Systems and methods to trigger workload migration between cloud-based resources and local resources

Patent metadata

Field	Value
Publication number	US-12223361-B2
Application number	US-202117391935-A
Country	US
Kind code	B2
Filing date	Aug 2, 2021
Priority date	Aug 2, 2021
Publication date	Feb 11, 2025
Grant date	Feb 11, 2025

Abstract

Embodiments of systems and methods are provided to trigger migration of a workload from cloud-based resources to local resources, or vice versa. In the disclosed embodiments, an orchestration service receives telemetry data from a client system associated with a user and cloud resource usage data corresponding to the user from a plurality of cloud service providers. Before the end of each cloud computing service billing cycle, the orchestration service: uses the cloud resource usage data and/or the telemetry data to determine a cloud resource usage, which is expected for the user at the end of the cloud computing service billing cycle; generates a trigger to migrate the user's workload from cloud-based resources to local resources, or vice versa, based on the expected cloud resource usage; and initiates migration of the user's workload if a trigger is generated. As such, the orchestration service can be used to effectively manage per-user cloud resource costs.

First claim

What is claimed is:

1. A computer-implemented method for triggering a migration of a workload, wherein the computer-implemented method is performed by execution of program instructions stored on an orchestration server communicatively coupled via a network to a plurality of cloud service providers and a plurality of user endpoint client devices that are each utilized by a respective user, and wherein the computer-implemented method comprises:
receiving telemetry data from a first one of the user endpoint client devices utilized by a respective first user, wherein the first user endpoint client device comprises local resources for running workloads locally for the first user on the first user endpoint client device;
receiving cloud resource usage data from each of the plurality of cloud service providers, wherein each cloud service provider comprises cloud-based resources for running workloads on remote servers operated by that cloud service provider, and wherein the cloud resource usage data comprises an amount of time the first user utilized the cloud-based resources provided by the cloud service providers during a cloud computing service billing cycle;
wherein before an end of the cloud computing service billing cycle, the computer-implemented method further comprises:
determining a per-day user workload run-rate for the first user using the cloud resource data received from the plurality of cloud service providers;
generating a user persona for the first user using the telemetry data received from the first user endpoint client device;
determining an expected cloud resource usage, which is expected for the first user at the end of the cloud computing service billing cycle, based on the per-day first user workload run-rate and/or the first user persona; and
generating a trigger to migrate the first user's workload based on the expected cloud resource usage;

wherein the telemetry data comprises the telemetry data collected by a client agent executing on the first endpoint client device; and
wherein if a trigger is generated to migrate the user's workload from the cloud-based resources to the local resources of the first user endpoint client device, the computer-implemented method further comprises:
initiating migration of the first user's workload by supplying policies to the cloud service providers and the first user endpoint client device to automatically migrate the first user's workload from a cloud instance of the workload to a local instance of the workload running on the first user endpoint client device, and

wherein if a trigger is generated to migrate the user's workload from the local resources of the first user endpoint client device to the cloud-based resources, the computer-implemented method further comprises:
initiating migration of the first user's workload by supplying policies to the first user endpoint client device and the cloud service providers to automatically migrate the user's workload from a local instance of the workload running on the first user endpoint client device to a cloud instance of the workload.

2. The computer-implemented method of claim 1 , wherein if a trigger is generated to migrate the first user's workload from the cloud-based resources to the local resources of the first user endpoint client device, the computer-implemented method further comprises initiating migration of the user's workload by supplying policies to the cloud service providers and the first user endpoint client device to automatically migrate the user's workload from a cloud instance of the workload to a local instance of the workload running on the first user endpoint client device.

3. The computer-implemented method of claim 1 , wherein if a trigger is generated to migrate the user's workload from the local resources of the first user endpoint client device to the cloud-based resources, the computer-implemented method further comprises initiating migration of the first user's workload by supplying policies to the first user endpoint client device and the cloud service providers to automatically migrate the user's workload from a local instance of the workload running on the first user endpoint client device to a cloud instance of the workload.

4. The computer-implemented method of claim 1 , wherein each cloud service provider provides a plurality of different cloud instance types for running workloads on the remote servers operated by the cloud service provider, and wherein the cloud resource usage data received from the plurality of cloud service providers comprises a number of usage hours accumulated by the user and a corresponding cloud instance rate for each cloud instance type used by the user during the cloud computing service billing cycle.

5. The computer-implemented method of claim 4 , wherein said determining a per-day user workload run-rate for the user comprises:
determining a per-day utilization of each cloud instance type by dividing the number of usage hours accumulated by the user for each cloud instance type by a total number of days in the cloud computing service billing cycle;
determining a per-day cost of each cloud instance type by multiplying the per-day utilization of each cloud instance type by the corresponding cloud instance rate; and
determining the per-day user workload run-rate for the user by combining the per-day cost of each cloud instance type used during the cloud computing service billing cycle.

6. The computer-implemented method of claim 1 , wherein said determining a cloud resource usage, which is expected for the user at the end of the cloud computing service billing cycle, comprises multiplying the per-day user workload run-rate for the user by a total number of days in the cloud computing service billing cycle.

7. The computer-implemented method of claim 6 , wherein said generating a trigger comprises generating a trigger to migrate the user's workload from the cloud-based resources to the local resources of the first user endpoint client device if the expected cloud resource usage is greater than a threshold value.

8. The computer-implemented method of claim 6 , wherein said generating a trigger comprises generating a trigger to migrate the user's workload from the cloud-based resources to the local resources of the first user endpoint client device if the expected cloud resource usage is greater than, less than or equal to a threshold value and the user persona is a lower tier user persona.

9. The computer-implemented method of claim 6 , wherein said generating a trigger comprises generating a trigger to migrate the workload from the local resources of the first user endpoint client device to the cloud-based resources based on one or more rules, which are used by the computer-implemented method to reduce utilization of the local resources of the first user endpoint client device when the expected cloud resource usage is less than a threshold value.

10. A system, comprising:
an orchestration server communicatively coupled via a network to a plurality of user endpoint client devices and a plurality of cloud service providers, wherein each respective endpoint client device is utilized by a respective user and comprises local resources for running workloads locally on the respective endpoint client device, and wherein each cloud service provider comprises cloud-based resources for running workloads on remote servers operated by the cloud service provider; and
an orchestration service stored within a computer readable storage medium and executed by a processing device of the orchestration server, wherein during a cloud computing service billing cycle the orchestration service is executed by the processing device to:
obtain telemetry data from a client agent execu

Assignees

• Dell Products Lp

Inventors

• Boyapalle Anantha K
• Iyer Vivek V

Classifications

• G06Q30/04

  Billing or invoicing · CPC title

• G06F2209/5022

  Workload threshold · CPC title

• G06F2209/5019

  Workload prediction · CPC title

• G06F9/505 Primary

  considering the load · CPC title

• G06F9/4856 Primary

  resumption being on a different machine, e.g. task migration, virtual machine migration (G06F9/5088 takes precedence) · CPC title