[go: up one dir, main page]

CN115622938A - Computing power service processing method, device and system - Google Patents

Computing power service processing method, device and system Download PDF

Info

Publication number
CN115622938A
CN115622938A CN202211349514.9A CN202211349514A CN115622938A CN 115622938 A CN115622938 A CN 115622938A CN 202211349514 A CN202211349514 A CN 202211349514A CN 115622938 A CN115622938 A CN 115622938A
Authority
CN
China
Prior art keywords
computing power
node
computing
service
multicast
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202211349514.9A
Other languages
Chinese (zh)
Other versions
CN115622938B (en
Inventor
李晶林
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
New H3C Technologies Co Ltd
Original Assignee
New H3C Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by New H3C Technologies Co Ltd filed Critical New H3C Technologies Co Ltd
Priority to CN202211349514.9A priority Critical patent/CN115622938B/en
Publication of CN115622938A publication Critical patent/CN115622938A/en
Application granted granted Critical
Publication of CN115622938B publication Critical patent/CN115622938B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/32Flooding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/42Centralised routing
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004Server selection for load balancing

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明实施例提供了一种算力服务处理方法、装置及系统,涉及通信技术领域,该系统包括至少一个算力服务节点、至少一个算力边缘节点以及至少一个路由节点;每个算力边缘节点,用于接收终端设备发送的算力服务请求报文,并向自身作为组播源的组播组中的算力服务节点发送组播查询报文;每个算力服务节点,用于响应于来自目标算力边缘节点的目标组播查询报文,向目标算力边缘节点发送查询响应报文,查询响应报文中包括该算力服务节点连接的算力服务器的算力资源信息;每个算力边缘节点,还用于基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向目标算力服务器转发算力服务请求报文。可以简化部署过程,降低部署成本。

Figure 202211349514

Embodiments of the present invention provide a computing power service processing method, device, and system, which relate to the field of communication technology. The system includes at least one computing power service node, at least one computing power edge node, and at least one routing node; each computing power edge node, used to receive the computing power service request message sent by the terminal device, and send a multicast query message to the computing power service node in the multicast group that serves as the multicast source; each computing power service node is used to respond Based on the target multicast query message from the target computing power edge node, send a query response message to the target computing power edge node, and the query response message includes the computing power resource information of the computing power server connected to the computing power service node; A computing power edge node is also used to select a target computing power server based on the received computing power resource information of each computing power server, and forward the computing power service request message to the target computing power server. It can simplify the deployment process and reduce the deployment cost.

Figure 202211349514

Description

算力服务处理方法、装置及系统Computing power service processing method, device and system

技术领域technical field

本发明涉及通信技术领域,特别是涉及一种算力服务处理方法、装置及系统。The present invention relates to the field of communication technologies, in particular to a computing power service processing method, device and system.

背景技术Background technique

目前部署算力网络的主要需求为,基于算力网络中各个能够提供算力服务的算力节点的算力资源信息,将用户的业务调度到最合适的算力节点中。因此,算力网络的控制面需要动态感知算力网络中的算力资源信息,从而确定各算力服务请求报文的转发路径。At present, the main requirement for deploying a computing power network is to dispatch the user's business to the most suitable computing power node based on the computing power resource information of each computing power node in the computing power network that can provide computing power services. Therefore, the control plane of the computing power network needs to dynamically perceive the computing power resource information in the computing power network, so as to determine the forwarding path of each computing power service request message.

