evenly distributed
Slide 48
Slide 48 text
No content
Slide 49
Slide 49 text
No content
Slide 50
Slide 50 text
…the overall industry
is improving its
software
development and
delivery practices
Slide 51
Slide 51 text
…low performers are
struggling to keep
up, widening the gap
Slide 52
Slide 52 text
No content
Slide 53
Slide 53 text
Increasing,
fragmented
regulation
Slide 54
Slide 54 text
◦ GDPR
◦ NDB + APP
◦ “the Netflix tax”
Slide 55
Slide 55 text
Good for society,
a PITA for us
Slide 56
Slide 56 text
No content
Slide 57
Slide 57 text
Frameworks for
accountability
Slide 58
Slide 58 text
Custom ASICs
are proliferating
Slide 59
Slide 59 text
No content
Slide 60
Slide 60 text
No content
Slide 61
Slide 61 text
No content
Slide 62
Slide 62 text
No content
Slide 63
Slide 63 text
Machine learning is
becoming
more accessible
Slide 64
Slide 64 text
Google
Neural
Machine
Translation
Source: NYT – The Great AI Awakening
Slide 65
Slide 65 text
Generative adversarial networks
Slide 66
Slide 66 text
Let’s talk about skills
for managing
uncertainty & ambiguity
National ↩︎
Organisational ↩︎
Team ↩︎
Occupational
Slide 73
Slide 73 text
Artifacts
Slide 74
Slide 74 text
physical
manifestations
of culture
Slide 75
Slide 75 text
ceremonies
Slide 76
Slide 76 text
org charts
Slide 77
Slide 77 text
desk layout
Slide 78
Slide 78 text
documentation
Slide 79
Slide 79 text
software
Slide 80
Slide 80 text
most visible parts of
an org’s culture
Slide 81
Slide 81 text
easiest part of a
culture to adopt
Slide 82
Slide 82 text
Values
Slide 83
Slide 83 text
conscious goals,
strategies, and
philosophies
Slide 84
Slide 84 text
rules that guide
how we interact
with people
Slide 85
Slide 85 text
rules that guide
how we do
our work
Slide 86
Slide 86 text
“we will dominate
the market”
Slide 87
Slide 87 text
“management is
available, and listen
to our concerns”
Slide 88
Slide 88 text
“we value quality
over delivery
speed”
Slide 89
Slide 89 text
“nobody will be
fired for making an
honest mistake”
Slide 90
Slide 90 text
values:
lived
vs
aspirational
Slide 91
Slide 91 text
Communication
We have an obligation to communicate.
Slide 92
Slide 92 text
Respect
We treat others as we would like to be
treated.
Slide 93
Slide 93 text
Integrity
We work with customers and prospects
openly, honestly, and sincerely.
Slide 94
Slide 94 text
Excellence
We are satisfied with nothing less than
the very best in everything we do.
Slide 95
Slide 95 text
Ethical
We conduct business affairs in
accordance with all applicable laws
and in a moral and honest manner.
Slide 96
Slide 96 text
No content
Slide 97
Slide 97 text
Work as imagined
vs
Work as done
Slide 98
Slide 98 text
Be clear about what
values are what
Slide 99
Slide 99 text
Assumptions
Slide 100
Slide 100 text
beliefs, perceptions,
thoughts, feelings
Slide 101
Slide 101 text
exist at an
unconscious level
Slide 102
Slide 102 text
hard to discern
Slide 103
Slide 103 text
“anyone can take
on leadership
responsibility”
Slide 104
Slide 104 text
“bad outcomes come
from bad people”
Slide 105
Slide 105 text
“it’s OK to withhold
information”
Slide 106
Slide 106 text
“individual
performance is
valued over team
performance”
Slide 107
Slide 107 text
“we can trust that team”
Slide 108
Slide 108 text
No content
Slide 109
Slide 109 text
Artifacts
Values
Assumptions
Slide 110
Slide 110 text
Our systems are
artifacts
Slide 111
Slide 111 text
Our processes are
artifacts
Slide 112
Slide 112 text
Artifacts
Values
Assumptions
Slide 113
Slide 113 text
Tools are a
snapshot of our
org’s culture
Slide 114
Slide 114 text
Tools are a
snapshot of our
org’s values and
assumptions
Slide 115
Slide 115 text
Artifacts influence
behaviour
Slide 116
Slide 116 text
Encode the org
behaviour you want
to see into your
artifacts
Slide 117
Slide 117 text
Change your org’s
values by changing
your artifacts
Slide 118
Slide 118 text
Artifact:
All changes go through
a CD pipeline.
Slide 119
Slide 119 text
Value:
We create fast feedback
loops to learn from
changes in production.
