Migrating legacy IT to the cloud can seem like a lot of work, we will share with you a simple framework that can help you migrate your applications at scale. Migration to AWS is the start of a journey that ends with increased business agility and innovation. Learn how you can stepwise adopt DevOps best practices and transform your IT and business culture to get the most out of your investments in cloud technology.
5. On-premises and back again
Prior to 2016: on AWS
In 2016: migrated to on-premises
Objective: reduce operational costs
Moved 80% of infra off AWS
Ran 1+ year on-premises
In 2017 we moved back to AWS.
6. Full control of environment
Control over upgrades
Lower operational costs*
Pros
Lot of stability issues
High operational burden
Hard to find talent
Cons
On-premises: our observations
7. Moving Back to AWS
Analyzed multiple cloud providers
Requirements: location, features, stability
Chose AWS after evaluation
Bonus: we were already familiar with AWS
8. Moving Back to AWS
Identified services and dependencies
Service by service
Infrastructure as code using terraform
Automation using puppet, salt
Parallel deployment to AWS, on-premises
Cut traffic over to AWS, stop on-premises
Zero downtime
Migration complete in 3 months
Huge increase in availability (99.999%)
9. Post-Migration Cost Optimization
Right size EC2 using
metrics
Reserved Instances for
Production
Right Size
Switch ELB to ALB
Consolidate CloudFront
distributions
Use Auto-Scaling
Use Managed Services
Optimize
More services, 30% less
cost (than previous AWS spend)
Constantly Optimizing
Infrastructure
Ongoing Cost Savings
Results
10. Take Aways
Managing your own infrastructure is operationally intensive –
AWS removes the need to do so
Migrating to AWS is not as difficult as you may think
Cost optimization is an on-going exercise
17. IT Map – Traditional IT
Infrastructure PMO Design Engineering Operations
Back-Office/Corporate Systems End User ComputingProducts and Services
Information Security Audit/Compliance
18. A Software Delivery Value Stream
Development &
Unit Testing
Design
Analysis and
Design
Feature
Project
Planning
Backlog
TEST
Env
UAT/Showcase
QA
Tested
Stories
Production
deployment
Change
Approval
Design PMO
EngineeringQAInfra, Risk
Ops
PROD
Env
20. What is DevOps?
In a DevOps organization, software developers and
operational engineers[*] work together as one team, that
shares responsibility for a product or service.
[*] Developers + Testing + Security + Ops + anyone else in the
value stream
21. The Three Ways of DevOps
Repeatable processes
Don’t pass defects on
Local vs Global optima
Optimize flow
Improve
Workflow
Identify all customers
Shorten feedback loops
Amplify feedback
Embed knowledge
Improve
Feedback
Reward calculated risks
Blameless post-mortems
Projects to improve systems
Inject faults, practice resilience
Continuous
Learning
22. The Three Engines of DevOps
No more “silos”
Shared responsibility
Ownership
Visibility
Communication
Culture
Infrastructure as Code
CI/CD
Micro-services
Monitoring and Logging
Practices
Automation
Source
Build
Test
Deploy
Monitor
Faster, smaller releases
Tools
24. Large Scale Migrations
Require different functions to work together
Rapidly learn and adapt
Promote best practices across a large org
Are an opportunity to re-organize
25. IT Map – Cloud First IT
Infrastructure PMO Design Engineering Operations
Back-Office/Corporate Systems End User ComputingProducts and Services
Information Security Audit/Compliance
Infrastructure PMO Design Engineering Operations
26. Cloud Center of Excellence
Governance
IT Map – Cloud First IT
Back-Office /
Corporate
Systems
End User
Computing
Products and Services
Information Security Audit/Compliance
PMO Design Engineering OperationsInfrastructure
Infrastructure PMO Design Engineering Operations
Business App Business App Business App Business App
Standards Training Review Tools
29. Adopting DevOps
Choose a project
Create DevOps team
Document ad-hoc processes
Establish baseline
Prioritize what to automate
Automate
Measure and compare
Extract Best Practices
Pilot
CCoE
Landing Zone
- Templates
- Processes & Best Practices
Education
Prioritized Projects
- Business Apps
- Infrastructure
- Dev Tools
Going Wider
34. Tenet: Improve Feedback
Is the service up?
Why is the website slow?
Why is component X failing?
Problems
Monitoring & Dashboards
Centralized Logging
Solutions
37. Provisioning
Model and provision your
AWS Cloud resources using a
common language
Operations Management
Control and optimize resources with
proper governance and
compliance
Monitoring and Logging
Gain insights into cloud
resources utilization, application
performance and operational health
Fully managed
configuration
management
Managed Services
Control For Your Cloud Environment