根据算力资源信息的收集、转发路径的编排和转发路径的分发机制的不同,算力网络的控制面的部署方式可以分为集中式、分布式和混合式三种。在集中式算力网络中,控制器可以按照应用需求,结合全网的算力资源信息和网络状态编排转发路径。在分布式算力网络中,各节点需通过分布式路由协议对本地的算力资源、网络状态和路由进行泛洪通告,进而分布式算力网络中的路由设备可以基于学习到的全网的算力资源计算转发路径。混合式算力网络的实现方式是集中式算力网络和分布式算力网络实现方式的结合。According to the collection of computing resource information, the arrangement of forwarding paths, and the distribution mechanisms of forwarding paths, the deployment methods of the control plane of the computing power network can be divided into three types: centralized, distributed, and hybrid. In a centralized computing power network, the controller can arrange the forwarding path according to the application requirements, combining the computing power resource information and network status of the entire network. In the distributed computing power network, each node needs to flood the local computing power resources, network status and routing through the distributed routing protocol, and then the routing devices in the distributed computing power network can Computing resources calculate the forwarding path. The implementation of the hybrid computing power network is a combination of the centralized computing power network and the distributed computing power network.

采用集中式的部署方式对控制器的要求较高,开发工作量和成本较高。若采用分布式的部署方式,则需将全网的设备均调整为具有泛洪通告算力资源能力的设备,需要对网络中已有的设备进行大幅调整。混合式的部署方式能一定程度缓解上述问题,但仍存在部署复杂的问题。The centralized deployment method has higher requirements on the controller, and the development workload and cost are higher. If the distributed deployment method is adopted, all the devices in the entire network need to be adjusted to devices capable of flooding and notifying computing power resources, and the existing devices in the network need to be greatly adjusted. The hybrid deployment method can alleviate the above problems to a certain extent, but there are still problems of complex deployment.

发明内容Contents of the invention

本发明实施例的目的在于提供一种算力服务处理方法、装置及系统,以解决算力服务处理系统部署复杂的问题。具体技术方案如下:The purpose of the embodiments of the present invention is to provide a computing power service processing method, device, and system to solve the problem of complex deployment of computing power service processing systems. The specific technical scheme is as follows:

第一方面,本申请实施例提供一种算力服务处理系统,所述系统包括至少一个算力服务节点、至少一个算力边缘节点以及至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;In the first aspect, an embodiment of the present application provides a computing power service processing system, the system includes at least one computing power service node, at least one computing power edge node, and at least one routing node, and each computing power service node is connected to at least one Computing server;

每个算力边缘节点,用于接收终端设备发送的算力服务请求报文,并向自身作为组播源的组播组中的算力服务节点发送组播查询报文,每个组播组对应的组播树中的叶子节点均为算力服务节点,非叶子节点为路由节点,根节点为算力边缘节点;Each computing power edge node is used to receive the computing power service request message sent by the terminal device, and send a multicast query message to the computing power service node in the multicast group that serves as the multicast source. Each multicast group The leaf nodes in the corresponding multicast tree are computing service nodes, the non-leaf nodes are routing nodes, and the root nodes are computing power edge nodes;

每个算力服务节点,用于响应于来自目标算力边缘节点的目标组播查询报文,向所述目标算力边缘节点发送查询响应报文,所述查询响应报文中包括该算力服务节点连接的算力服务器的算力资源信息;Each computing power service node is configured to send a query response message to the target computing power edge node in response to a target multicast query message from the target computing power edge node, and the query response message includes the computing power The computing resource information of the computing server connected to the service node;

每个算力边缘节点,还用于基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向所述目标算力服务器转发算力服务请求报文。Each computing power edge node is further configured to select a target computing power server based on the received computing power resource information of each computing power server, and forward the computing power service request message to the target computing power server.

可选的,所述算力服务请求报文用于请求指定服务类型的算力服务;Optionally, the computing power service request message is used to request a computing power service of a specified service type;

每个算力边缘节点,具体用于在接收到算力服务请求报文后,从自身作为组播源的多个组播组中,查找支持所述指定服务类型的指定组播组,向所述指定组播组中的算力服务节点发送组播查询报文。Each computing power edge node is specifically used to search for a specified multicast group that supports the specified service type from multiple multicast groups that it serves as a multicast source after receiving a computing power service request message, and send The computing power service node in the specified multicast group sends a multicast query message.

