Opportunistic placement of compute in an edge network

1. A node in an edge network, comprising:

a processor; and

memory to store instructions, which when executed by the processor, cause the node to:

access a service level agreement related to a workload, the workload to be orchestrated for a user equipment by the node;

modify an implementation of a machine learning model based on the service level agreement;

implement the machine learning model to identify resource requirements to execute the workload in a manner to satisfy the service level agreement;

initiate resource assignments from a resource provider, the resource assignments to satisfy the resource requirements;

construct a resource hierarchy from the resource assignments;

initiate execution of the workload using resources from the resource hierarchy; and

monitor and adapt execution of the workload based on the resource hierarchy in response to the execution of the workload.

2. The node of claim 1 , wherein the instructions to modify the implementation of the machine learning model cause the node to:

access user preferences; and

apply weights to inputs for the machine learning model to fit resource costs to the user preferences.

3. The node of claim 2 , wherein the user preferences are stored in a spider diagram.

4. The node of claim 1 , wherein the instructions to modify the implementation of the machine learning model cause the node to:

evaluate trusted computing capabilities of nodes that host resources; and

use the evaluated trusted computing capabilities of nodes in a cost function to identify resource requirements to execute the workload in a manner to satisfy the service level agreement.

5. The node of claim 1 , wherein to evaluate trusted computing capabilities, the node is to:

receive attestation from a node, the attestation indicating that the node is capable of trusted computing operations.

6. The node of claim 1 , wherein the instructions to initiate resource assignments from the resource provider cause the node to:

implement a cost function to identify resources for the resource assignments, the cost function based on a latency map, a storage hierarchy, a computation map, or a resource redundancy depth.

7. The node of claim 1 , wherein the instructions to construct the resource hierarchy cause the node to:

construct a primary resource map including resources to be used initially to begin execution the workload; and

construct a resiliency resource map including resources that can be used when a resource from the primary resource map becomes unavailable, unresponsive, or fails;

wherein the primary resource map and the resiliency resource map are integrated into the resource hierarchy.

8. The node of claim 1 , wherein the instructions to construct the resource hierarchy cause the node to:

construct the resource hierarchy using a latency-based metric of each of the resources from the resource assignments.

9. The node of claim 1 , wherein the instructions to construct the resource hierarchy cause the node to:

implement a cost function to construct the resource hierarchy, the cost function used to weigh a plurality of properties including:

resource response time, resource availability, resource durability, resource resiliency, resource consistency, or resource integrity.

10. The node of claim 9 , wherein the cost function is weighted based on user preferences, administrator policies, client policies, or the service level agreement.

11. The node of claim 10 , wherein the user preferences are represented in a spider diagram.

12. The node of claim 1 , wherein the instructions to initiate execution of the workload cause the node to:

initiate a portion of workload execution at the user equipment.

13. The node of claim 12 , wherein the service level agreement includes an indication that the user equipment is available for use as a resource to execute the workload.

14. The node of claim 1 , wherein the instructions to monitor and adapt execution of the workload cause the node to:

migrate the execution of the workload from resources provided by a first resource provider to resources provided by a second resource provider.

15. The node of claim 1 , wherein the instructions to monitor and adapt execution of the workload cause the node to:

migrate orchestration of the workload from a first resource management controller to a second resource management controller, where the second resource management controller has a closer latency-based measurement to the user equipment than the first resource management controller.

16. The node of claim 1 , wherein the instructions to monitor and adapt execution of the workload cause the node to:

measure resource usage of the execution of the workload; and

adapt execution of the workload when the resource usage exceeds a threshold resource usage.

17. A method, comprising:

accessing a service level agreement related to a workload, the workload to be orchestrated for a user equipment by a node in an edge network;

modifying an implementation of a machine learning model based on the service level agreement;

implementing the machine learning model to identify resource requirements to execute the workload in a manner to satisfy the service level agreement;

initiating resource assignments from a resource provider, the resource assignments to satisfy the resource requirements;

constructing a resource hierarchy from the resource assignments;

initiating execution of the workload using resources from the resource hierarchy; and

monitoring and adapt execution of the workload based on the resource hierarchy in response to the execution of the workload.

18. The method of claim 17 , wherein modifying the implementation of the machine learning model comprises:

evaluating trusted computing capabilities of nodes that host resources; and

using the evaluated trusted computing capabilities of nodes in a cost function to identify resource requirements to execute the workload in a manner to satisfy the service level agreement.

19. The method of claim 17 , wherein evaluating trusted computing capabilities comprises:

receiving attestation from a node, the attestation indicating that the node is capable of trusted computing operations.

20. The method of claim 17 , wherein initiating resource assignments from the resource provider comprises:

implementing a cost function to identify resources for the resource assignments, the cost function based on a latency map, a storage hierarchy, a computation map, or a resource redundancy depth.

21. The method of claim 17 , wherein constructing the resource hierarchy comprises:

constructing a primary resource map including resources to be used initially to begin execution the workload; and

constructing a resiliency resource map including resources that can be used when a resource from the primary resource map becomes unavailable, unresponsive, or fails;

wherein the primary resource map and the resiliency resource map are integrated into the resource hierarchy.

22. The method of claim 17 , wherein constructing the resource hierarchy comprises:

constructing the resource hierarchy using a latency-based metric of each of the resources from the resource assignments.

23. At least one machine-readable medium including instructions, which when executed by a node, cause the node to:

access a service level agreement related to a workload, the workload to be orchestrated for a user equipment by the node;

modify an implementation of a machine learning model based on the service level agreement;

implement the machine learning model to identify resource requirements to execute the workload in a manner to satisfy the service level agreement;

initiate resource assignments from a resource provider, the resource assignments to satisfy the resource requirements;

construct a resource hierarchy from the resource assignments;

orchestrate execution of the workload using resources from the resource hierarchy; and

monitor and adapt execution of the workload based on the resource hierarchy in response to the execution of the workload.

24. The at least one machine-readable medium of claim 23 , wherein the instructions to monitor and adapt execution of the workload cause the node to:

migrate the execution of the workload from resources provided by a first resource provider to resources provided by a second resource provider.

25. The at least one machine-readable medium of claim 23 , wherein the instructions to monitor and adapt execution of the workload cause the node to:

migrate orchestration of the workload from a first resource management controller to a second resource management controller, where the second resource management controller has a closer latency-based measurement to the user equipment than the first resource management controller.