Computing Node Management System and Method for Managing a Plurality of Computing Nodes

Patent metadata

Field	Value
Publication number	US-2024048612-A1
Application number	US-202318488687-A
Country	US
Kind code	A1
Filing date	Oct 17, 2023
Priority date	Apr 17, 2021
Publication date	Feb 8, 2024
Grant date	—

Abstract

This application provides a computing node management system and a method for managing a plurality of computing nodes, to reduce costs to be paid by a cloud vendor. The computing node management system in this application includes a central offload node and a plurality of computing nodes. A riser card is inserted into each computing node. Each riser card is connected to the central offload node. Each riser card establishes a communication channel between the central offload node and the computing node into which the riser card is inserted. The central offload node is configured to provide resources for the plurality of computing nodes through the communication channels.

First claim

What is claimed is:

1 . A computing node management system, wherein the system comprises
a central offload node and a plurality of computing nodes, a riser card is inserted into each computing node, each riser card is connected to the central offload node, each riser card establishes a communication channel between the central offload node and the computing node into which the riser card is inserted, and the central offload node is configured to provide resources for the plurality of computing nodes through the communication channels.

2 . The system according to claim 1 , wherein the communication channel comprises an inter-node communication sub-channel and an intra-node communication sub-channel, the inter-node communication sub-channel is provided between the riser card and the central offload node, and the intra-node communication sub-channel is provided between the riser card and the computing node into which the riser card is inserted.

3 . The system according to claim 2 , wherein the resource comprises a storage resource; the central offload node is connected to a storage device that provides the storage resource; a first virtual device is deployed on a first riser card in the plurality of riser cards; a first cloud instance in a first computing node into which the first riser card is inserted is bound to the first virtual device; the central offload node allocates, to the first virtual device through a first inter-node communication sub-channel, the storage resource of the storage device connected to the central offload node; and the first cloud instance accesses, through a first intra-node communication sub-channel, the first virtual device to which the storage resource is allocated.

4 . The system according to claim 3 , wherein the central offload node is further configured to forward an input/output IO request between the storage device and the first cloud instance.

5 . The system according to claim 4 , wherein the central offload node is further configured to perform processing on the IO request, and the processing comprises at least one of the following: data encryption and decryption, data verification, data transmission speed control, and data quality-of-service control.

6 . The system according to claim 2 , wherein the resource comprises a computing resource; the central offload node is connected to a computing device that provides the computing resource; a second virtual device is deployed on a second riser card in the plurality of riser cards; a second cloud instance in a second computing node into which the second riser card is inserted is bound to the second virtual device; the central offload node allocates, to the second virtual device through a second inter-node communication sub-channel, the computing resource of the computing device connected to the central offload node; and the second cloud instance accesses, through a second intra-node communication sub-channel, the second virtual device to which the computing resource is allocated.

7 . The system according to claim 6 , wherein the computing device is a processor configured for heterogeneous computing.

8 . The system according to claim 2 , wherein the resource comprises a network resource; the central offload node is connected to a network device that provides the network resource; a third virtual device is deployed on a third riser card in the plurality of riser cards; a third cloud instance in a third computing node into which the third riser card is inserted is bound to the third virtual device; the central offload node allocates, to the third virtual device through a third inter-node communication sub-channel, the network resource of the network device connected to the central offload node; and the third cloud instance accesses, through a third intra-node communication sub-channel, the third virtual device to which the network resource is allocated.

9 . The system according to claim 8 , wherein the central offload node is further configured to forward a network packet between the network device and the third cloud instance.

10 . The system according to claim 8 , wherein the central offload node is further configured to forward network packets between cloud instances on the plurality of computing nodes.

11 . A method for managing a plurality of computing nodes, wherein the method is applied to a central offload node, a riser card is inserted into each of the plurality of computing nodes, each riser card is connected to the central offload node, each riser card establishes a communication channel between the central offload node and the computing node into which the riser card is inserted, and the method comprises:
providing, by the central offload node, resources for the plurality of computing nodes through the communication channels; and separately managing, by the central offload node, the plurality of computing nodes through the communication channels.

12 . The method according to claim 11 , wherein the communication channel comprises an inter-node communication sub-channel and an intra-node communication sub-channel, the inter-node communication sub-channel is provided between the riser card and the central offload node, and the intra-node communication sub-channel is provided between the riser card and the computing node into which the riser card is inserted.

13 . The method according to claim 12 , wherein the resource comprises a storage resource; the central offload node is connected to a storage device that provides the storage resource; a first virtual device is deployed on a first riser card in the plurality of riser cards; a first computing node into which the first riser card is inserted or a first cloud instance running on the first computing node is bound to the first virtual device through a first intra-node communication sub-channel; and the providing, by the central offload node, resources for the plurality of computing nodes through the communication channels comprises: allocating, by the central offload node to the first virtual device through a first inter-node communication sub-channel, the storage resource of the storage device connected to the central offload node.

14 . The method according to claim 12 , wherein the resource comprises a computing resource; the central offload node is connected to a computing device that provides the computing resource; a second virtual device is deployed on a second riser card in the plurality of riser cards; a second computing node into which the second riser card is inserted or a second cloud instance running on the second computing node is bound to the second virtual device through a second intra-node communication sub-channel; and the providing, by the central offload node, resources for the plurality of computing nodes through the communication channels comprises: allocating, by the central offload node to the second virtual device through a second inter-node communication sub-channel, the computing resource of the computing device connected to the central offload node.

15 . The method according to claim 12 , wherein the resource comprises a network resource; the central offload node is connected to a network device that provides the network resource; a third virtual device is deployed on a third riser card in the plurality of riser cards; a third computing node into which the third riser card is inserted or a third cloud instance running on the third computing node is bound to the third virtual device through a third intra-node communication sub-channel; and the providing, by the central offload node, resources for the plurality of computing nodes through the commu

Assignees

• Huawei Cloud Computing Tech Co Ltd

Inventors

• Ye Chuan
• Liu Jinsong

Classifications

• G06F9/54

  Interprogram communication · CPC title

• G06F9/5072

  Grid computing · CPC title

• G06F13/4022

  using switching circuits, e.g. switching matrix, connection or expansion network (G06F13/4009 takes precedence) · 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

• H04L67/1031 Primary

  Controlling of the operation of servers by a load balancer, e.g. adding or removing servers that serve requests · CPC title