For many years, software architecture was described as the “parts that are hard to change later.” Modern advances in architecture have shown that if architects build evolvability into the architecture, change becomes easier. Neal Ford describes a family of software architectures that support evolutionary change, how we can build evolvable systems, and how to retrofit existing ones. He discusses three key concepts that support evolutionary architectures. Incremental change covers engineering practices to support continuous delivery and DevOps. Neal explains how fitness functions build protective, testable scaffolding around critical parts to guide the architecture as it evolves and to help developers assess how well the architecture adheres to a particular goal. Third, he describes how to determine the level of architectural coupling, which affects the ability to evolve the architecture. Neal illustrates ways to retrofit existing systems to support architectural evolution and shares tips on migrating between architectural styles. Leave with a new appreciation of software architecture and practical ways to make your systems not nearly so hard to change.

1. K2
Keynote
6/7/2017 10:00:00 AM
K2 Modern Evolutionary Software
Architectures
Presented by:
Neal Ford
ThoughtWorks
Brought to you by:
350 Corporate Way, Suite 400, Orange Park, FL 32073
888-­‐268-­‐8770 ·∙ 904-­‐278-­‐0524 - info@techwell.com - https://www.techwell.com/

2. Neal Ford
ThoughtWorks
Neal Ford is director, software architect, and meme wrangler at ThoughtWorks, a
software company and a community of passionate, purpose-led individuals. He
thinks disruptively to deliver technology to address the toughest challenges, all
while seeking to revolutionize the IT industry and create positive social change. A
speaker at hundreds of developer conferences worldwide, Neal is an
internationally recognized expert on software development and delivery,
particularly where agile engineering techniques and software architecture
intersect. Neal has authored magazine articles, seven books (and counting), and
dozens of video presentations. Neal's topics include software architecture,
continuous delivery, functional programming, and cutting-edge software
innovations. Check out Neal's website at nealford.com.

3. Building Evolutionary
Architectures
@neal4d
nealford.com
with Rebecca Parsons & Pat Kua

4. Rebecca Parsons
Pat Kua
Neal Ford
Photos by Martin Fowler:
http://martinfowler.com/albums/ThoughtWorkers/

5. What is Software
Architecture?

8. Entire architectural Scope
requirements

9. Entire architectural Scope
time
performance
legality
scalability
security
data
auditability
requirements

10. sam
pling
accessibility reliability repeatability
accountability extensibility reproducibility
accuracy failure transparency resilience
adaptability fault-tolerance responsiveness
administrability fidelity reusability
affordability flexibility robustness
agility inspectability safety
auditability installability scalability
autonomy integrity seamlessness
availability interchangeability self-sustainability
compatibility interoperability serviceability
composability learnability supportability
configurability maintainability securability
correctness manageability simplicity
credibility mobility stability
customizability modifiability standards compliance
debugability modularity survivability
degradability operability sustainability
determinability orthogonality tailorability
demonstrability portability testability
dependability precision timeliness
deployability predictability traceability
discoverability process capabilities transparency
distributability producibility ubiquity
durability provability understandability
effectiveness recoverability upgradability
efficiency relevance usability
https://en.wikipedia.org/wiki/List_of_system_quality_attributes
ility

11. evolvability
accessibility reliability repeatability
accountability extensibility reproducibility
accuracy failure transparency resilience
adaptability fault-tolerance responsiveness
administrability fidelity reusability
affordability flexibility robustness
agility inspectability safety
auditability installability scalability
autonomy integrity seamlessness
availability interchangeability self-sustainability
compatibility interoperability serviceability
composability learnability supportability
configurability maintainability securability
correctness manageability simplicity
credibility mobility stability
customizability modifiability standards compliance
debugability modularity survivability
degradability operability sustainability
determinability orthogonality tailorability
demonstrability portability testability
dependability precision timeliness
deployability predictability traceability
discoverability process capabilities transparency
distributability producibility ubiquity
durability provability understandability
effectiveness recoverability upgradability
efficiency relevance usability
evolvability

12. Once I’ve built an
architecture, how can I
prevent it from gradually
degrading over time?

13. How is long term planning
possible when things
change unexpectedly?

14. Dynamic Equilibrium

15. Architecture is the decisions
that you wish you could get right early
in a project.
— Ralph Johnson
things that people perceive
as hard to change.

