Multi-level workflow scheduling using metaheuristic and heuristic algorithms
Techniques described herein relate to a method for deploying workflows. The method may include receiving, by a global orchestrator of a device ecosystem, a request to execute a workflow; decomposing, by the global orchestrator, the workflow into a plurality of workflow portions; executing, by the global orchestrator, a metaheuristic algorithm to generate a result comprising a plurality of domains of the device ecosystem in which to execute the plurality of workflow portions; and providing, by the global orchestrator, the plurality of workflow portions to respective local orchestrators of the plurality of domains based on the result of executing the metaheuristic algorithm.
1. A method for deploying workflows, the method comprising:
receiving, by a global orchestrator of a device ecosystem, a request to execute a workflow;
decomposing, by the global orchestrator of the device ecosystem, the workflow into a plurality of workflow portions;
executing, by the global orchestrator of the device ecosystem, a metaheuristic algorithm to generate a result comprising a plurality of domains of the device ecosystem in which to execute the plurality of workflow portions;
providing, by the global orchestrator of the device ecosystem, the plurality of workflow portions to respective local orchestrators of the plurality of domains based on the result of executing the metaheuristic algorithm, wherein a workflow portion of the plurality of workflow portions is provided to a local orchestrator of a domain of the plurality of domains along with additional workflow portion requirements; and
composing, by the local orchestrator of the domain of the plurality of domains, a plurality of devices in the domain of the plurality of domains for performing the workflow portion of the plurality of workflow portions, wherein the composing comprises performing a local optimization based on the additional workflow portion requirements.
2. The method of claim 1 , wherein decomposing the workflow comprises performing a trace analysis based on the workflow.
3. The method of claim 2 , wherein results of the trace analysis comprise a critical path associated with the workflow.
4. The method of claim 1 , wherein the local optimization comprises reserving resources of the domain of the plurality of domains for execution of the workflow portion of the plurality of workflow portions.
5. The method of claim 1 , wherein the local optimization comprises using a best fit composition based on a prior workflow portion execution.
6. The method of claim 1 , further comprising:
transmitting, by the local orchestrator of the domain of the plurality of domains and to the global orchestrator of the device ecosystem, workflow portion execution metrics associated with execution of the workflow portion of the plurality of workflow portions.
7. The method of claim 1 , wherein decomposing the workflow comprises identifying an anchor point for the workflow.
8. The method of claim 7 , further comprising, before executing the metaheuristic algorithm:
pruning a global graph representing the device ecosystem based on the anchor point for the workflow to obtain a pruned global graph,
wherein the metaheuristic algorithm is executed using the pruned global graph.
9. The method of claim 1 , wherein the metaheuristic algorithm is constrained by one parameter selected from a group of parameters consisting of time and cost.
10. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for deploying workflows, the method comprising:
receiving, by a global orchestrator of a device ecosystem, a request to execute a workflow;
decomposing, by the global orchestrator of the device ecosystem, the workflow into a plurality of workflow portions;
executing, by the global orchestrator of the device ecosystem, a metaheuristic algorithm to generate a result comprising a plurality of domains of the device ecosystem in which to execute the plurality of workflow portions;
providing, by the global orchestrator of the device ecosystem, the plurality of workflow portions to respective local orchestrators of the plurality of domains based on the result of executing the metaheuristic algorithm, wherein a workflow portion of the plurality of workflow portions is provided to a local orchestrator of a domain of the plurality of domains along with additional workflow portion requirements; and
composing, by the local orchestrator of the domain of the plurality of domains, a plurality of devices in the domain of the plurality of domains for performing the workflow portion of the plurality of workflow portions, wherein the composing comprises performing a local optimization based on the additional workflow portion requirements.
11. The non-transitory computer readable medium of claim 10 , wherein decomposing the workflow comprises performing a trace analysis based on the workflow.
12. The non-transitory computer readable medium of claim 11 , wherein results of the trace analysis comprise a critical path associated with the workflow.
13. The non-transitory computer readable medium of claim 10 , wherein the local optimization comprises reserving resources of the domain of the plurality of domains for execution of the workflow portion of the plurality of workflow portions.
14. The non-transitory computer readable medium of claim 10 , wherein the local optimization comprises using a best fit composition based on a prior workflow portion execution.
15. The non-transitory computer readable medium of claim 10 , wherein the method performed by executing the computer readable program code further comprises:
transmitting, by the local orchestrator of the domain of the plurality of domains and to the global orchestrator of the device ecosystem, workflow portion execution metrics associated with execution of the workflow portion of the plurality of workflow portions.
16. The non-transitory computer readable medium of claim 10 , wherein decomposing the workflow comprises identifying an anchor point for the workflow.
17. The non-transitory computer readable medium of claim 16 , wherein the method performed by executing the computer readable program code further comprises, before executing the metaheuristic algorithm:
pruning a global graph representing the device ecosystem based on the anchor point for the workflow to obtain a pruned global graph,
wherein the metaheuristic algorithm is executed using the pruned global graph.
18. A system for deploying workflows, the system comprising:
a global orchestrator for a device ecosystem, the global orchestrator comprising a processor and memory, wherein the global orchestrator is operatively connected to a plurality of local orchestrators, and configured to: receive a request to execute a workflow;
decompose the workflow into a plurality of workflow portions;
execute a metaheuristic algorithm to generate a result comprising a plurality of domains of the device ecosystem in which to execute the plurality of workflow portions;
provide the plurality of workflow portions to respective local orchestrators of the plurality of domains based on the result of executing the metaheuristic algorithm, wherein a workflow portion of the plurality of workflow portions is provided to a local orchestrator of a domain of the plurality of domains along with additional workflow portion requirements; and
composing, by the local orchestrator of the domain of the plurality of domains, a plurality of devices in the domain of the plurality of domains for performing the workflow portion of the plurality of workflow portions, wherein the composing comprises performing a local optimization based on the additional workflow portion requirements.