Infrastructure as Code: Fueling the Fire for Faster Application Delivery- Forrester, 2015

Collaborate more
PEOPLE Share common goals
Focus on improvement
BRINGING PEOPLE TOGETHER

Eliminate waste
PROCESS Increase efficiency
Streamline feedback
DELIVERING VALUE FASTER

Enhance productivity
TOOLS Enable collaboration
Facilitate experimentation
EXECUTING A DEVOPS STRATEGY
DevOps Benefits for IT Pros
Spend 29 percent more
time on new work
Deploy code 200x
more frequent

https://puppet.com/resources/white-paper/2016-state-of-devops-report
 Shorten cycle times Improve quality Optimize resources Deliver apps with digital-
and deliver value faster and availability and eliminate waste era velocity
Delivering value to Business the BizDevOps way

Plan + Requirements Monitor + learn

Track
(Continuous Improvement)

Development Collaboration Production

Develop Release
+ test
 Focus on planning Focus on delivering
Compete, not collaborate Collaborate to win
Static hierarchies Fluid and flexible teams
Individual productivity Collective value creation
Efficiency of process Effectiveness of outcomes
Assumptions, not data Experiment, learn and respond
“DevOps is “DevOps is
development treating your
and operations infrastructure
collaboration” as code”

“DevOps “DevOps
is using is feature
automation” switches”

“DevOps “Kanban
is small for Ops?”
deployments”
 - Donovan Brown
DevOps Senior Program Manager at Microsoft
One team responsible And responsible for the
for the complete stack complete lifecycle
PLAN RELEASE
Data
Internal or external data
incl. security and privacy

Application
Custom Application or
COTS Application

D E V E LO P + T E S T M O N I TO R + L E A R N
Infrastructure
From the marketplace or
a deployment template
managed by central IT Completely Automated
Automated Release Pipeline (app & configuration code)
Based on Microsoft’s experiences with our
own transformation to a DevOps culture,
we have identified DevOps principles and
related DevOps practices that will lead to a
true shift to DevOps in organizations of any
size.
 PRACTICES:
Application Performance Management

PRACTICES:
Infrastructure as Code
Automated Testing Continuous Deployment
PRACTICES:
Continuous Integration Release Management
Usage Monitoring PRACTICES:
Continuous Deployment Configuration Management
Telemetry Collection Code Reviews
Release Management Automated Recovery
Testing in Production Automated Testing
Stakeholder Feedback Continuous Measurement

INFRA-
F LOW O F T E A M A U TO N O M Y B AC K LO G EVIDENCE MANAGING PRODUCTION
STRUCTURE as
C U S TO M E R and ENTERPRISE groomed with gathered in TECHNICAL FIRST
a FLEXIBLE
VA L U E ALIGNMENT LEARNING PRODUCTION D E BT MINDSET
RESOURCE

PRACTICES:
PRACTICES:
Application Performance Management
PRACTICES: Testing in Production Infrastructure as Code
Automated Testing Usage Monitoring Continuous Delivery
Continuous Integration User Telemetry Release Management
Continuous Deployment Stakeholder feedback Configuration Management
Release Management Feature flags Automated Recovery
DevOps Lifecycle
1 Plan 4 Monitor + Learn

Development Production

2 Develop + Test 3 Release

Plan

PROJECT STARTS

MANAGE WORK

DEVELOP+TEST

TRACK PROGRESS PLAN

 The Scrum Team

Source: Scrum Alliance

Develop + Test
Once the iteration starts, developers
turn great ideas into features…

WRITE CODE

UNIT
TESTING

VERSION
CONTROL

BUILD

BUILD
VERIFICATION

RELEASE
 MASTER BRANCH

CREATE A ADD OPEN A PULL DISCUSS AND

DEPLOY MERGE
BRANCH COMMITS REQUEST REVIEW CODE
 Value

cspkg

Measure
Release

Cloud
Load Testing

IN TEGRATION TES TIN G S TAGIN G

EN VIRON M EN T EN VIRON M EN T

MONITOR+LEARN

AUT OMAT E D FUN CTION A L PRE- PRODUC TI ON

T ES TIN G EN VIRON M EN T EN VIRON M EN T
 Value
cspkg

Measure
Infrastructure as Code

Value

Measure
Release Management

Value

Measure
Load Testing and Auto Scale

Value

Measure
Monitor + Learn
PLAN THE N EXT ITERATION

PLAN

FEEDBA CK

MON ITOR
Hypothesis Driven Development
Application Performance Monitoring
Value

Measure
Automated Recovery
Rollback
Value

Measure
Value

Measure
Performance Indicators High IT Performers Medium IT Performers Low IT Performers

