devops has been popular in IT ever since emerging cloud technology. to make IT more agile, we need to keep setup goal and measure performance with adopting new cloud native tools.
6. The History of IT Organizational Systems
• 1960-1980s: Fragile “Cowboys”
– Optimize around the hardware
• 1990-2000s: Robust “ITIL”
– Optimize and protect around changes or traumas
• 2010s+: Resilient/Anti-Fragile “DevOps”
– Optimize on continuously improving the flow of value
Adapted from: Sergio Graziosi, Social Science Research Unit, Institute of Education, London
7. Origin of DevOps movement around 2009
• The Velocity Conference movement - “10 Deploys A Day”
• The “infrastructure as code” movement (Mark Burgess and Luke
Kanies)
• The “Agile infrastructure” movement (Andrew Shafer)
• The Agile system administration movement (Patrick DeBois)
• The Lean Startup movement by Eric Ries
• The continuous integration and release movement by Jez Humble
• The widespread availability of cloud and PaaS (platform as a
service) technologies
Top11ThingsToKnowAboutDevOp
9. Defining DevOps
• DevOps is a professional movement to adopt modern roles,
cultural behaviors, practices, design patterns and technology
for acquiring, developing and operating software services with
increased agility and reliability.
• DevOps integrates developers and operations teams in order
to improve collaboration and productivity by automating
infrastructure, automating workflows and continuously measuring
application performance - rackspace -
10. Team culture for transformation
http://christopher5106.github.io/cloud/2015/12/26/
future-cloud-programming-AWS-lambda-functions-
jekyll-github-pages-newsletter-emailing.html
Adrian cockcroft @ nginxconf 2014
11. New Approaches to IT Process and Culture
https://www.rackspace.com/devops
13. What does DevOps-Oriented IT look like?
Increased Level of Optionality
Products, not projects
Features, changes are built, tested, deployed as an end to end flow
Most changes do not require dedicated analysis & approval
Quality is built in, changes are reversible
Production code is releasable at any time
All products are automatically test-driven, and testable in production
Software is runnable and portable on any cloud, any time
12 Factors: make development equivalent to production
Services can be swapped without affecting the whole
Microservices, APIs, cloud native scale-out
14. What does DevOps-Oriented IT look like?
Low-Cost, Sacrificial Elements
Teams are small, product-oriented, cross-functional
Releases are inexpensive, frequent, and fast
Deployments are automated, continuous, standardized
Servers are prefabricated, replicated, and disposable
Services are published, on demand, and resilient
15. DevOps Metrics in IT Performance
• Throughput Measures
– Deployment frequency
– Deployment Lead time
• Stability Measures
– Mean Time To Recover (MTTR)
16. DevOps Metrics in IT Performance
Source – 2016_state_of_devops_report
17. Benchmark Company Performance: Faster, Higher
Quality, Higher Results 2015
Source - https://puppetlabs.com/2015-devops-report
30x more frequent deployments
60x better change success rate
168x faster mean time to recovery (MTTR)
200x faster lead times to change
200% more likely to exceed mission, profitability,
productivity goals
50% higher market capitalization growth over 3 years
18. Benchmark Company Performance: Faster, Higher
Quality, Higher Results 2016
Source – 2016_state_of_devops_report
30x => 200x more frequent deployments
60x better change success rate
168x => 24x faster mean time to recovery (MTTR)
200x => 2555x faster lead times to change
200% more likely to exceed mission, profitability,
productivity goals
50% higher market capitalization growth over 3 years
19. ROI of DevOps
• Savings
– Cost of excess rework = Technical staff size * Average salary *
benefits multiplier * percentage of technical staff time spent on
excess rework
– Cost of downtime = Deployment frequency * Change failure
rate * Mean time to recover * hourly cost of outage
• Value
– Do better from the saved engineering time
Source – 2016_state_of_devops_report
20. ROI of DevOps: Do better from saving
Source – 2016_state_of_devops_report
23. The First Way: Systems Thinking
• Flow: Extending Delivery to Operations
– Continuous Delivery through to Production
– Treat Operations as a Platform – “done” is “running in production”
24. The Second Way: Amplify Feedback Loops
• Feedback: Extending Operations into Delivery
– Developers have insight into production
– Fail fast, fail often fail cheaply; Blameless post mortems
– Cloud Native Platform – deployments are cheap, servers are disposable
25. The Third Way: Culture of Continual
Experimentation And Learning
• Continuous Improvement:
– Iterative Experimentation & Learning throughout the organization
– Embed delivery knowledge into operations
▪ Cross-functional teams ; context carried throughout process
– Embed operations knowledge into delivery
▪ Delivery culture of “code for production quality”
26. How can an organization evolve towards a DevOps
Culture?
• Continuous Improvement Program
• Cloud Native Architecture & Technology
• Mapping Roles to Business Value
• Cloud Native Tools
27. Crafting a Continuous Improvement Program
• Metrics
– Track Production mean-time to recover (MTTR), lead times, rework
• Program
– Wardley maps, Value stream mapping, waste identification
– Iterative improvement and migration to cloud native technology
• Sync with Delivery Products
– Waste removal & reducing lead times every month
• Impact the Culture Iteratively:
– Top Down Mandate and Steering
– Bottom Up Empowerment and Decision Making
– Building Successes into a Force Multiplier
28. DevOps Metrics in IT Performance
• Throughput Measures
– Deployment frequency
– Deployment Lead time
• Stability Measures
– Mean Time To Recover (MTTR)
puppet_2015-state-of-devops-report.pdf
32. Map your organization roles to business value
INFRASTRUCTURE
SITE RELIABILITY
PLATFORM
Innova6on; Plan, design, develop and test
business capabili4es as deployable ar4facts
Produc6on Apps; config, deployment, QA,
monitoring, scaling
App Pla>orm; upgrades, patches, capacity
planning, service mgmt., scale pla>orm
Infra Pla>orm; Rack and stack, networking,
data storage, etc.