可选的,所述目标组播查询报文用于请求查询指定服务类型的算力资源信息;Optionally, the target multicast query message is used to request to query computing power resource information of a specified service type;

每个算力服务节点,具体用于响应于接收到的目标组播查询报文,从连接的算力服务器中,筛选出支持所述指定服务类型的指定算力服务器,向所述目标算力边缘节点发送所述查询响应报文,所述查询响应报文中包括所述指定算力服务器的算力资源信息。Each computing power service node is specifically used to respond to the received target multicast query message, select the designated computing power server that supports the specified service type from the connected computing power servers, and provide the target computing power The edge node sends the query response message, and the query response message includes the computing power resource information of the designated computing power server.

可选的,所述组播查询报文用于请求查询指定服务类型的算力资源信息;Optionally, the multicast query message is used to request to query computing power resource information of a specified service type;

每个算力服务节点,具体用于响应于接收到的目标组播查询报文,从连接的算力服务器中,筛选出支持所述指定服务类型且剩余算力资源满足预设条件的指定算力服务器,向所述目标算力边缘节点发送所述查询响应报文,所述查询响应报文中包括所述指定算力服务器的算力资源信息。Each computing power service node is specifically used to respond to the received target multicast query message, and select the designated computing power server that supports the specified service type and the remaining computing power resources meet the preset conditions from the connected computing power servers. The power server sends the query response message to the target computing power edge node, and the query response message includes the computing power resource information of the designated computing power server.

可选的,optional,

每个算力服务节点,还用于周期性获取与自身连接的各算力服务器的算力资源信息。Each computing power service node is also used to periodically obtain the computing power resource information of each computing power server connected to itself.

可选的,所述系统还包括汇聚节点;Optionally, the system also includes a sink node;

每个算力服务节点,还用于在接入到所述算力服务处理系统中时,向所述汇聚节点发送注册信息,所述注册信息包括该算力服务节点连接的算力服务器的服务类型;Each computing power service node is further configured to send registration information to the aggregation node when accessing the computing power service processing system, and the registration information includes the service of the computing power server connected to the computing power service node type;

所述汇聚节点,用于响应于来自目标算力服务节点的目标注册信息,基于所述目标注册信息携带的服务类型,将所述目标算力服务节点加入到与所述目标注册信息携带的服务类型匹配的目标组播组中,并更新所述目标组播组对应的组播树。The converging node is configured to add the target computing power service node to the service carried in the target computing power service node based on the service type carried in the target computing power service node in response to the target registration information from the target computing power service node. type matching target multicast group, and update the multicast tree corresponding to the target multicast group.

第二方面,本申请实施例提供一种算力服务处理方法,应用于算力服务处理系统中的任一算力边缘节点,所述算力服务处理系统中还包括至少一个算力服务节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;所述方法包括:In the second aspect, the embodiment of the present application provides a computing power service processing method, which is applied to any computing power edge node in the computing power service processing system, and the computing power service processing system further includes at least one computing power service node and At least one routing node, each computing power service node is connected to at least one computing power server; the method includes:

接收终端设备发送的算力服务请求报文;Receive the computing power service request message sent by the terminal device;

向自身作为组播源的组播组中的算力服务节点发送组播查询报文,所述组播组对应的组播树中的叶子节点均为算力服务节点,非叶子节点为路由节点,根节点为所述算力边缘节点;Send a multicast query message to the computing power service node in the multicast group that serves as the multicast source. The leaf nodes in the multicast tree corresponding to the multicast group are all computing power service nodes, and the non-leaf nodes are routing nodes. , the root node is the computing power edge node;

接收所述组播组中的算力服务节点发送的查询响应报文,所述查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息;receiving a query response message sent by a computing power service node in the multicast group, where the query response message includes computing power resource information of a computing power server connected to the computing power service node;

基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向所述目标算力服务器转发所述算力服务请求报文。Based on the received computing power resource information of each computing power server, a target computing power server is selected, and the computing power service request message is forwarded to the target computing power server.

