Dev Ops
Dev Ops
Prepared by:
OUTLINE
1. DELOITTE’S VIEW ON DEVOPS
2. HOW DELOITTE DOES DEVOPS
3. CASE STUDY
4. APPENDIX
DELOITTE’S VIEW ON DEVOPS
DEVOPS IS A TRANSFORMATIONAL APPROACH TO
MODERNIZE SOFTWARE DELIVERY
Development wants rapid change
Operations wants stability and quality
DevOps helps achieve both objectives through creating and transforming shared culture, processes, and technology
Customer Feedback
Regulatory requirements
How fast can you adapt?
How often do your changes cause unanticipated failures?
Industry innovations
How disruptive are your changes to customer experience?
Do your systems tell you anything?
Are alerts actionable? How soon after reporting/ noticing a problem will the
customer see the resolution?
How resilient are your systems?
How fast can you resolve?
How fast can you spot a failure?
© Deloitte LLP and affiliated entities. 6
DEVOPS SUPPORTS TRANSITION INTO RIGHT SPEED IT
Right Speed IT delivers the ability to provide the
whole spectrum of IT services at variable pace,
with appropriate governance and processes to
support. There is seamless interaction between
the Business and all aspects of IT
Maturity
Time
Change cycles & process
Right Speed IT allows the IT organization to introduce the differentiation in change cycles (shorter more
iterative), investment planning and governance and processes that are required to deliver against a
digital strategy and embed the ability to realize innovation within the organization. This turns IT into an
enabler, rather than simply a cost center. Getting the right governance and process in place is key
© Deloitte LLP and affiliated entities.
however.
7
APPLYING DEVOPS PRINCIPLES TRANSFORMS THE
SOFTWARE DEVELOPMENT LIFECYCLE
The Dev:
Runs IT as a business
The Ops:
Evolves IT to enhance the business
Empowerment
Modern Architecture
© Deloitte LLP and affiliated entities. 8
HIGH PERFORMANCE TECHNOLOGY TEAMS
2017 STATE OF DEVOPS REPORT – SPONSORED BY DELOITTE
How often are features delivered to Production? High performing teams deploy How long does it take a feature to go from code to
On Demand. production? The very best take a little over 1 hour.
Personal
Has a clear concept of where the organization is going
Recognition
and where it should be in the next five years. Deployment Lead Time for
Demonstrated care and
Frequency Changes consideration of followers’
Supportive personal needs and feelings.
Vision
Leadership
How long does it take to restore service when an incident Mean Time to Change Failure
occurs? High performing teams can restore service in just Recover Rate
over 1 hour. What percentage of changes result in
downgraded performance? This happens to top
performing teams 0-15%.
9
DEVOPS TRENDS – THE NEAR FUTURE
TODAY’S INNOVATIONS ARE TOMORROW’S NORMS
Automating the Automation Drift Detection
In an ideal world, all server in all environments are
By utilizing configuration management tools and infrastructure as code
identical. This simplifies maintenance and makes
entire DevOps systems can be replicated and version controlled
applications independent of any environment properties.
Drift Detection allows for tracking the deviation of
Serverless Computing infrastructure across and within environments.
The concept of a server is no longer present to the user. When an
execution is required, the instructions are sent to a web service Text
provider who takes care of all the details needed to maintain a
stable and secure infrastructure. Ephemeral Pipelines
Pipelines on demand! Utilizing dynamic infrastructure
pipelines are created as required, either from the need to
build or deploy. Once the task is complete the pipeline self
DevOps + Big Data Text Text destructs. This not only reduces infrastructure usage costs
Application logs, console outputs, and test reports are a few but also maintenance cost as well.
examples of the data generated from DevOps pipeline. Utilizing
Text
techniques from BIG DATA analysis a pipeline can synthesis and
analysis this information. These intelligent pipelines are capable of Hybrid Cloud
self diagnosis and triaging errors. Have on premise servers but want the convenient interface of a cloud provider?
Services such Platform9 can modernize how you interact with your servers
© Deloitte LLP and affiliated entities. creating cloud like benefits with your on premise infrastructure.
10
HOW DELOITTE DOES DEVOPS
FULL DEVOPS POTENTIAL IS REALIZED ONLY WHEN
TRANSFORMATION IS HOLISTIC
Deloitte brings the highly connected DevOps, Cloud, Digital, and Agile offerings together to offer
a comprehensive business transformation
Deployment Dependency
Configuration
New Code Change Continuous
Infra
Management
Integration
Application Monitoring
Auto Env
Performance App Infra
Test Management Auto
Management Lifecycle Lifecycle
Automation Scaling
Continuous Test
Dependency Data Automated
Continuous
Delivery New Infra Change
Lifecycle Procurement
Deployment
(e.g. patch)
On-Call Support Management – ChatOps – Business Performance Management – Pipeline Monitoring – Self-Healing Systems
© Deloitte LLP and affiliated entities. 14
Advanced Capabilities
BUILDING DEVOPS CAPABILITIES
Intermediate Capabilities
INCREMENTALLY IS KEY
Continuous Deployment
Continuous Delivery Core Capabilities
Infrastructure as a Service
Platform as a Service
Application Containerization Source Code Management
Container Orchestration
Test data Management Automated Build Management
Automated Env Management
Centralized Log Mgmt Build Artifact and Retention Management
Performance Profiling Automated Unit Testing
Configuration Management Automated Functional Testing
ChatOps for Incident Mgmt
Infrastructure Monitoring Continuous Integration
Self-Healing Environments
APM (Performance Mgmt) Consistent/Standardized Environments
Log Insight/Analytics
Drift Detection Static Code Analysis
Ops Dashboarding
Service Virtualization
Capability Characteristics
Versioned and Built from Sources, Self-Served to End Users, Integrated Toolchain, Transparent Pipeline, Treated as a Platform
© Deloitte LLP and affiliated entities.
