Cloud Native is about Culture, not Containers

A presentation at GOTO Berlin 2019 in October 2019 in Berlin, Germany by Holly Cummins

Slide 1

Slide 1

Cloud Native is About Culture, Not Containers (how to not fail at cloud native) Austin Copenhagen Dubai London Madrid Melbourne Munich Milan New York Holly Cummins IBM Garage @holly_cummins Nice Raleigh San Francisco São Paulo Singapore Tokyo Toronto

Slide 2

Slide 2

Slide 3

Slide 3

what is cloud native? #IBMGarage @holly_cummins

Slide 4

Slide 4

#IBM #IBMGarage @holly_cummins

Slide 5

Slide 5

Credit to @bibryam #IBM #IBMGarage @holly_cummins

Slide 6

Slide 6

(a great article, btw) #IBM #IBMGarage @holly_cummins

Slide 7

Slide 7

(a great article, btw) #IBM #IBMGarage @holly_cummins

Slide 8

Slide 8

#IBM #IBMGarage @holly_cummins

Slide 9

Slide 9

microservices containers dynamically orchestrated #IBM #IBMGarage @holly_cummins

Slide 10

Slide 10

; o t go #IBM e v i t a n d u o l c e h “t co mputing g n o r w s i n o i t a fo un d … .” e v i t a n d u o l abo ut c #IBMGarage @holly_cummins

Slide 11

Slide 11

; o t go #IBM e v i t a n d u o l c e h “t co mputing g n o r w s i n o i t a fo un d … .” e v i t a n d u o l abo ut c #IBMGarage Holly @holly_cummins

Slide 12

Slide 12

#IBM #IBMGarage @holly_cummins

Slide 13

Slide 13

build great products faster #IBM #IBMGarage @holly_cummins

Slide 14

Slide 14

why? #IBMGarage @holly_cummins

Slide 15

Slide 15

what problem are we trying to solve? #IBM #IBMGarage @holly_cummins

Slide 16

Slide 16

#IBM #IBMGarage @holly_cummins

Slide 17

Slide 17

CV-driven development. #IBM #IBMGarage @holly_cummins

Slide 18

Slide 18

“my CV looks dull” is not a good reason to go cloud native #IBM #IBMGarage @holly_cummins

Slide 19

Slide 19

why cloud? #IBM #IBMGarage @holly_cummins

Slide 20

Slide 20

cost #IBM #IBMGarage @holly_cummins

Slide 21

Slide 21

el a s t i c i t y #IBM #IBMGarage @holly_cummins

Slide 22

Slide 22

spee d #IBM #IBMGarage @holly_cummins

Slide 23

Slide 23

exotic capabilities #IBM #IBMGarage @holly_cummins

Slide 24

Slide 24

why cloud native? #IBM #IBMGarage @holly_cummins

Slide 25

Slide 25

#IBM #IBMGarage @holly_cummins

Slide 26

Slide 26

12 factors #IBM #IBMGarage @holly_cummins

Slide 27

Slide 27

12 factors #IBM how to write a cloud application so you don’t get electrocuted #IBMGarage @holly_cummins

Slide 28

Slide 28

#IBM #IBMGarage @holly_cummins

Slide 29

Slide 29

#IBM #IBMGarage @holly_cummins

Slide 30

Slide 30

cloud native is not a synonym for ‘microservices’ #IBM #IBMGarage @holly_cummins

Slide 31

Slide 31

#IBM #IBMGarage @holly_cummins

Slide 32

Slide 32

if ‘cloud native’ has to be a synonym for anything, it would be ‘idempotent’ #IBM #IBMGarage @holly_cummins

Slide 33

Slide 33

if ‘cloud native’ has to be a synonym for anything, it would be ‘idempotent’ which definitely needs a synonym #IBM #IBMGarage @holly_cummins

Slide 34

Slide 34

e s a b od o g a rs are e n i a t con #IBM #IBMGarage @holly_cummins

Slide 35

Slide 35

it’s not a competition to see how many you can have #IBM e s a b od o g a rs are e n i a t con #IBMGarage @holly_cummins

Slide 36

Slide 36

you do not need intra-app http communication to be cloud native #IBM #IBMGarage @holly_cummins

Slide 37

Slide 37

#IBM #IBMGarage @holly_cummins

Slide 38

Slide 38

complexity adds expense #IBM #IBMGarage @holly_cummins

Slide 39

Slide 39

unnecessary complexity adds unnecessary expense #IBM #IBMGarage @holly_cummins

Slide 40

Slide 40

#IBM #IBMGarage @holly_cummins

Slide 41

Slide 41

space pencil #IBM #IBMGarage @holly_cummins

Slide 42

Slide 42

space pencil $128.89 #IBM #IBMGarage @holly_cummins

Slide 43

Slide 43

space pencil $128.89 #IBM #IBMGarage space pen @holly_cummins

Slide 44

Slide 44

space pencil $128.89 #IBM #IBMGarage space pen $2.39 @holly_cummins

Slide 45

Slide 45

space pencil $128.89 space pen $2.39 +medical bills #IBM #IBMGarage @holly_cummins

