Resource management method, host, and endpoint

Patent metadata

Field	Value
Publication number	US-2017199767-A1
Application number	US-201715471202-A
Country	US
Kind code	A1
Filing date	Mar 28, 2017
Priority date	Dec 8, 2014
Publication date	Jul 13, 2017
Grant date	—

Abstract

Embodiments provide a resource management technology that may be applied to a host, where the host includes a CPU, an endpoint connected to the CPU, and an I/O device connected to the endpoint. A method includes: allocating, by the CPU, a target endpoint to a target process, where a virtual device is disposed on the target endpoint; obtaining, by the target endpoint, a performance specification of the target process, and adjusting a performance parameter of the virtual device according to the performance specification, where the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint; and when the target process needs to access a resource, obtaining, from the I/O device, a resource that satisfies the performance specification of the target process, and providing the obtained resource to the target process for use.

First claim

What is claimed is:

1 . A resource management method, applied to a host, wherein the host comprises a Central Processing Unit (CPU) and an endpoint in communication with the CPU, the endpoint is connected to an Input/Output (I/O) device, the method comprising:
allocating, by the CPU, a target endpoint to a target process, wherein a virtual device is disposed on the target endpoint; obtaining, by the target endpoint, a performance specification of the target process, and adjusting a performance parameter of the virtual device according to the performance specification, wherein the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint; and when the target process needs to access a resource, sending, by the CPU, a resource access request, of the target process, for the virtual device to the target endpoint; and after receiving the resource access request, obtaining, by the target endpoint from the I/O device, a resource that satisfies the performance specification of the target process, and providing the obtained resource to the target process for use.

2 . The resource management method according to claim 1 , wherein the obtaining, by the target endpoint, a performance specification of the target process comprises:
obtaining, by the target endpoint, the performance specification from the CPU.

3 . The resource management method according to claim 1 , wherein the obtaining, by the target endpoint, a performance specification of the target process comprises:
collecting, by the target endpoint, statistics about performance of a resource occupied when the target process runs in history, and generating the performance specification according to a statistical result.

4 . The resource management method according to claim 1 , wherein:
the target process runs in a virtual machine.

5 . The resource management method according to claim 1 , wherein resources of multiple I/O devices from multiple endpoints form an I/O resource pool together, or resources of multiple I/O devices from a same endpoint form an I/O resource pool together, and the obtaining, from the I/O device, a resource that satisfies the performance specification of the target process comprises:
obtaining, from the resource pool, the resource that satisfies the performance specification of the target process.

6 . The resource management method according to claim 1 , wherein the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint comprises:
for performance that can be shared by multiple processes, satisfying a maximum requirement for this performance parameter in all performance specifications of all processes that use the target endpoint; and for performance that cannot be shared by the multiple processes, satisfying a summation of requirement for performance parameters in all performance specifications of all processes that use the target endpoint.

7 . The resource management method according to claim 1 , wherein:
performance that can be shared by multiple processes comprises a latency time; and wherein performance that cannot be shared by the multiple processes comprises input/output per second IOPS and bandwidth.

8 . The resource management method according to claim 1 , wherein the method further comprises: migrating a process to second host, wherein the second host comprises a second CPU, a second endpoint, and a second I/O resource pool, and the migrating comprising:
sending, by the CPU, the performance specification of the process to the second CPU; sending, by the another CPU, the performance specification to the second endpoint; generating, by the second endpoint, a second virtual device, wherein the second virtual device satisfies the performance specification; sending, by the CPU, description information of the process to the second CPU, and generating, by the second CPU, a second process according to the description information; and sending, by the second CPU, a second resource access request, of the second process, for the second virtual device to the second endpoint when the second process needs to access a resource; and after receiving the second resource access request, obtaining, by the second endpoint, a resource from the second I/O resource pool according to the performance specification, and providing the resource to the second process for use.

9 . A host comprising:
a Central Processing Unit (CPU); an endpoint in communication with the CPU and an Input/Output (I/O) resource; a target endpoint in communication with the CPU and an I/O device; and a memory storage comprising instructions in communication with the CPU, wherein the CPU executes the instructions to:
allocate the target endpoint to a process;
send a resource access request, of a target process, for a virtual device to the target endpoint when the target process needs to access a resource; and
wherein the target endpoint performs operations to:
receive the resource access request;
generate the virtual device;
obtain a performance specification of the target process;
adjust a performance parameter of the virtual device according to the performance specification, wherein the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint; and
obtain, after receiving the resource access request, an I/O resource from the I/O device according to the performance specification of the target process, and provide the resource to the target process for use.

10 . The host according to claim 9 , wherein the target endpoint further performs operations to obtain the performance specification from the CPU.

11 . The host according to claim 9 , wherein the target endpoint further performs operations to: collect statistics about performance of a resource occupied when the target process runs in history, and generate the performance specification for the endpoint according to a statistical result.

12 . The host according to claim 9 , wherein:
the target process runs in a virtual machine, and the virtual machine runs in a hypervisor of the CPU.

13 . The host according to claim 9 , wherein:
the I/O device is in communication with at least two endpoints, generates an I/O resource pool together, and provides a resource by using the I/O resource pool.

14 . The host according to claim 9 , wherein
resources of multiple I/O devices form an I/O resource pool together; and wherein the I/O device in communication with one endpoint, and the I/O device provides resources to form the I/O resource pool, and provides a resource by using the I/O resource pool.

15 . The host according to claim 9 , wherein that the adjusted virtual device satisfies a total requirement of performance specifications of all processes that use the target endpoint wherein the target endpoint further performs operations to:
for performance that can be shared by multiple processes, satisfy a maximum requirement for this performance parameter in all performance specifications of all processes that use the target endpoint; and for performance that cannot be shared by the multiple processes, satisfy a total summation of requirement for performance parameters in all performance specifications of all processes that use the target endpoint.

16 . The host according to claim 9 , wherein the host and second host constitute a resource management system, and the second host comprises second CPU, second endpoint, and second I/O resource pool, w

Assignees

• Huawei Tech Co Ltd

Inventors

• Yu Baifeng
• Gu Jiongjiong
• Lin Muhui
• Yu Zhou
• Mao Lingzhi

Classifications

• G06F11/3017

  where the computing system is implementing multitasking (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• G06F11/3442

  for planning or managing the needed capacity · CPC title

• G06F11/3452

  Performance evaluation by statistical analysis · CPC title

• G06F9/5044 Primary

  considering hardware capabilities · 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