4. Everything is a Model
No need to orchestrate API calls
No code to maintain
Decoupled Binding Model
Containers talk to Services
Services forward traffic to POD
Advanced Lifecycle Management
Rolling updates and canary deployments
POD Auto-scaling
Container + Infrastructure
Infrastructure Provisioning on demand
Support for GCE, AWS, OpenStack
Why Kubernetes?
10. Containers behave differently per environment
Most deployment issues are related to configuration differences/incompatibilities.
Not all containers work on all container management platforms.
Make deployment commutative and test for it.
Invest in a fast repo
Pulling containers is the most time consuming.
Add a test for the size of your containers.
Understand network model of your container management system
Network overlays, bridges, intra containers communication.
Log everything
Logs are the new ssh
What we learned?
12. Authentication/Authorization:
Move towards a native Authorization model
SAML Authentication
SAML Claims for namespace membership
Quota Management UI
Management:
Trigger for rolling updates and canary deployments
Email Notifications for lifecycle events
Integration:
Audit Trail and event archiving
Git authentication support
Roadmap
Editor's Notes
451 report on DevOps, Q1 2014
Gartner, Seven Steps to Start Your Devops Initiative, Sept 2014
451 report on DevOps, Q1 2014
Gartner, 2014 CIO Survey
Cloud application lifecycle is broken
More than 60% of businesses would like to deploy more frequently
51% of enterprises cite business and strategy demanding to reduce application cycles
Only 16% of businesses are using model-driven automation tools
Docker file contains enough info for you to create an image that contains the binaries of an application. At a low level, the inputs/outputs of that application.
Docker file contains bare minimum to deploy. But it’s not rich enough to contain meta data. Not enough visibility of changes; rich parameterization.
For example -
Docker file contains enough info for you to create an image that contains the binaries of an application. At a low level, the inputs/outputs of that application.
Docker file contains bare minimum to deploy. But it’s not rich enough to contain meta data. Not enough visibility of changes; rich parameterization.
For example -
Docker file contains enough info for you to create an image that contains the binaries of an application. At a low level, the inputs/outputs of that application.
Docker file contains bare minimum to deploy. But it’s not rich enough to contain meta data. Not enough visibility of changes; rich parameterization.
For example -
Docker file contains enough info for you to create an image that contains the binaries of an application. At a low level, the inputs/outputs of that application.
Docker file contains bare minimum to deploy. But it’s not rich enough to contain meta data. Not enough visibility of changes; rich parameterization.
For example -