16. What if we build
architectures that expect
change?

17. Definition:
An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.

18. Perspectives on Architecture
Technical Architecture

19. Evolvability of Architectures

20. Big Ball of Mud
classes
coupling connections 0dimensions :

21. Layered Architecture
1dimensions :
opportunities: 4

22. Microkernel
1dimensions :

23. REST
1dimensions :

24. Domain Perspective
0dimensions :

25. Microservices
ndimensions :evolutionary architecture

26. An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.
evolutionary architecture
Definition:

27. Composability

28. Composability

29. Composability

30. Definition:
An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.
evolutionary architecture

31. Incremental Change
production
Components are deployed.
Features are released.
Applications consist of routing.

32. Incremental Change
production

33. Definition:
An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.
evolutionary architecture

35. Perspectives on Architecture
Data Architecture

36. Perspectives on Architecture
Security Architecture

37. Perspectives on Architecture
Domain Architecture

38. Microservices
Domain Architecture
+

39. Fitness Functions
⩊
a particular type of objective function that is
used to summarize…how close a given
design solution is to achieving the set aims.

40. Architecture Fitness Functions
⩊
metrics
tests

41. Definition:
An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.
evolutionary architecture

42. Agenda
definition
incremental change
fitness functions
appropriate coupling

43. Fitness Function
a particular type of objective
function that is used to
summarize…how close a given
design solution is to achieving
the set aims.

44. Architecture Fitness Functions
metrics
tests

45. Architecture Fitness Functions
metrics tests

46. Fitness Function
atomic
holistic
batch continuous

47. Fitness Function
atomic
holistic
batch continuous

48. Fitness Function
atomic
holistic
batch
continuous

49. Cyclic Dependency Function
clarkware.com/software/JDepend.html
application

50. Coupling Fitness Function
application

51. Fitness Function
atomic
holistic
batch
continuous

52. Fitness Function
atomic
holistic
batch continuous

53. Fitness Function
atomic
holistic
batch
continuous

54. holistic
batch
Holistic fitness functions must run in a specific (shared) context.

55. Consumer Driven Contracts
martinfowler.com/articles/consumerDrivenContracts.html

56. Fitness Function
atomic
holistic
batch
continuous

57. Fitness Function
atomic
holistic
batch continuous

58. Fitness Function
atomic
holistic
batch
continuous

59. atomic
continuous
monitoring logging

60. Fitness Function
atomic
holistic
batch
continuous

61. Fitness Function
atomic
holistic
batch continuous

62. Fitness Function
atomic
holisticbatch
continuous

63. holistic
resilience-ility?

64. Fitness Function
atomic
holisticbatch
continuous

65. Fitness Function
atomic
holistic
batch continuous

66. System-wide Fitness Function

67. Fitness Function Fit

68. Guided Evolution

69. Agenda
definition
incremental change
fitness functions
appropriate coupling

70. Prerequisites

71. Deployment Pipeline

72. Deployment Pipeline

73. Incremental Change
Engine of evolutionary architecture
v ∝ c
where
c = cycle time
v = maximum speed of new generations

74. Agenda
definition
incremental change
fitness functions
appropriate coupling

75. Customer
Code Reuse (Over Time)
checkout shippingSOA

76. checkout shippingmicroservices
Code Reuse (Over Time)

77. The more reusable code is,
the less usable it is.

78. Decentralized Data Management

79. Decentralized Data Management
Transactions are
temporal coupling.

80. Decentralized Data Management
Limit
transactional
contexts.

81. Evolutionary Database Design
http://databaserefactoring.com/

82. Evolving Columns

83. Transition

84. Expand/Contract Pattern
app v230
compatible with db
v14
DB
version
15
DB
version
14
app v205
compatible with db
v13 and v14
app v234
compatible with db
v14
app v241
compatible with db
v14 and 15
DB
version
13
app v248
compatible with db
v15
Time
app v205
deployed
migrate
db to v14
app v230
deployed
app v234
deployed
app v234
deployed
app v248
deployed
migrate
db to v15

85. Decentralized Governance

86. Decentralized Governance

87. Decentralized Governance

88. “Goldilocks” Governance
Choose technology stacks
appropriate to problem scale.

