IO bandwidth control method, IO access request processing method, apparatus, and system

Patent metadata

Field	Value
Publication number	US-10887248-B2
Application number	US-201715799196-A
Country	US
Kind code	B2
Filing date	Oct 31, 2017
Priority date	Jul 12, 2016
Publication date	Jan 5, 2021
Grant date	Jan 5, 2021

Abstract

An input output (IO) bandwidth control method, an IO access request processing method, an apparatus, and a system relate to the field of storage technologies, where the IO bandwidth control method, executed by a name node, includes determining an IO bandwidth of each data node in at least one data node and an IO bandwidth of a first tenant, and instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant based on the IO bandwidth of each data node and the IO bandwidth of the first tenant, where the at least one IO bandwidth is in a one-to-one correspondence with the at least one data node, and each IO bandwidth in the at least one IO bandwidth is greater than 0 and is less than or equal to an IO bandwidth of a corresponding data node.

First claim

What is claimed is:

1. An input output (IO) bandwidth control method implemented by a name node in a distributed file system, wherein the IO bandwidth control method comprises:
determining an IO bandwidth of at least one data node in the distributed file system;
determining an IO bandwidth of a first tenant according to a weight of each tenant in a plurality of tenants and a sum of the IO bandwidth of each data node in the at least one data node, wherein the IO bandwidth of the first tenant is shared by a plurality of users, wherein each user of the plurality of users corresponds to a plurality of applications, wherein the applications corresponding to the users that belong to the first tenant share the IO bandwidth of the first tenant, and wherein the weight of each tenant is based on a service priority of the tenant and an IO bandwidth required by a service of the tenant; and
instructing, based on the IO bandwidth of each data node in the at least one data node and the IO bandwidth of the first tenant, the at least one data node to allocate at least one IO bandwidth to the first tenant,
wherein the at least one IO bandwidth is in a one-to-one correspondence with the at least one data node,
wherein each IO bandwidth in the at least one IO bandwidth is less than or equal to an IO bandwidth of a corresponding data node, and
wherein each IO bandwidth in the at least one IO bandwidth is greater than zero.

2. The IO bandwidth control method according to claim 1 , wherein the at least one data node comprises a first data node, wherein an IO bandwidth allocated by the first data node to the first tenant according to an instruction of the name node is a first IO bandwidth, wherein the first IO bandwidth is an IO bandwidth in the at least one IO bandwidth corresponding to the first data node, and wherein after instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant, the IO bandwidth control method further comprises:
determining that an IO bandwidth that is in the at least one IO bandwidth and that is used by the first tenant is at least one utilized bandwidth, wherein the at least one utilized bandwidth is in a one-to-one correspondence with the at least one IO bandwidth; and
either instructing the at least one data node to adjust a corresponding IO bandwidth in the at least one IO bandwidth, or instructing the first data node to adjust the first IO bandwidth according to the at least one utilized bandwidth.

3. The IO bandwidth control method according to claim 1 , wherein after instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant, the IO bandwidth control method further comprises:
determining that an IO bandwidth that is in the at least one IO bandwidth and that is used by the first tenant is at least one utilized bandwidth, wherein the at least one utilized bandwidth is in a one-to-one correspondence with the at least one IO bandwidth; and
instructing a second data node to allocate an IO bandwidth to the first tenant when a difference between a sum of the at least one IO bandwidth and a sum of the at least one utilized bandwidth is less than or equal to a first threshold, wherein the at least one data node does not comprise the second data node.

4. The IO bandwidth control method according to claim 1 , wherein after instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant, the IO bandwidth control method further comprises:
determining that an IO bandwidth that is in the at least one IO bandwidth and that is used by the first tenant is at least one utilized bandwidth, wherein the at least one utilized bandwidth is in a one-to-one correspondence with the at least one IO bandwidth; and
instructing a third data node to adjust an IO bandwidth allocated by the third data node to the first tenant to zero when a difference between a sum of the at least one IO bandwidth and a sum of the at least one utilized bandwidth is greater than a first threshold, wherein the at least one data node comprises the third data node.

5. The IO bandwidth control method according to claim 1 , wherein the at least one data node comprises a first data node, wherein an IO bandwidth allocated by the first data node to the first tenant according to an instruction of the name node is a first IO bandwidth, wherein the first IO bandwidth is an IO bandwidth in the at least one IO bandwidth corresponding to the first data node, and wherein after instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant, the IO bandwidth control method further comprises:
determining an IO bandwidth that is in the first IO bandwidth and that is used by the first tenant; and
instructing, according to the IO bandwidth that is in the first IO bandwidth and that is used by the first tenant, the first data node to adjust the first IO bandwidth.

6. A name node in a distributed file system, wherein the name node comprises:
a processor; and
a memory coupled to the processor and configured to store instructions, wherein the instructions cause the processor to be configured to:
determine an input output (IO) bandwidth of at least one data node in the distributed file system;
determine an IO bandwidth of a first tenant according to a weight of each tenant in a plurality of tenants and a sum of the IO bandwidth of each data node in the at least one data node, wherein the IO bandwidth of the first tenant is shared by a plurality of users, wherein each user of the plurality of users corresponds to a plurality of applications, wherein the applications corresponding to the users that belong to the first tenant share the IO bandwidth of the first tenant, wherein the weight of each tenant is based on a service priority of the tenant and an IO bandwidth required by a service of the tenant; and
instruct, based on the IO bandwidth of each data node in the at least one data node and the IO bandwidth of the first tenant, the at least one data node to allocate at least one IO bandwidth to the first tenant,
wherein the at least one IO bandwidth is in a one-to-one correspondence with the at least one data node,
wherein each IO bandwidth in the at least one IO bandwidth is less than or equal to an IO bandwidth of a corresponding data node, and
wherein each IO bandwidth in the at least one IO bandwidth is greater than 0.

7. The name node according to claim 6 , wherein the at least one data node comprises a first data node, wherein an IO bandwidth allocated by the first data node to the first tenant according to an instruction of the name node is a first IO bandwidth, wherein the first IO bandwidth is an IO bandwidth in the at least one IO bandwidth corresponding to the first data node, and wherein the instructions further cause the processor to be configured to:
determine that an IO bandwidth that is in the at least one IO bandwidth and that is used by the first tenant is at least one utilized bandwidth after instructing the at least one data node to allocate the at least one IO bandwidth to the first tenant, wherein the at least one utilized bandwidth is in a one-to-one correspondence with the at least one IO bandwidth; and
either instruct the at least one data node to adjust a corresponding IO bandwidth in the at least one IO bandwidth, or instruct the first data node to adjust the first IO bandwidth according to the at least one utilized bandwidth.

8. The name node according to claim 6 , wherein the instructions further cause the processor to be configured to:
determine that an IO bandwidth that is in the at least one IO bandwidth and that is used by the first tenant is at least one utilized bandwidth after instructing the at least one data node

Assignees

• Huawei Tech Co Ltd

Inventors

• Li Yi
• Li Jing
• Zhong Yong

Classifications

• H04L47/783 Primary

  Distributed allocation of resources, e.g. bandwidth brokers · CPC title

• H04L43/0876 Primary

  Network utilisation, e.g. volume of load or congestion level · CPC title

• H04L67/60

  Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources (admission control or resource allocation H04L47/70) · CPC title

• G06F9/5011

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title

• H04L43/16

  Threshold monitoring · CPC title