Implementing dynamic adjustment of resources allocated to SRIOV remote direct memory access adapter (RDMA) virtual functions based on usage patterns

Patent metadata

Field	Value
Publication number	US-9535740-B1
Application number	US-201514836309-A
Country	US
Kind code	B1
Filing date	Aug 26, 2015
Priority date	Aug 26, 2015
Publication date	Jan 3, 2017
Grant date	Jan 3, 2017

Abstract

A method, system and computer program product are provided for implementing dynamic adjustment of resources allocated to Single Root Input/Output Virtualization (SRIOV) Remote Direct Memory Access (RMDA) virtual functions (VFs) in Cloud Software Defined Server environments. A hardware management console (HMC), and a hypervisor are used to implement resource allocation to the SRIOV RDMA VFs based on resource usage. The hypervisor checks resource usage for the resource allocations of the SRIOV RDMA VF relative to lower and upper threshold values. Responsive to identifying the resource usage below the lower threshold or above the upper threshold, the hypervisor sends an event to the HMC, and the HMC starts a resource redistribution process.

First claim

What is claimed is:

1. A method for implementing dynamic adjustment of resources allocated to Single Root Input/Output Virtualization (SRIOV) Remote Direct Memory Access (RMDA) virtual functions (VFs) in Cloud Software Defined Server environments comprising:
using a hardware management console (HMC) and a hypervisor to implement resource allocation
said hypervisor, checking resource usage for the resource allocations of a Single Root Input/Output Virtualization (SRIOV) Remote Direct Memory Access (RMDA) VF relative to lower and upper threshold values;
responsive to identifying the resource usage below the lower threshold or above the upper threshold for the RMDA VF, said hypervisor sending an event to the HMC,
said HMC starting a resource redistribution process including said HMC querying usage statistics from said hypervisor, responsive to receiving said event; for each active VF, allocating based on priority from an adapter unallocated resource pool, and when needed, allocating based on priority from an adapter allocated resource pool, updating resources used attribute for the RMDA VF, and sending updated limits to said hypervisor;
said HMC responsive to usage statistics from said hypervisor, making request to said hypervisor to allocate and deallocate VF resources; and
said hypervisor, making resource changes in the SRIOV RDMA adapter and system, and notifying said HMC, and a logical partition (LPAR) owning the SRIOV RMDA VF of the change.

2. The method as recited in claim 1 , includes said HMC allocating additional resources from free resources to a particular VF requiring more resources.

3. The method as recited in claim 1 , includes said hypervisor collecting VF statistics and determining a VF having more traffic based on the collected VF statistics; and said hypervisor notifying said HMC.

4. The method as recited in claim 3 , includes said hypervisor dynamically allocating and deallocating resources to and from a VF.

5. The method as recited in claim 1 , includes said HMC deallocating excess resources from a VF and returning the excess resources to free resource pool responsive to resources used by a particular VF being lower than a predefined threshold.

6. The method as recited in claim 1 , includes said hypervisor maintaining usage of resources by a Virtual Function (VF) using a Bitmap.

7. The method as recited in claim 1 , includes said hypervisor using a plurality of resource pools for maintaining usage of resources.

8. The method as recited in claim 1 , includes said hypervisor maintaining usage of resources using an adapter allocated used resource pool of resources allocated and actively being used by VFs; an adapter allocated unused resource pool of resources allocated and not being used by VFs; and an adapter unallocated resource pool of all unallocated resources.

9. The method as recited in claim 1 , includes said HMC responsive to a reallocation resource request, allocating resources for each active VF starting with a highest priority VF.

10. The method as recited in claim 1 , includes providing an intelligent software provisioning layer monitoring events from said hypervisor; said intelligent software provisioning layer dynamically reducing unused resources for the VFs based on thresholds and idle periods for resources of a predefined amount of time for which there is no activity.

11. A system for implementing dynamic adjustment of resources allocated to Single Root Input/Output Virtualization (SRIOV) Remote Direct Memory Access (RMDA) virtual functions (VFs) in Cloud Software Defined Server environments comprising:
a Single Root Input/Output Virtualization (SRIOV) Remote Direct Memory Access (RMDA) adapter comprising a plurality of virtual functions (VFs),
a hardware management console (HMC) and a hypervisor;
a processor using said hardware management console (HMC) and said hypervisor to implement resource allocation;
said hypervisor, checking resource usage for the resource allocations of a SRIOV Remote Direct Memory Access (RMDA) VF relative to lower and upper threshold values;
responsive to identifying the resource usage below the lower threshold or above the upper threshold, said hypervisor sending an event to the HMC,
said HMC starting a resource redistribution process including said HMC querying usage statistics from said hypervisor, responsive to receiving said event; for each active VF, allocating based on priority from an adapter unallocated resource pool, and when needed, allocating based on priority from an adapter allocated resource pool, updating resources used attribute for the RMDA VF, and sending updated limits to said hypervisor;
said HMC responsive to usage statistics from said hypervisor, making a request to said hypervisor to allocate and deallocate VF resources; and
said hypervisor, making resource changes in the SRIOV RDMA adapter and system, and notifying said HMC, and a logical partition (LPAR) owning the SRIOV RMDA VF of the change.

12. The system as recited in claim 11 , wherein said management function and said hypervisor include control code tangibly embodied in a non-transitory machine readable medium used for implementing dynamic resource allocation.

13. The system as recited in claim 11 , includes said HMC allocating additional resources from free resources to a particular VF requiring more resources.

14. The system as recited in claim 11 , includes said HMC deallocating additional resources from a particular VF to free resources.

15. The system as recited in claim 11 , includes said hypervisor dynamically allocating and deallocating resources to and from a VF.

16. The system as recited in claim 11 , includes said HMC responsive to a reallocation resource request, allocating resources for each active VF starting with a highest priority VF.

17. The system as recited in claim 11 , includes said hypervisor maintaining usage of resources by a Virtual Function (VF) using a Bitmap.

18. The system as recited in claim 11 , includes said hypervisor maintaining usage of resources using an adapter allocated used resource pool of resources allocated and actively being used by VFs; an adapter allocated unused resource pool of resources allocated and not being used by VFs; and an adapter unallocated resource pool of all unallocated resources.

19. The system as recited in claim 11 , includes providing an intelligent software provisioning layer monitoring events from said hypervisor; said intelligent software provisioning layer sending requests to said HMC for dynamically reducing unused resources for the VFs based on thresholds and idle periods for resources of a predefined amount of time for which there is no activity.

20. The system as recited in claim 11 , includes said hypervisor collecting VF statistics and determining a VF having more traffic based on the collected VF statistics; and said hypervisor notifying said HMC.

Assignees

• IBM

Inventors

• Graham Charles S
• Kommineni Prathima
• Schimke Timothy J
• Challa Narsimha R

Classifications

• G06F2009/45583

  Memory management, e.g. access or allocation · CPC title

• G06F9/45558 Primary

  Hypervisor-specific management and integration aspects · 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

• G06F2009/45591

  Monitoring or debugging support · CPC title

• G06F2009/4557

  Distribution of virtual machine instances; Migration and load balancing · CPC title