5. NEW WORLD IT Employees at work
Factories +
supply chain
IoT connected
things
Online
marketing
Continuous
supply tracking
Just-in-time
production
Online sales
+ delivery
Social media
18. Microservices
Five Years Agoto Functions
Amazon
Kinesis
Amazon API
Gateway
Amazon SNS
Amazon S3
Amazon
DynamoDB
Amazon
SQS
Standard building brick
services provide standardized
platform capabilities
19. Amazon SNS
Amazon S3
Amazon API
Gateway
Amazon
SQS
Amazon
Kinesis
Amazon
DynamoDB
Microservices
to Functions
Business Logic
Glue between
the bricks
Standard building brick
services provide standardized
platform capabilities
20. Amazon SNS
Amazon S3
Amazon API
Gateway
Amazon
SQS
Amazon
Kinesis
Amazon
DynamoDB
Microservices
to Functions
21. Amazon SNS
Amazon S3
Amazon API
Gateway
Amazon
SQS
Amazon
Kinesis
Amazon
DynamoDB
Microservices
to Functions
22. Amazon SNS
Amazon S3
Amazon API
Gateway
Amazon
SQS
Amazon
Kinesis
Amazon
DynamoDB
Microservices
to FunctionsEphemeral
27. Amazon SNS
Amazon S3
Amazon API
Gateway
Amazon
SQS
Amazon
Kinesis
Amazon
DynamoDB
Microservices
to
Functions
Ephemeral
When the system is
idle, it shuts down and
costs nothing to run
55. Cloud Migration
Pay as you go
DATACENTER
Applications and data
Pay up front and
depreciate over
3 years
Pay a month later
for the number of
seconds used
57. File tickets and wait
for every step
Self-service,
on-demand, no delays
!
VS
!
58. File tickets and wait
for every step
Self-service,
on-demand, no delays
!
VS
!!
59. File tickets and
wait for every step
Self service,
on-demand, no delays
!
VS
!!
Deploy by filing a
ticket and waiting
weeks or months
Deploy by making an
API call self-service
within minutes
60. Cloud Native Principle
Self-service, API-driven, automated
Move from request tickets at every
step to a tracking ticket that records
what happened
70. Elasticity
DATACENTER
Hard to get over 10 percent
utilization—need extra capacity in
case of peak
CLOUD
Target over 40 percent
utilization—no capacity overload
issues
71. Automatic scaling for predictable heavy workloads
Serverless for spiky workloads with idle periods
72. Cloud Native Principle
Turn it off when it’s idle
Many times higher utilization
Huge cost savings
Avoids capacity overloads
102. Years
Millions of dollars
Hundreds of dev years
Lock-in
Lawyers and contracts
Old World
Monolithic on-prem waterfall lock-in
Weeks
Hundreds of dollars
A few dev weeks
Refactoring
Self-service
New World
Agile cloud-native micro-dependencies
103. Bottom Line
ROI for choosing, using, losing has
changed radically. Stop talking about
lock-in, it’s just refactoring dependencies
The cost of each dependency is far lower
Frequency of refactoring is far higher
Investment and return are much more incremental
119. Who has a backup datacenter?
What’s the best description of it?
1. Availability theater—never tried to use it
2. Infrequent partial testing
3. Regular tests during maintenance
4. Frequent failovers during production to
prove that no-one can tell it’s happening
120. Route updates and customer requests
to specific regions and services
Replicate data and re-route requests
during incidents
Switching mechanism must be far
more reliable than redundant
elements you are switching between
124. Microservices limit ‘blast radius’
for software incidents
Circuit breakers limit damage
Bulkheads prevent it from spreading
DITTO—Do Idempotent Things To Others
Avoid update and delete semantics
127. I wonder why it
did that?
Let’s reboot it.
Whoops!
Now it’s
really hosedUnexpected application
behavior often causes
people to intervene and
make the situation worse
People
128. A fire drill is a boring
routine where we make
everyone take the
stairs and assemble in
the parking lot
People
Training
!
129. Fire drills save lives in
the event of a real fire,
because people are
trained for how to react
People
Training