可选的,所述向组播组中的算力服务节点发送组播查询报文,包括:Optionally, the sending the multicast query message to the computing power service node in the multicast group includes:

从多个组播组中,查找支持所述指定服务类型的指定组播组;Finding a specified multicast group that supports the specified service type from multiple multicast groups;

向所述指定组播组中的算力服务节点发送所述组播查询报文。Sending the multicast query message to the computing service nodes in the specified multicast group.

第三方面,本申请实施例提供一种算力服务处理方法,应用于算力服务处理系统中的任一算力服务节点,所述算力服务处理系统中还包括至少一个算力边缘节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;所述算力服务节点为组播组对应的组播树的叶子节点,所述组播树的非叶子节点为路由节点,根节点为算力边缘节点;所述方法包括:In the third aspect, the embodiment of the present application provides a computing power service processing method, which is applied to any computing power service node in the computing power service processing system, and the computing power service processing system further includes at least one computing power edge node and At least one routing node, each computing power service node is connected to at least one computing power server; the computing power service node is a leaf node of the multicast tree corresponding to the multicast group, and the non-leaf nodes of the multicast tree are routing nodes , the root node is a computing power edge node; the method includes:

接收目标算力边缘节点发送的组播查询报文,所述组播查询报文为所述目标算力边缘节点接收到终端设备的算力服务请求报文后发送的;receiving a multicast query message sent by the target computing power edge node, where the multicast query message is sent by the target computing power edge node after receiving the computing power service request message of the terminal device;

响应于所述组播查询报文,向所述目标算力边缘节点发送查询响应报文,所述查询响应报文中包括所述算力服务节点连接的算力服务器的算力资源信息,以使得所述目标算力边缘节点基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向所述目标算力服务器转发所述算力服务请求报文。In response to the multicast query message, send a query response message to the target computing power edge node, the query response message includes computing power resource information of the computing power server connected to the computing power service node, to The target computing power edge node selects a target computing power server based on the received computing power resource information of each computing power server, and forwards the computing power service request message to the target computing power server.

可选的,所述组播查询报文用于请求查询指定服务类型的算力资源信息,所述查询响应报文中包括支持所述指定服务类型的指定算力服务器的算力资源信息;Optionally, the multicast query message is used to request to query computing power resource information of a specified service type, and the query response message includes computing power resource information of a specified computing power server supporting the specified service type;

在所述向所述目标算力边缘节点发送查询响应报文之前,所述方法还包括:Before sending the query response message to the target computing power edge node, the method further includes:

从连接的算力服务器中,筛选出支持所述指定服务类型的指定算力服务器,获取所述指定算力服务器的算力资源信息。From the connected computing power servers, select a designated computing power server that supports the specified service type, and obtain computing power resource information of the designated computing power server.

可选的,所述组播查询报文用于请求查询指定服务类型的算力资源信息,所述查询响应报文中包括支持所述指定服务类型,且剩余算力资源满足预设条件的指定算力服务器的资源信息;Optionally, the multicast query message is used to request to query computing power resource information of a specified service type, and the query response message includes specifying that the specified service type is supported and the remaining computing power resources meet preset conditions. Resource information of computing power server;

在所述向所述目标算力边缘节点发送查询响应报文之前,所述方法还包括:Before sending the query response message to the target computing power edge node, the method further includes:

从自身连接的服务器中,筛选出支持所述指定服务类型且剩余算力资源满足预设条件的指定算力服务器,获取所述指定算力服务器的算力资源信息。From the servers connected to itself, select a designated computing power server that supports the specified service type and whose remaining computing power resources meet the preset conditions, and obtain computing power resource information of the designated computing power server.

