DISTRIBUTED STORAGE RESOURCE MANAGEMENT IN A HYPER CONVERGED INFRASTRUCTURE

• US 20180248949A1
• Filed: 05/05/2017
• Published: 08/30/2018
• Est. Priority Date: 02/25/2017
• Status: Active Grant

• Current Assignee: Vmware LLC (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: SIDDAPPA, GIREESHA UDAGANI, NAYAK, SAMDEEP
• Granted Patent: US 10,462,221 B2
• Time in Patent Office: Days
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1012 based on compliance of requirements or conditions with available server resources
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

A distributed storage resource management in a hyper converged infrastructure is created. The system accesses a virtualized infrastructure comprising a plurality of host machines, each of said plurality of host machines having resources, the resources comprising: a central processing unit (CPU), a memory, and a storage, each of the plurality of host machines communicatively coupled with one or more virtual machines. The resources in each of the host machines are disaggregated. The resources are aggregated based on a nature of the resource into a common pool of shared resources. A request is received for a workload application having a resource requirement. The resource requirement of the workload application is matched with resources in the common pool of shared resources. The matched resources in the common pool of shared resources are assigned to the workload application.

Claims

1. A method for distributed resource management in a hyper converged infrastructure, the method comprising:

accessing a virtualized infrastructure comprising;

a plurality of host machines, each of said plurality of host machines having resources, the resources comprising;

a central processing unit (CPU), a memory, and a storage, each of the plurality of host machines communicatively coupled with one or more virtual machines;

disaggregating the resources in each of the host machines;

aggregating the resources based on a nature of the resource into a common pool of shared resources;

receiving a request for a workload application having a resource requirement;

matching the resource requirement of the workload application with resources in the common pool of shared resources; and

assigning the matched resources in the common pool of shared resources to the workload application.

2. The method as recited by claim 1, wherein the distributed resource management is a distributed storage resource management and the common pool of shared resources is a common pool of shared storage resources.

3. The method as recited by claim 1, wherein the distributed resource management is a distributed memory resource management and the common pool of shared resources is a common pool of shared memory resources.

4. The method as recited by claim 1, wherein the distributed resource management is a distributed CPU resource management and the common pool of shared resources is a common pool of shared CPU resources.

5. The method as recited by claim 1, further comprising:

spreading the matching of the resource requirement of the workload application across resources on multiple host machines to reduce loss due to a host machine loss.

6. The method as recited by claim 1, wherein the plurality of host machines are communicatively coupled via a host bus adapter, the method further comprising:

utilizing a four queue depth capacity value for the host bust adapter when the workload application has a low performance requirement.

7. The method as recited by claim 1, wherein the plurality of host machines are communicatively coupled via a host bus adapter, the method further comprising:

utilizing a twelve queue depth capacity value for the host bust adapter when the workload application has a high performance requirement.

8. The method as recited by claim 1, wherein disaggregating the resources in each of the host machines comprises:

discovering the resources available across a rack of host machines,

the discovering performed by a virtual resource manager (VRM) acting as an out-of-band management agent.

9. The method as recited by claim 8, wherein discovering the resources available across a rack of host machines further comprises:

discovering a capacity of said resources;

discovering a usage of said resources;

discovering a power state of said resources; and

discovering a location of the host machines maintaining the resources.

10. A non-transitory computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by one or more processors of a device, cause the one or more processors to;

access a virtualized infrastructure comprising;

a plurality of host machines, each of said plurality of host machines having resources, the resources comprising;

a central processing unit (CPU), a memory, and a storage, each of the plurality of host machines communicatively coupled with one or more virtual machines;

disaggregate the resources in each of the host machines;

aggregate the resources based on a nature of the resource into a common pool of shared resources;

receive a request for a workload application having a resource requirement;

match the resource requirement of the workload application with resources in the common pool of shared resources; and

assign the matched resources in the common pool of shared resources to the workload application.

11. The non-transitory computer-readable medium of claim 10, further comprising one or more instructions that, when executed by one or more processors, cause the one or more processors to:

spread the matching of the resource requirement of the workload application across resources on multiple host machines to reduce loss due to a host machine loss.

12. The non-transitory computer-readable medium of claim 10, wherein the plurality of host machines are communicatively coupled via a host bus adapter, the method further comprising one or more instructions that, when executed by one or more processors, cause the one or more processors to:

utilize a four queue depth capacity value for the host bust adapter when the workload application has a low performance requirement.

13. The non-transitory computer-readable medium of claim 10, wherein the plurality of host machines are communicatively coupled via a host bus adapter, the method further comprising one or more instructions that, when executed by one or more processors, cause the one or more processors to:

utilizing a twelve queue depth capacity value for the host bust adapter when the workload application has a high performance requirement.

14. The non-transitory computer-readable medium of claim 10, further comprising one or more instructions that, when executed by one or more processors, cause the one or more processors to:

discover the resources available across a rack of host machines,

the discovering performed by a virtual resource manager (VRM) acting as an out-of-band management agent.

15. The non-transitory computer-readable medium of claim 10, wherein the common pool of shared resources is selected from two or more of the group consisting of:

a common pool of shared storage resources,

a common pool of shared memory resources, and

a common pool of shared CPU resources.

16. A non-transitory computer readable storage medium having computer readable program code embedded therein that when executed causes a computing system to perform a method for distributed resource management in a hyper converged infrastructure, the method comprising:

accessing a virtualized infrastructure comprising;

a plurality of host machines, each of said plurality of host machines having resources, the resources comprising;

a central processing unit (CPU), a memory, and a storage, each of the plurality of host machines communicatively coupled with one or more virtual machines;

disaggregating the resources in each of the host machines;

aggregating the resources based on a nature of the resource into a plurality of common pools of shared resources;

receiving a request for a workload application having a resource requirement;

matching the resource requirement of the workload application with resources in the plurality of common pools of shared resources; and

assigning the matched resources in the plurality of common pools of shared resources to the workload application.

17. The non-transitory computer readable storage medium as recited by claim 16, further comprising:

spreading the matching of the resource requirement of the workload application across resources on multiple host machines to reduce loss due to a host machine loss.

18. The non-transitory computer readable storage medium as recited by claim 16, wherein the plurality of host machines are communicatively coupled via a host bus adapter, the method further comprising:

utilizing a four queue depth capacity value for the host bust adapter when the workload application has a low performance requirement; and

utilizing a twelve queue depth capacity value for the host bust adapter when the workload application has a high performance requirement.

19. The non-transitory computer readable storage medium as recited by claim 16, wherein disaggregating the resources in each of the host machines comprises:

discovering the resources available across a rack of host machines,

the discovering performed by a virtual resource manager (VRM) acting as an out-of-band management agent.

20. The non-transitory computer readable storage medium as recited by claim 16, wherein the common pools of shared resources are selected from two or more of the group consisting of:

a common pool of shared storage resources,

a common pool of shared memory resources, and

a common pool of shared CPU resources.