Slide 120
Slide 120 text
Artifact:
Developers and
managers do on-call
Slide 121
Slide 121 text
Value:
Performance,
availability and
sustainability are
everyone’s
responsibility
Slide 122
Slide 122 text
Artifact:
Our ceremonies
include and engage
non-technical disciplines
Slide 123
Slide 123 text
Value:
Nobody has all the
answers. We succeed by
working together.
Slide 124
Slide 124 text
But the tools are only a
means to an end
Slide 125
Slide 125 text
The goal is
transforming our
ways of working
Slide 126
Slide 126 text
No content
Slide 127
Slide 127 text
Harnessing
mirroring
Slide 128
Slide 128 text
"Organizations which design
systems are constrained to
produce designs which are
copies of the communication
structures of these
organizations."
– Melvin Conway
Slide 129
Slide 129 text
Mirroring
Slide 130
Slide 130 text
“In a complex system, the technical
architecture and the division of labor
will “mirror” one another in the sense
that the network structure of one will
correspond to the structure of the other.”
Slide 131
Slide 131 text
Two separate
research traditions
studying mirroring
Slide 132
Slide 132 text
1. Computer science
Conway’s law
Slide 133
Slide 133 text
2. Management
Org + product design &
orgs + products as
complex systems
Slide 134
Slide 134 text
What is mirroring?
Slide 135
Slide 135 text
Two networks
Slide 136
Slide 136 text
Organisational
Slide 137
Slide 137 text
Technical
Slide 138
Slide 138 text
Organisational Technical
Mirroring
Slide 139
Slide 139 text
We do this to
solve problems
Slide 140
Slide 140 text
We do this to
take people to where
the problems are
Slide 141
Slide 141 text
Who owns
this system?
? ?
?
Slide 142
Slide 142 text
We do this because
it’s economical
Slide 143
Slide 143 text
Organization design:
an information
processing view
Galbraith, 1974
Slide 144
Slide 144 text
As uncertainty increases,
the amount of
information that must be
processed by decision
makers increases
Slide 145
Slide 145 text
The org can respond by
reducing the need to
process information
Slide 146
Slide 146 text
The org can respond by
increasing the capacity to
process information
Slide 147
Slide 147 text
The org can respond by
Increase the capacity to
process information
devops
Slide 148
Slide 148 text
Creation of lateral relations:
Direct contact
Liaison roles
Task forces
Teams
Integrating roles
Managerial linking roles
Matrix organisation
Slide 149
Slide 149 text
Creation of lateral relations:
Matrix organisation
Teams
Slide 150
Slide 150 text
Creation of lateral relations:
Matrix organisation
Teams
Dev Ops Design
Frontend Backend App eng Infra UX/UI Research
Slide 151
Slide 151 text
Dev Ops Design
Frontend Backend App eng Infra UX/UI Research
Why do we stop at
dev and ops?
Slide 152
Slide 152 text
Dev Ops Design
Frontend Backend App eng Infra UX/UI Research
We can also include:
support
marketing
design
analytics
legal
finance
Slide 153
Slide 153 text
Dev Ops Design
Frontend Backend App eng Infra UX/UI Research
What happens
if we don’t?
Slide 154
Slide 154 text
Architectural Innovation: The
Reconfiguration of Existing
Product Technologies and the
Failure of Established Firms
Henderson & Clark, 1990
Slide 155
Slide 155 text
3 year study
of semiconductor
photolithographic
alignment equipment
industry
Slide 156
Slide 156 text
field based study,
high rate-of-change industry
Slide 157
Slide 157 text
4 waves of innovation
between 1962-1986
Slide 158
Slide 158 text
4 waves of innovation
new leader after each:
Kulicke ↩︎
Kasper ↩︎
Perkin-Elmer ↩︎
GCA ↩︎
Nikon
Slide 159
Slide 159 text
4 waves of innovation
each incumbent
could not course correct
Slide 160
Slide 160 text
4 waves of innovation
each incumbent
invested heavily
in new technology
Slide 161
Slide 161 text
4 waves of innovation
each incumbent
structured organisation
and communication
based on product
architecture
Slide 162
Slide 162 text
4 waves of innovation
What about this
makes sense?
Slide 163
Slide 163 text
Henderson & Clark’s
framework for defining innovation
Based on Schumpeter, 1942
Slide 164
Slide 164 text
Strongly mirrored
Broken mirror
Slide 165
Slide 165 text
Henderson & Clark’s
framework for defining innovation
Based on Schumpeter, 1942
Mirror
Don’t
mirror
Slide 166
Slide 166 text
Mapping strategy
Slide 167
Slide 167 text
Strategy is not just
having a plan –
Slide 168
Slide 168 text
– it’s understanding
how you react in a
complex environment
Slide 169
Slide 169 text
Your criteria for
making decisions
when you
face uncertainty
Slide 170
Slide 170 text
“Everyone has a plan
until they get punched
in the mouth”
– Mike Tyson
Slide 171
Slide 171 text
Mapping 101
Slide 172
Slide 172 text
Why do we map?