第四方面,本申请实施例提供一种算力服务处理装置,应用于算力服务处理系统中的任一算力边缘节点,所述算力服务处理系统中还包括至少一个算力服务节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;所述装置包括:In the fourth aspect, the embodiment of the present application provides a computing power service processing device, which is applied to any computing power edge node in the computing power service processing system, and the computing power service processing system further includes at least one computing power service node and At least one routing node, each computing power service node is connected to at least one computing power server; the device includes:

接收模块,用于接收终端设备发送的算力服务请求报文;The receiving module is used to receive the computing power service request message sent by the terminal device;

发送模块,用于向所述算力边缘节点作为组播源的组播组中的算力服务节点发送组播查询报文,所述组播组对应的组播树中的叶子节点均为算力服务节点,非叶子节点为路由节点,根节点为所述算力边缘节点;A sending module, configured to send a multicast query message to a computing power service node in a multicast group with the computing power edge node as a multicast source, and the leaf nodes in the multicast tree corresponding to the multicast group are computing Power service nodes, non-leaf nodes are routing nodes, and the root node is the computing power edge node;

所述接收模块,还用于接收所述组播组中的算力服务节点发送的查询响应报文,所述查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息;The receiving module is further configured to receive a query response message sent by a computing power service node in the multicast group, where the query response message includes computing power resource information of a computing power server connected to the computing power service node;

选择模块,用于基于所述接收模块接收到的各算力服务器的算力资源信息,选择一个目标算力服务器;A selection module, configured to select a target computing power server based on the computing power resource information of each computing power server received by the receiving module;

所述发送模块,还用于向所述选择模块选择的所述目标算力服务器转发所述算力服务请求报文。The sending module is further configured to forward the computing power service request message to the target computing power server selected by the selecting module.

可选的,所述发送模块,具体用于:Optionally, the sending module is specifically used for:

从多个组播组中,查找支持所述指定服务类型的指定组播组;Finding a specified multicast group that supports the specified service type from multiple multicast groups;

向所述指定组播组中的算力服务节点发送所述组播查询报文。Sending the multicast query message to the computing service nodes in the specified multicast group.

第五方面,本申请实施例提供一种算力服务处理装置,应用于算力服务处理系统中的任一算力服务节点,所述算力服务处理系统中还包括至少一个算力边缘节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;所述算力服务节点为组播组对应的组播树的叶子节点,所述组播树的非叶子节点为路由节点,根节点为算力边缘节点;所述装置包括:In the fifth aspect, the embodiment of the present application provides a computing power service processing device, which is applied to any computing power service node in the computing power service processing system, and the computing power service processing system further includes at least one computing power edge node and At least one routing node, each computing power service node is connected to at least one computing power server; the computing power service node is a leaf node of the multicast tree corresponding to the multicast group, and the non-leaf nodes of the multicast tree are routing nodes , the root node is a computing power edge node; the device includes:

接收模块,用于接收目标算力边缘节点发送的组播查询报文,所述组播查询报文为所述目标算力边缘节点接收到终端设备的算力服务请求报文后发送的;The receiving module is configured to receive a multicast query message sent by the target computing power edge node, the multicast query message is sent after the target computing power edge node receives the computing power service request message of the terminal device;

发送模块,用于响应于所述组播查询报文,向所述目标算力边缘节点发送查询响应报文,所述查询响应报文中包括所述算力服务节点连接的算力服务器的算力资源信息,以使得所述目标算力边缘节点基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向所述目标算力服务器转发所述算力服务请求报文。A sending module, configured to send a query response message to the target computing power edge node in response to the multicast query message, where the query response message includes the computing power of the computing power server connected to the computing power service node. Power resource information, so that the target computing power edge node selects a target computing power server based on the received computing power resource information of each computing power server, and forwards the computing power service request message to the target computing power server .

