Cloud Computing: Remote services on demand
- Automated provisioning
- Self Service - automatic provisioning of which device service given to whom etc.
- Multitenancy - consistent access by different users
- Elasticity - Scale up and down
- Telemetry - resource monitoring by consumer and provider
- Scale out rather than up - horizontal scaling rather than vertical
- Create clones rather than upgrading the single device.
- For vertical scaling, once up difficult or should I say impossible to scale down
1. Private Cloud - Private to organization and they may even have their own hardware
2. Public Cloud - cloud service providers
3. Hybrid Cloud - private cloud for standard resources and public cloud for user interface.
OpenStack is IAAS and OpenShift is PAAS
- IAAS can be implemented by two ways
- Generic Infrastructure - RHEV-M (Red Hat Enterprise Virtualization)
- Cloud based - RHELOSP (Red Hat Enterprise Linux OpenStack Platform)
Traditional Virtualization: Stateful
Cloud Infrastructure: Greater Scalability with shared resources
- Login into Horizon Interface for RHELOSP
- Go to Project -> Orchestration -> Stacks
- Click on Launch Stack
- Template Source - URL - http://classroom.example.com/materials/heat/scaling.template
- Environment Source - override values in template
Environment.yaml
parameters:
key_name: <<>>
flavor: <<>>
image: <<>>
- Click on next
- Give a Stack Name
- Check rollback on failure
- Leave rest details to default
- Launch the stack
- Stack is created - now make it automatic scalable.
- Click on compute - instances
- Select an instance - click on console tab - Access console
- Run the script
$ ./launchme.sh
- The script runs for 5 minutes and configures the scaling module and then goes into a cooldown period.