ROLES
Cross-Func4onal
(Prod. Owner, Dev, QA)
Applica4on Operators
Pla>orm Operators
Engineering
(Storage, Security, Network, etc.)
AREAS OF FOCUS
BUSINESS CAPABILITY
33. Platform Support for CD Processes
● Baseline environments consistent
● Inject environment-specific config and external dependencies
● Consistent API across environments and operations
● Promote apps with same process
● Support for advanced operations
Production
Integration
Staging
Development
Instant App
Environment
Automated
Deployment
Autoscaling A/B Testing
Canary /
Zero
Downtime
Deployment
35. Platform Support for CD Processes
Distributed revision
control and source
code management.
Collaborative software
development
Build and test
software projects
continuously and
incrementally.
Hundreds of
compatible plugins
Share binaries and
manage distributions.
Manage artifact
lifecycle. Avoid
license violations
Develop, test, QA and production on the same platform. Simple, developer friendly
commands and APIs. Operational benefits for every app. Built-in ecosystem
services. Deploy, operate and scale on any IAAS
AUTOMATION.
Integrate tools and automate
processes from testing to builds and
deployment
SPEED.
Release more frequently with smaller
bits will reduce complexity and
improve time-to-market
QUALITY.
Reduce feedback loop using test-
driven development to surface
problems sooner and be responsive
AGILITY.
Push updates on regular basis with no
downtime to improve customer
experience and time to market
Commit Code
Change
Automate
Build & Test
(Unit Test, Static Code
Analysis)
Store Binaries
& Build
Artifacts
Automated
Integration
Testing
Acceptance,
Performance &
Load
Zero Downtime
Upgrade to
Production
BuildPipeline
OperationsToolChain
GitlLab Jenkins Development Test + UAT + Staging Production
38. DevOps Tools for CLOUD NATIVE
Today, these old divisions are breaking down, with the IT and
developer roles merging and following a series of systematic
principles:
1. Infrastructure as code
2. Continuous deployment
3. Automation
4. Monitoring
5. Security
45. Cloud Native tool’s value
45
CI
Pivotal
Labs
Tracing
Metrics
API/ID
Data Flow
ETL
Integration
Stream/Event
Model
VALUE
LINE
Google AWS Azure VMW Openstack
OSembedded
CPI
Container Orchestration
Cloud Orchestration
Java Microservices
Spring Boot
Spring Cloud Netflix OSS
.Net Microservices
with Spring Cloud
46. Full Application lifecycle management
DevelopmentTest + UAT + Staging
Production
Jfrog Artifactory
CloudBees Jenkins
Spring Cloud Spring Bootrepository Git repo
concourse
47. Cloud Native Framework
Application Framework
Runtime Platform
Infrastructure Automation
Infrastructure
Culture
Dev
Dev IT Ops
IT Ops
IT Ops
Tools
BOSH
AWS VMWare OpenStackAzure
Spring Cloud Spring Boot
DevOps
Agile
48. Significant ROI Through Automation
Continuously Deliver High Quality Software Using Pivotal Cloud Foundry
CUT COSTS
GET
EARLIER ROI
MUCH HIGHER
QUALITY
FASTER TIME-
TO-MARKET
SHIP CODE 30x
TIMES FASTER!
90% LOWER
OPERATIONAL
COSTS
50% FEWER
FAILED
DEPLOYMENTS
& RESTORE 12x
FASTER!
GET CODE OUT
THE DOOR
SOONER AND
UPDATE IT
MORE OFTEN
49. Why Pivotal Cloud Foundry?
Use a Mature Platform and Focus on Business Innovation with NO-vendor Lock-in
FOCUS ON INNOVATION,
NOT INFRASTRUCTURE ECOSYSTEM AND BACKING
THE LEADING PAAS
DISTRIBUTION
DIY (and OSS) can inhibit 6me-to-
market, and be costly / risky. Many
customers move to commercial PaaS.
• Pivotal Cloud Foundry is Proven at Scale
• 250+ People Dedicated to PCF R&D
• PCF is IaaS Agnos4c and Highly Extensible;
It Will Meet Your Specific Needs
• Leave the Pla>orm to Us - Spend Your
Time Innova4ng With SoZware
Led by a non-profit founda6on with 52
members. Pivotal leads the way with a
90%+ of commits.
• Broad Industry Acceptance
• 2,123 OSS Code Developers
• Proven Scale in Produc4on Contexts
• Very Well Funded, Stable and Growing
• Pivotal’s Leadership and Influence
• Pre-Engineered Infrastructure Solu4ons
Produc6ve and powerful Operator and
Developer features. Expansive list of
Services and partners (e.g. CloudBees).
• $40m of Year One Product Revenue
• Large Financial Services Customers
• Produc4on-Class Data (e.g. MySQL),
Mobile and App (e.g. Ne>lix OSS) Services
• Rich, IaaS Agnos4c, “Day 2” Capabili4es
• Awesome Developer CLI and Console
50. Summary
• DevOps practices, patterns, and technology provide a new paradigm for
IT management
• Time to market, focusing business improvements
• Process and Culture Should Map to Need
– No such things as a “one size fits all” IT organization
– Metrics, Improve Program, CI Pipeline
• DevOps Team requires cloud-native tools for productivity and agility in
improvement and learning
– Small Experiments, Fail Fast & Inexpensive Feedback
– Iterative improvement and migration to cloud native technology