GridHypervisor
The aim of GridHypervisor is to enable the large-scale, reliable, efficient and dynamic deployment and re-allocation of virtual machines between different administrative domains, each with its own security policy and local virtualization management technology.
Functionality
GridHypervisor technology will perform management and provisioning of virtual machines (VMs) in a grid infrastructure following the next steps:
- Physical Resource Discovery and Selection: Based on a set of requirements a list of appropriate resources will be obtained by accessing to an information service. Then a single resource is selected among the candidate resources in the list based on a pre-defined set of scheduling policies.
- Remote System Preparation: The image is transferred and the selected resource is prepared for VM execution.
- Start and Migration: The VM is submitted to the selected resource. However, the VM could be relocated on a different resource if a performance slowdown (SLAs) is detected or a better resource is discovered.
- Monitoring: The VM evolution is monitored over time.
- Termination: Its owner is notified and some completion tasks are performed.
GridHypervisor will rely on Globus components, mainly on the Virtual Workspaces Service (VWS). Globus VWS allows an authorized Grid client to start, shutdown and inspect an environment described by the workspace meta-data on a specified resource quota. WMS provides a WSRF interface and the security required in a Grid environment for submitting and controlling VMs on remote resources. In this sense, WMS could be used as a front-end to local VM Managers.
See also
- Globus Toolkit
- Virtual Machine
- Web Services Resource Framework
- Global Grid Forum