Slide 46

Slide 46

accidental complexity #IBM essential complexity #IBMGarage @holly_cummins

Slide 47

Slide 47

accidental complexity essential complexity “this will impress my boss” complexity #IBM #IBMGarage @holly_cummins

Slide 48

Slide 48

2010 the dawn of cloud native #IBM #IBMGarage @holly_cummins

Slide 49

Slide 49

#IBM #IBMGarage @holly_cummins

Slide 50

Slide 50

behaves well on the cloud #IBM #IBMGarage @holly_cummins

Slide 51

Slide 51

behaves well on the cloud written for the cloud #IBM #IBMGarage @holly_cummins

Slide 52

Slide 52

this is all how we run our application, not what’s in it #IBM #IBMGarage @holly_cummins

Slide 53

Slide 53

speed #IBM #IBMGarage @holly_cummins

Slide 54

Slide 54

speed #IBM #IBMGarage @holly_cummins

Slide 55

Slide 55

what’s the point of getting the same old stuff to market faster? #IBM #IBMGarage @holly_cummins

Slide 56

Slide 56

what’s the point of being able to respond to the market, if you don’t? #IBM #IBMGarage @holly_cummins

Slide 57

Slide 57

what’s the point of architecture that can go faster, if you don’t go faster? #IBM #IBMGarage @holly_cummins

Slide 58

Slide 58

how to fail at cloud native #IBMGarage @holly_cummins

Slide 59

Slide 59

“each of our microservices has duplicated the same object model … with twenty classes and seventy fields” #IBM #IBMGarage @holly_cummins

Slide 60

Slide 60

Microservice #IBM #IBMGarage @holly_cummins

Slide 61

Slide 61

Microservice Domain #IBM #IBMGarage @holly_cummins

Slide 62

Slide 62

Microservice Domain #IBM #IBMGarage @holly_cummins

Slide 63

Slide 63

“every time we change code, something breaks” #IBM #IBMGarage @holly_cummins

Slide 64

Slide 64

distributed monolith #IBM #IBMGarage @holly_cummins

Slide 65

Slide 65

cloud-native spaghetti is still spaghetti (Image: Cloudy with a Chance of Meatballs.) #IBM #IBMGarage @holly_cummins

Slide 66

Slide 66

just because a system runs across 6 containers doesn’t mean it’s decoupled #IBM #IBMGarage @holly_cummins

Slide 67

Slide 67

distributed != decoupled #IBM #IBMGarage @holly_cummins

Slide 68

Slide 68

#IBM #IBMGarage @holly_cummins

Slide 69

Slide 69

Courtesy NASA/ JPL-Caltech #IBM #IBMGarage @holly_cummins

Slide 70

Slide 70

© 2019 IBM Corporation #IBMGarage @holly_cummins

Slide 71

Slide 71

© 2019 IBM Corporation #IBMGarage @holly_cummins

Slide 72

Slide 72

Slide 73

Slide 73

metric units

Slide 74

Slide 74

metric units imperial units

Slide 75

Slide 75

metric units imperial units distributing did not help

Slide 76

Slide 76

microservices need consumer-driven contract tests #IBM #IBMGarage @holly_cummins

Slide 77

Slide 77

“our tests aren’t automated” #IBM #IBMGarage @holly_cummins

Slide 78

Slide 78

“we don’t know if our code works” #IBM #IBMGarage @holly_cummins

Slide 79

Slide 79

systems will behave in unexpected ways #IBM #IBMGarage @holly_cummins

Slide 80

Slide 80

documentation can be wrong #IBM #IBMGarage @holly_cummins

Slide 81

Slide 81

dependency updates can change behaviour #IBM #IBMGarage @holly_cummins

Slide 82

Slide 82

#IBM #IBMGarage @holly_cummins

Slide 83

Slide 83

#IBM #IBMGarage @holly_cummins

Slide 84

Slide 84

“Had we done end-to-end testing, we believe this error would have been caught.” Arthur Stephenson Chief Investigator #IBM #IBMGarage @holly_cummins

Slide 85

Slide 85

“we can’t ship until we have more confidence in the quality” #IBM #IBMGarage @holly_cummins

Slide 86

Slide 86

microservices need automated integration tests #IBM #IBMGarage @holly_cummins

Slide 87

Slide 87

not a good CI/CD indicator a good CI/CD indicator “we don’t know when the build is broken” #IBM #IBMGarage @holly_cummins

Slide 88

Slide 88

a good build radiator #IBM #IBMGarage @holly_cummins

Slide 89

Slide 89

#IBM #IBMGarage @holly_cummins

Slide 90

Slide 90

“oh yes, that build has been broken for a few weeks…” #IBM #IBMGarage @holly_cummins

Slide 91

Slide 91

how to brick a spaceprobe #IBM #IBMGarage @holly_cummins

Slide 92

Slide 92

“we couldn’t get the automated checks to work, so we bypassed them” #IBM #IBMGarage @holly_cummins