Slide 173
Slide 173 text
Build collective
context
Slide 174
Slide 174 text
Inform strategic
technology decisions
Slide 175
Slide 175 text
Understand and
visualise tradeoffs
Slide 176
Slide 176 text
– Sidney Dekker
“multiple overlapping and partially
contradictory descriptions of the same act
are always possible, and even necessary, to
approximate the complexity of reality”
It’s a tool that clarifies:
* Relationships
* Position in value chain
* Evolution/maturity
Slide 183
Slide 183 text
It’s a tool that clarifies:
* Hot spots
* Where new initiatives fit
Slide 184
Slide 184 text
Evolution/maturity
Visibility to customers
Slide 185
Slide 185 text
Start with user needs
Slide 186
Slide 186 text
Then add
most visible systems
Slide 187
Slide 187 text
Solid lines represent
dependencies
Slide 188
Slide 188 text
Add dependent systems
Slide 189
Slide 189 text
No content
Slide 190
Slide 190 text
No content
Slide 191
Slide 191 text
No content
Slide 192
Slide 192 text
No content
Slide 193
Slide 193 text
No content
Slide 194
Slide 194 text
No content
Slide 195
Slide 195 text
No content
Slide 196
Slide 196 text
Direction of travel
Slide 197
Slide 197 text
Visualise
opportunities
Slide 198
Slide 198 text
Keep in mind:
* Subjective
* Visibility informs priority
* Maturity informs investment
* Don’t replace arch diagrams
* Maps change over time
Slide 199
Slide 199 text
As the complexity of
a system increases,
the accuracy of any
single agent's own
model of that system
decreases rapidly.
Woods' Theorem
Slide 200
Slide 200 text
medium.com/wardleymaps
Slide 201
Slide 201 text
learn.hiredthought.com/p/wardley-mapping
Slide 202
Slide 202 text
• What are complex
system
• Simple, rugged, and
dancing landscapes
• The interesting in-
between
• Explore/exploit
Slide 203
Slide 203 text
stella.report
Slide 204
Slide 204 text
Managing risk
Slide 205
Slide 205 text
No content
Slide 206
Slide 206 text
No content
Slide 207
Slide 207 text
No content
Slide 208
Slide 208 text
Likelihood
% chance the thing
will happen in the
next 12 months
Slide 209
Slide 209 text
Impact
$ cost of impact
expressed as range*
*90% confidence interval
Slide 210
Slide 210 text
Before
Slide 211
Slide 211 text
After
Slide 212
Slide 212 text
Use maps to
understand your
risk landscape
Slide 213
Slide 213 text
1. Understand
aggregate risk
position
Slide 214
Slide 214 text
2. Compare to
appetite
Slide 215
Slide 215 text
2. Compare to
appetite
Probability of exceeding loss
0%
25%
50%
75%
100%
Loss exceeded
10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000
Current Appetite Residual
Slide 216
Slide 216 text
Probability of exceeding loss
0%
25%
50%
75%
100%
Loss exceeded
10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000
Current Appetite Residual
Slide 217
Slide 217 text
3. Invest in:
◦ Reducing uncertainty
◦ Mitigation
Slide 218
Slide 218 text
The biggest risk?
Slide 219
Slide 219 text
We only look for
answers in our field
Slide 220
Slide 220 text
Our niche becomes
obsolete
Slide 221
Slide 221 text
We used to laugh at
the box huggers
Slide 222
Slide 222 text
But building your own
PaaS in AWS?
Slide 223
Slide 223 text
The 2019 version of
being a box hugger
Slide 224
Slide 224 text
COTS PaaS meets
user needs of
95% of teams
Slide 225
Slide 225 text
Your niche is
going to disappear.
Slide 226
Slide 226 text
– Edward Deming
"It is not necessary to
change. Survival is not
mandatory."
Slide 227
Slide 227 text
Move up the stack
Slide 228
Slide 228 text
Understand the
business model
Slide 229
Slide 229 text
Learn skills to
navigate
uncertainty &
ambiguity
Slide 230
Slide 230 text
Kill your heroes
Slide 231
Slide 231 text
Thank you!
❤ the talk? Let @auxesis know.
Slide 232
Slide 232 text
•
Learn more:
◦ Algorithmic Impact Assessment report 2018
◦ Stella Report
◦ Wardley Mapping for busy people
◦ Wardley Maps book
◦ The Great Courses: Understanding complexity
Slide 233
Slide 233 text
•
Learn more:
◦ How to measure anything
◦ How to measure anything in cybersecurity
◦ Organisational culture and leadership
(Schein)
Slide 234
Slide 234 text
•
Learn more:
◦ Organization design: an information
processing view (Galbraith, 1974)
◦ Architectural Innovation: The
Reconfiguration of Existing Product
Technologies and the Failure of Established
Firms (Henderson and Clarke, 1990)