Autonomic workflow management in dynamically federated, hybrid cloud infrastructures

1. An autonomic workflow framework comprising:

a federated cloud comprising a plurality of clouds, wherein each cloud in the federated cloud comprises one or more nodes;

a workflow manager configured to receive a workflow, wherein the workflow comprises a plurality of stages, wherein the workflow is associated with a user objective, wherein the user objective comprises one or more of the following:

a budget constraint,

a time constraint, and

a privacy constraint: and

an autonomic scheduler in communication with the workflow manager, wherein the autonomic scheduler is configured to provision one or more of the one or more nodes to process the stages to achieve the user objective.

2. The autonomic workflow framework of claim 1 , wherein:

the workflow manager is further configured to submit the stages to an autonomic cloud engine space; and

each cloud comprises one or more application agents that are each configured to retrieve one or more of the stages from the autonomic cloud engine space.

3. The autonomic workflow framework of claim 2 , wherein the application agents are each configured to process the stages according to a consumption policy.

4. The autonomic workflow framework of claim 3 , wherein the application agents are each configured to process the stages using a master/worker paradigm.

5. The autonomic workflow framework of claim 2 , wherein the application agents are each further configured to send one or more completed tasks for the stages to the workflow manager.

6. The autonomic workflow framework of claim 1 , further comprising a resource manager in communication with the autonomic scheduler, wherein the resource manager is configured to:

collect resource status information from the one or more nodes of the one or more clouds in the federated cloud; and

provide at least a portion of the collected resource status information to the autonomic scheduler.

7. The autonomic workflow framework of claim 1 , wherein the autonomic scheduler is further configured to monitor a resource status of the nodes, wherein the resource status comprises one or more of the following:

CPU availability of a node;

memory availability of a node; and

network bandwidth of a node.

8. The autonomic workflow framework of claim 1 , wherein the autonomic scheduler is further configured to manage one or more of the following:

one or more clouds joining the federated cloud; and

one or more clouds leaving the federated cloud.

9. The autonomic workflow framework of claim 1 , wherein the autonomic scheduler is further configured to:

monitor a progress of workflow processing; and

adjust the provisioning in response to determining that processing will not achieve the user objective.

10. The autonomic workflow framework of claim 1 , wherein the workflow manager is further configured to:

determine whether one or more completed tasks associated with one or more stages are received within a time period; and

in response to determining that that the one or more completed tasks associated with the one or more stages are not received within the time period, determining whether the user objective is violated.

11. The autonomic workflow framework of claim 1 , wherein the plurality of clouds comprises at least one public cloud.

12. The autonomic workflow framework of claim 1 , wherein the plurality of clouds comprises at least one private cloud.

13. An autonomic workflow framework comprising:

a workflow manager configured to receive a workflow, wherein the workflow comprises a plurality of stages, wherein the workflow is associated with a user objective, wherein the user objective comprises one or more of the following:

a budget constraint,

a time constraint, and

a privacy constraint;

an autonomic scheduler in communication with the workflow manager, wherein the autonomic scheduler is configured to provision one or more of the nodes to process each stage in order to achieve the user objective;

a resource manager in communication with the autonomic scheduler;

a federated cloud comprising a plurality of clouds, wherein each cloud in the federated cloud comprises:

one or more nodes,

a cloud agent in communication with the resource manager, and

one or more application agents; and

an autonomic engine application space that is accessible by the autonomic scheduler, the workflow manager and each application agent.

14. The autonomic workflow framework of claim 13 , wherein:

each cloud agent is configured to send a join request to the autonomic scheduler, wherein the join request comprises a request for the cloud corresponding to the cloud agent to join the federated cloud and a list of addresses of its node; and

the autonomic scheduler is further configured to register the cloud corresponding to the cloud agent that sends the join request into the federated cloud.

15. The autonomic workflow framework of claim 13 , wherein the resource manager is configured to:

receive node availability information from at least one cloud agent, wherein the node availability information indicates whether one or more nodes of the cloud associated with the at least one cloud agent are available to process a stage; and

send at least a portion of the node availability information to the autonomic scheduler.

16. The autonomic workflow framework of claim 13 , wherein the autonomic scheduler is configured to send a list identifying the provisioned nodes of at least one of the clouds to the cloud agent of the at least one cloud; the cloud agent of the at least one cloud is configured to use the list to manage processing by the provisioned nodes of the at least one cloud; and the cloud agent is configured to manage processing of one or more stages by the provisioned nodes.

17. The autonomic workflow framework of claim 13 , wherein the plurality of clouds comprises at least one public cloud.

18. The autonomic workflow framework of claim 13 , wherein the plurality of clouds comprises at least one private cloud.