Slide 93

Slide 93

“we’ve scheduled the architecture board review for a month after the project ships” #IBM #IBMGarage @holly_cummins

Slide 94

Slide 94

#IBM #IBMGarage @holly_cummins

Slide 95

Slide 95

navigators warned something was wrong #IBM #IBMGarage @holly_cummins

Slide 96

Slide 96

navigators warned something was wrong they didn’t fill in the right form #IBM #IBMGarage @holly_cummins

Slide 97

Slide 97

navigators warned something was wrong they didn’t fill in the right form so nothing was done #IBM #IBMGarage @holly_cummins

Slide 98

Slide 98

does the process add value? #IBM #IBMGarage @holly_cummins

Slide 99

Slide 99

extreme programming is the right kind of rigour #IBM #IBMGarage @holly_cummins

Slide 100

Slide 100

extreme programming is the right kind of rigour test-driven development #IBM #IBMGarage @holly_cummins

Slide 101

Slide 101

extreme programming is the right kind of rigour test-driven development pair programming #IBM #IBMGarage @holly_cummins

Slide 102

Slide 102

extreme programming is the right kind of rigour test-driven development pair programming optimise for feedback #IBM #IBMGarage @holly_cummins

Slide 103

Slide 103

“but it’s in the plan” #IBM #IBMGarage @holly_cummins

Slide 104

Slide 104

“but it’s not in the plan” #IBM #IBMGarage @holly_cummins

Slide 105

Slide 105

Slide 106

Slide 106

lots of bulkheads

Slide 107

Slide 107

it was too big #IBM #IBMGarage @holly_cummins

Slide 108

Slide 108

lookouts saw the iceberg but the ship wasn’t nimble enough to avoid it #IBM #IBMGarage @holly_cummins

Slide 109

Slide 109

NASA Langley 1961 #IBM #IBMGarage @holly_cummins

Slide 110

Slide 110

IBM 7090 (large!) #IBM #IBMGarage NASA Langley 1961 @holly_cummins

Slide 111

Slide 111

door (normal-sized) #IBM IBM 7090 (large!) #IBMGarage NASA Langley 1961 @holly_cummins

Slide 112

Slide 112

door (normal-sized) #IBM IBM 7090 (large!) #IBMGarage NASA Langley 1961 @holly_cummins

Slide 113

Slide 113

plans are always wrong success is in how you respond #IBM #IBMGarage @holly_cummins

Slide 114

Slide 114

“we can’t ship until every feature is complete” #IBM #IBMGarage @holly_cummins

Slide 115

Slide 115

how not to drive a car #IBM #IBMGarage @holly_cummins

Slide 116

Slide 116

they often couldn’t see the explorer #IBM #IBMGarage @holly_cummins

Slide 117

Slide 117

feedback is good business #IBM #IBMGarage @holly_cummins

Slide 118

Slide 118

feedback is good engineering #IBM #IBMGarage @holly_cummins

Slide 119

Slide 119

an mvp hurts #IBM #IBMGarage @holly_cummins

Slide 120

Slide 120

if you’re not embarrassed by your first release it was too late #IBM #IBMGarage @holly_cummins

Slide 121

Slide 121

experiments can mean failure #IBM #IBMGarage @holly_cummins

Slide 122

Slide 122

experiments can meanfailure failure experiments mean #IBM #IBMGarage @holly_cummins

Slide 123

Slide 123

The titanic staircase and lifeboats

Slide 124

Slide 124

users will have weird behaviours #IBM #IBMGarage @holly_cummins

Slide 125

Slide 125

optimise for recovery #IBM #IBMGarage @holly_cummins

Slide 126

Slide 126

unrecoverable #IBM #IBMGarage @holly_cummins

Slide 127

Slide 127

back in ms no data loss manual intervention bricked #IBM fast, but data lost handoffs #IBMGarage @holly_cummins

Slide 128

Slide 128

business recoverability #IBM #IBMGarage @holly_cummins

Slide 129

Slide 129

remember, users will have weird behaviours #IBM #IBMGarage @holly_cummins

Slide 130

Slide 130

speed #IBM #IBMGarage @holly_cummins

Slide 131

Slide 131

slow is demoralising for teams #IBM #IBMGarage @holly_cummins

Slide 132

Slide 132

fast is good business #IBM #IBMGarage @holly_cummins

Slide 133

Slide 133

A late change in requirements is a competitive advantage. –Mary Poppendieck #IBM #IBMGarage @holly_cummins

Slide 134

Slide 134

more feedback → more accuracy #IBM #IBMGarage @holly_cummins

Slide 135

Slide 135

cloud rescued developers from tedium and toil #IBM #IBMGarage @holly_cummins

Slide 136

Slide 136

cloud native should feel fun #IBM #IBMGarage @holly_cummins

Slide 137

Slide 137

100

Slide 138

Slide 138

® @holly_cummins https://www.ibm.com/legal/us/en/copytrade.shtml 101