On demand (multiple Between once per week Between once per month
Deployment frequency
deploys per day) and once per month and once every 6 months

Between one week and Between one month and

Lead time for changes Less than one hour
one month 6 months

Mean time to recover (MTTR) Less than one hour Less than one day Less than one day*

Change failure rate 0-15% 31-45% 16-30%

 "…Of course, computing needs will go far beyond
the data center. Essentially wherever is data,
compute will migrate to data and so we are going
and taking Azure to Azure Stack to Azure IoT Edge,
to Azure Sphere. This is that one ubiquitous
distributed computing fabric. One programming
model that is event driven, server less so you can
write an application that truly works across all of
this…"

Satya Nadella, Microsoft CEO, July 2018

39
 Hybrid DevOps

Use cloud native DevOps strategies to drive innovation through hybrid applications

Portable cloud native Use the same tools Extend Azure for
applications and across cloud boundaries specialized needs
infrastructure
Hybrid DevOps

1. Engineer makes changes to

application code and ARM
template.
2. Code and ARM template are
checked into Visual Studio
Team Services Git.
3. Continuous integration
triggers application build and
unit tests.
4. Continuous deployment
trigger orchestrates
deployment of application
artifacts with environment-
specific parameters.
Hybrid DevOps
 Cross Cloud Scaling

Scale your application components and infrastructure seamlessly across

cloud boundaries with custom compliance needs

Scale application to Increased Flexibility and Consistent experiences

meet your customer Resiliency with Azure across cloud boundaries
demands easily Stack consistency
 Cross Cloud Scaling

Azure

1. Large numbers of users

attempt to access the web
app.
5 2. Traffic manager returns the
Azure Stack DNS name.
3. Users use the Azure Stack
web app.
4. Once a threshold is reached,
1 4 a Function starts the Azure
2 Web App and enabled the
Azure Traffic Manager Route.
5. Traffic is routed to Azure,
3 which can automatically
scale App Service.
Azure Stack
Cross Cloud Scaling
 Data Sovereignty and Data Gravity

Keep your cloud-native data and applications available while

meeting custom compliance requirements

Bring applications Create robust Meet regulatory, data

and services to where applications that span privacy, and compliance
data resides across cloud boundaries requirements
Data Sovereignty and Data Gravity

Azure Azure Stack

1. User enters data into Azure-based
web app.
2. Application commits data to database
over VNet to VNet VPN connection to
Azure Stack.
1 2 3. Data is stored in SQL database on VM.

3
 AI at the edge

Build AI enabled hybrid applications that span cloud boundaries

Bring intelligence and Train models at scale Get faster insights

computing power to – deploy on-prem with inference @ the
where your data edge
originates
 AI at the edge

1. Data scientists train a model using

Azure ML workbench and an
HDInsight cluster. The model is
Azure Azure Stack
containerized and put in to an Azure
Container Registry.
2. The model is deployed to a
Kubernetes cluster on Azure Stack.
3. End users provide data that is scored
2 3 against the model.
4. Insights and anomalies from scoring
1 7 are placed into a queue.
5. A Function sends compliant data and
5 4 anomalies to Azure Storage
6. Globally-relevant and compliant
6 insights are available in the global
app.
7. Data from edge scoring is used to
improve the model.
AI at the edge
 Geo-distributed applications

Route customers to the application that meets their needs

Deploy modern Provide compliant Enable seamless

serverless applications services for users connections with Azure
and microservices worldwide. Services
globally
 Geo-distributed applications

Azure

3 1. Users attempt to access

application
Azure Stack (EU) 2. DNS queries Traffic Manager.
3. Traffic Manager returns the
1 application (Azure or Azure
Stack) that’s closest to the
2 user.

Azure Stack (Africa)

Azure Stack
Geo-distributed applications