Distributed artificial intelligence fabric controller

Patent metadata

Field	Value
Publication number	US-12353920-B2
Application number	US-202117644961-A
Country	US
Kind code	B2
Filing date	Dec 17, 2021
Priority date	Dec 29, 2020
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

In general, this disclosure describes techniques for configuring and provisioning, with a distributed artificial intelligence (AI) fabric controller, network resources in an AI fabric for use by AI applications. In one example, the AI fabric controller is configured to discover available resources communicatively coupled to a cloud exchange; obtain a set of candidate solutions, each candidate solution of the set of candidate solutions comprising an AI application and a configuration of resources for use by the AI application; filter, based on one or more execution metrics corresponding to each of the candidate solutions, the set of candidate solutions to generate a filtered set of candidate solutions; generate provisioning scripts for the filtered set of candidate solutions; execute the provisioning scripts to provision resources for each candidate solution in the filtered set of candidate solutions; and create an execution environment for each candidate solution in the filtered set of candidate solutions.

First claim

What is claimed is:

1. A computing device comprising:
processing circuitry coupled to a memory;
an artificial intelligence (AI) fabric controller configured for execution by the processing circuitry, wherein the AI fabric controller comprises:
a monitoring service configured to discover available resources communicatively coupled to a cloud exchange and to monitor the available resources;
an analysis service configured to:
obtain a set of candidate solutions, each candidate solution of the set of candidate solutions comprising an AI application and a configuration of resources for use by the AI application, wherein the configuration of resources is different for each of the candidate solutions, and wherein each of the candidate solutions is associated with corresponding execution metrics, and
filter, based on one or more of the corresponding execution metrics, the set of candidate solutions to generate a filtered set of candidate solutions, the filtered set of candidate solutions comprising a plurality of candidate solutions;

a planning service configured to generate provisioning scripts for the filtered set of candidate solutions; and
an execution service configured to:
execute the provisioning scripts to provision resources for each candidate solution in the filtered set of candidate solutions; and
create an execution environment for each candidate solution in the filtered set of candidate solutions.

2. The computing device of claim 1 , wherein the AI fabric controller is configured to determine the corresponding execution metrics associated with a corresponding candidate solution by executing an instance of the AI application for the corresponding candidate solution of the candidate solutions in accordance with the configuration of resources associated with the corresponding candidate solution.

3. The computing device of claim 1 , wherein the AI fabric controller is configured to select, from the filtered set of candidate solutions, an optimal candidate solution based on the execution metrics associated with each of the candidate solutions and to deploy the optimal candidate solution to a production environment.

4. The computing device of claim 1 , wherein the monitoring service adds resource data describing an available resource as a node to a graph of available resources and monitors the available resources associated with the nodes in the graph.

5. The computing device of claim 1 , wherein the monitoring service monitors one or more of resource utilization, resource availability, resource capacity and resource performance for each corresponding available resource of the available resources.

6. The computing device of claim 1 , wherein the analysis service is triggered to execute in response to a new resource being added as an available resource.

7. The computing device of claim 1 , wherein the execution metrics associated with a corresponding solution comprise one or more of a machine learning model accuracy, a resource cost, and a privacy setting.

8. The computing device of claim 1 , wherein the analysis service orders the set of candidate solutions according to an infrastructure cost associated with each respective candidate solution of the set of candidate solutions.

9. The computing device of claim 1 , wherein a candidate solution comprises the AI application configured to operate in a centralized mode, and wherein the planning service is configured to:
determine that the AI application can execute in a federated mode;
in response to the determination that the AI application can execute in a federated mode, generating an additional candidate solution corresponding to operating the AI application in the federated mode; and
adding the additional candidate solution to the set of candidate solutions.

10. The computing device of claim 1 , wherein to create the execution environment for each candidate solution comprises to create the execution environment using a container orchestration platform.

11. A method comprising:
discovering, by one or more processors, available resources communicatively coupled to a cloud exchange;
obtaining, by the one or more processors, a set of candidate solutions, each candidate solution of the set of candidate solutions comprising an artificial intelligence (AI) application and a configuration of resources for use by the AI application, wherein the configuration of resources is different for each of the candidate solutions, and wherein each of the candidate solutions is associated with corresponding execution metrics, and
filtering, by the one or more processors, based on one or more of the corresponding execution metrics, the set of candidate solutions to generate a filtered set of candidate solutions, the filtered set of candidate solutions comprising a plurality of candidate solutions;
generating, by the one or more processors, provisioning scripts for the filtered set of candidate solutions;
executing, by the one or more processors, the provisioning scripts to provision resources for each candidate solution in the filtered set of candidate solutions; and
creating an execution environment for each candidate solution in the filtered set of candidate solutions.

12. The method of claim 11 , further comprising executing an instance of the AI application for a corresponding candidate solution of the candidate solutions in accordance with the configuration of resources associated with the corresponding candidate solution to determine the corresponding execution metrics associated with the corresponding candidate solution.

13. The method of claim 11 , further comprising:
selecting, from the filtered set of candidate solutions, an optimal candidate solution based on the execution metrics associated with each of the filtered set of candidate solutions; and
deploying the optimal candidate solution to a production environment.

14. The method of claim 11 , further comprising adding resource data describing an available resource as a node to a graph of available resources; and
monitoring the available resources associated with the nodes in the graph.

15. The method of claim 14 , wherein monitoring the available resources comprises monitoring one or more of resource utilization, resource availability, resource capacity and resource performance for each corresponding available resource of the available resources.

16. The method of claim 11 , wherein the execution metrics associated with a corresponding solution comprise one or more of a machine learning model accuracy, a resource cost, and a privacy setting.

17. The method of claim 11 , further comprising ordering the set of candidate solutions according to an infrastructure cost associated with each respective candidate solution of the set of candidate solutions.

18. The method of claim 11 , wherein a candidate solution comprises the AI application configured to operate in a centralized mode, and wherein the method further comprises:
determining that the AI application can execute in a federated mode;
in response to determining that the AI application can execute in a federated mode, generating an additional candidate solution corresponding to operating the AI application in the federated mode; and
adding the additional candidate solution to the set of candidate solutions.

19. The method of claim 11 , wherein creating the execution environment for each candidate solution comprises creating the execution environment using a container orchestration platform.

20. A non-transitory computer readable medium comprising

Assignees

• Equinix Inc

Inventors

• Voruganti Kaladhar
• Li Danjue
• Wang Lanfa
• Arisoylu Mustafa
• Pasula Ravi Kiran
• Elder Rodney Martin
• Liu Zoe

Classifications

• G06F2209/508

  Monitor · CPC title

• G06F11/3409

  for performance assessment · CPC title

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• G06F9/5077

  Logical partitioning of resources; Management or configuration of virtualized resources (specific details on emulation or internal functioning of virtual machines G06F9/455) · CPC title

• G06F9/5072 Primary

  Grid computing · CPC title