DEVOPS TRANSFORMATION – STARTING ON THE RIGHT FOOT
DEVOPS IS NOT A ONE SIZE FITS ALL
What to transform? How to transform?
“Any improvement
not made at the
constraint is an Ad Ad
op
Pilot op Pilot t Pilot
illusion” t
DELOITTE
TRANSFORMATION OFFERING IMPLEMENTATION & EXECUTION
§ Operating Model Development DEVOPS DEVOPS § System Integration and AMS via DevOps
§ Process Design, Tooling Selection, and TRANSFORMATION IMPLEMENTATION Techniques
Pipeline Build (Version control, CI/CD, etc.) & EXECUTION § Environment management as a services (EAAS)
§ Sourcing and Service model (vendor § Tooling Configuration & Implementation
selection, etc.)
§ Organizational Change Management
© Deloitte LLP and affiliated entities. 17
CASE STUDY
A RECENT DELOITTE DEVOPS IMPLEMENTATION ON
FULL APPLICATION STACK
Deloitte established an end-to-end CI/CD pipeline based on DevOps practices for a major P&C insurer in a
green field environment
Pipeline Overview
• The program was delivered using an agile approach with multiple scrum teams operating in a synchronized 4-week sprint cycle
• The DevOps pipeline is currently being used to support ~ 15 teams with around 100 on-demand, integrated dev/test
environments to enable high velocity for the Development and the QA team (~150 deployments daily into dev/ test
environments)
• Development and QA teams (with team members spread across 4 continents) can initiate their own integrated environment
builds across the technology “stack” i.e., infrastructure to configuration
• Successful builds of a dev/test environment are completed ~ 30 mins with a high success rate
• Pipeline is being maintained and supported by a small team (approx. 5) for the entire program
• Automated testing includes staged sanity and a nightly regression suite of over 1000 test cases
• Small production releases take place on weekly basis with bigger releases going in monthly
19
LARGE P&C INSURANCE CLIENT– DEVOPS PIPELINE
Developer Manual Trigger
Check-out Engineer/ Tester
Check-in Code Version Code Continuous
Control Integration
System Server Time Trigger/ Event
Auto-trigger Clo
ud
Pro
vid
er
Perform SDLC Function Provision Environment
Ser
vice
Virt
ual
CI/ CD Pipeline izat
ion
Ove
rlay
Ne
two
rk
Build latest dev Run unit tests Deploy latest dev Run sanity Promote build Run Promote build Run regression Promote to Promote to
commit build into OD env tests in OD into next env functional into next env tests performance production
env tests env then PPS environment
Supporting Tools
Artifact Management Configuration Management Service Discovery Secrets Management Centralized Log Management Performance Monitoring
LARGE P&C INSURANCE CLIENT – ENV ORCHESTRATION
Large OD (Future) – Prod Region
Composed of Pre-Prod and Production
…
21
LARGE P&C INSURANCE CLIENT – SELF-SERVE
CAPABILITIES
Self-serve Builds and Deployments Allow Development to be faster Dashboards Enable Transparency and reduce knowledge dependency on
and Operations to focus on improvements SMEs
22
Get in touch. Jody McDermott
Bay Adelaide Centre, East Tower
Fawzy Manaa
Bay Adelaide Centre, East Tower
Abeer Rahman
Bay Adelaide Centre, East Tower
8 Adelaide Street West, Suite 200 8 Adelaide Street West, Suite 200 8 Adelaide Street West, Suite 200
Look forward to hearing from you on next steps Toronto, ON, M5H 0A9 Toronto, ON, M5H 0A9 Toronto, ON, M5H 0A9
+1 647 464 8222 +1 416 525 1813 +1 416 526 7245
[email protected] [email protected] [email protected]
APPENDIX
WHAT A MODERN DEVOPS PIPELINE LOOKS LIKE
Continuous Deployment Continuous Verification
Write Test Framework
Dynamically Provision
Tests
… Infrastructure
Invoke
Dev/QA Test
Monitoring PROD DEV Schedule On/
Config. Scripts Suites
Report From
Configuration
Dev & Ops
Deploy CI Server
CI Server Env X
Promote
DevOps
Pipeline
Continuous Integration Continuous Development
Team Collab.
Tool
Commit Build View 4) Collaborate
25
BREAKING CORE DEVOPS PRACTICES DOWN
DevSecOps
Infrastructure Lifecycle
Data Lifecycle 26
DEVOPS TOOLING LANDSCAPE
https://xebialabs.com/periodic-table-of-devops-tools/
Technical Capability Best of Breed Work Planning
Code Management
• Work/defect tracking – Centralize tracking and reporting of tasks and work assignments
• Source code control – Record changes to files and allow these changes to be managed using
automated or manual triggers
• Code scan – Automate the analysis of source code, identifying code quality challenges, and
non-compliance to coding standards
Continuous Integration
• Build dependency management – Allows for specification and management of direct build process
dependencies
• Automated build – Automate the creation of an application deployable artifact and set of activities in a
single package which can be deployed and made available to an environment
• Application configuration management – Centralized management of configuration, injecting them into
the appropriate artifacts during build or deployment
Quality Management
• Automated unit testing – Automate the execution and tracking of tests which validate the individual
classes or components that comprise the application
• Automated non-functional testing – Automate validation tests to ensure that application adheres to
requirements around the application’s operation (e.g., performance)
• Automated functional testing – Automate validation tests to ensure the application adheres to business
features, defined by requirements/acceptance criteria
Dynamic Infrastructure
• Environment configuration management – Centralize management of environment injecting
configuration during application build or deployment
• Infrastructure automation – Automate provisioning and management of infrastructure assets built into
the runtime environment