89. Shift to Domain-centric Architectures

90. Shift to Domain-centric Architectures

91. Shift to Domain-centric Architectures

92. Incidentally Coupled Teams
user interface
server-side
DBA

93. Conway’s Law
organizations which design
systems ... are constrained to
produce designs which are copies
of the communication structures
of these organizations
en.wikipedia.org/wiki/Conway%27s_law
Melvin Conway, 1968

94. Incidentally Coupled Teams
user interface
server-side
DBA

95. Autonomous Teams
Orders Shipping
Inverse Conway Maneuver
Catalog

96. Low Efferent Coupling between Teams
n(n-1)
2

97. Architectural Quantum

98. Architectural Quantum

99. Architectural Quantum

100. Architectural Quantum

101. Utilizing
Evolutionary Architecture

102. 1. Choose Dimensions

103. Entire architectural Scope
time
performance
legality
scalability
security
data
auditability
requirements

104. Entire architectural Scope
fitness functions
time
performance
legality
scalability
security
data
auditability
requirements

105. Entire architectural Scope
fitness functions
time
performance
legality
scalability
security
data
auditability
requirements

106. Utilizing
Evolutionary Architecture

107. 2. Identify Fitness Functions
atomic
holistic
/
automated / manual

108. Utilizing
Evolutionary Architecture

109. 3. Apply Incremental Change

110. Utilizing
Evolutionary Architecture

111. Utilizing
Evolutionary Architecture

112. Agenda
definition
incremental change
fitness functions
appropriate coupling

113. Why should a company
decide to build an
evolutionary architecture?

114. Why should a company
decide to build an
evolutionary architecture?
Predictable versus Evolvable
Scale
Cycle Time as a Business Metric
Isolating “—ilities” at the Quantum Level
Longer Lasting Useful Systems
Advanced Business Capabilities

115. Why should a company
decide to build an
evolutionary architecture?

116. Why would a company choose not
to build an evolutionary
architecture?

117. Why would a company choose not
to build an evolutionary
architecture?
Can’t Evolve a Ball of Mud
Other Architectural Characteristics
Dominate
Sacrificial Architecture
Planning on Closing the Business Soon

118. Why should a company
decide to build an
evolutionary architecture?
Predictable versus Evolvable
Scale
Cycle Time as a Business Metric
Isolating “—ilities” at the Quantum Level
Longer Lasting Useful Systems
Advanced Business CapabilitiesAdvanced Business Capabilities

119. Hypothesis and Data
Driven Development

120. Move
Fast
&
Fix
Things

121. https://github.com/github/scientist

122. ▫︎It decides whether or not to run the try block,
▫︎Randomizes the order in which use and try blocks are
run,
▫︎Measures the durations of all behaviors,
▫︎Compares the result of try to the result of use,
▫︎Swallows (but records) any exceptions raised in the try
block
▫︎Publishes all this information.

126. Bugs Found; Resolution
▫︎faster conflict return because shell script exited
immediately; replicated in library
▫︎index write was causing O(n) problem; inlined into
memory
▫︎the ancestor had a file with a given filemode, whilst one
side of the merge had removed the file and the other
side had changed the filemode; bug in git!
▫︎Git incorrectly successfully merged files w/ 768 conflicts;
fixed git shell script
▫︎new library was skipping an entire step; bug found &
fixed

127. Definition:
An evolutionary architecture supports
incremental, guided change as a first
principle across multiple dimensions.
evolutionary architecture

128. http://evolutionaryarchitecture.com

129. nealford.com
@neal4d
nealford.com/books
nealford.com/videos
www.oreilly.com/software-architecture-video-training-series.html

130. Agenda
definition
incremental change
fitness functions
appropriate coupling