可选的,所述组播查询报文用于请求查询指定服务类型的算力资源信息,所述查询响应报文中包括支持所述指定服务类型的指定算力服务器的算力资源信息;Optionally, the multicast query message is used to request to query computing power resource information of a specified service type, and the query response message includes computing power resource information of a specified computing power server supporting the specified service type;

所述装置还包括:The device also includes:

获取模块,用于从连接的算力服务器中,筛选出支持所述指定服务类型的指定算力服务器,获取所述指定算力服务器的算力资源信息。The acquisition module is configured to filter out a designated computing power server that supports the specified service type from the connected computing power servers, and obtain computing power resource information of the designated computing power server.

可选的,所述组播查询报文用于请求查询指定服务类型的算力资源信息,所述查询响应报文中包括支持所述指定服务类型,且剩余算力资源满足预设条件的指定算力服务器的资源信息;Optionally, the multicast query message is used to request to query computing power resource information of a specified service type, and the query response message includes specifying that the specified service type is supported and the remaining computing power resources meet preset conditions. Resource information of computing power server;

所述装置还包括:The device also includes:

获取模块,用于从自身连接的服务器中,筛选出支持所述指定服务类型且剩余算力资源满足预设条件的指定算力服务器,获取所述指定算力服务器的算力资源信息。The acquisition module is configured to select, from the servers connected to itself, a designated computing power server that supports the designated service type and whose remaining computing power resources meet preset conditions, and obtain computing power resource information of the designated computing power server.

第六方面,本申请实施例提供一种电子设备,包括处理器、通信接口、存储器和通信总线,其中,处理器,通信接口,存储器通过通信总线完成相互间的通信;In a sixth aspect, an embodiment of the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus;

存储器,用于存放计算机程序;memory for storing computer programs;

处理器,用于执行存储器上所存放的程序时,实现第二方面或第三方面任一所述的方法步骤。The processor is configured to implement the method steps described in any one of the second aspect or the third aspect when executing the program stored in the memory.

第七方面,本申请实施例还提供一种计算机可读存储介质,所述计算机可读存储介质内存储有计算机程序,所述计算机程序被处理器执行时实现第二方面或第三方面任一所述的方法步骤。In the seventh aspect, the embodiment of the present application also provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any one of the second aspect or the third aspect is implemented. The method steps described.

第八方面,本申请实施例提供了一种计算机程序产品,包括计算机程序,所述计算机程序在被处理器执行时实现根据第二方面或第三方面任一所述的方法。In an eighth aspect, an embodiment of the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method according to any one of the second aspect or the third aspect is implemented.

采用本申请实施例,算力边缘节点接收到终端设备发送的算力服务请求报文后,可向自身作为组播源的组播组中的算力服务节点发送组播查询报文,以便于获取各算力服务节点连接的算力服务器的算力资源信息,进而就无需控制器统一接收各算力服务节点的算力资源,降低了对控制器的要求。且算力服务处理系统中的各节点也无需互相泛洪通告算力资源,如此,就无需要求算力服务处理系统中的所有节点均支持算力网络服务。本申请实施例中,只需算力边缘节点和算力服务节点支持算力网络服务,其余路由节点只是进行普通的报文转发,所以在部署算力服务处理系统时,无需大幅调整大量的路由节点,只需将个别边缘节点部署为支持算力网络服务的设备,可以减少开发工作量,简化部署过程,降低部署成本。Using the embodiment of this application, after receiving the computing power service request message sent by the terminal device, the computing power edge node can send a multicast query message to the computing power service node in the multicast group that serves as the multicast source, so that The computing power resource information of the computing power server connected to each computing power service node is obtained, so that the controller does not need to uniformly receive the computing power resources of each computing power service node, which reduces the requirements for the controller. Moreover, each node in the computing power service processing system does not need to flood and notify each other of computing power resources. In this way, there is no need to require all nodes in the computing power service processing system to support computing power network services. In the embodiment of this application, only computing power edge nodes and computing power service nodes need to support computing power network services, and the rest of the routing nodes only perform ordinary message forwarding, so when deploying computing power service processing systems, there is no need to greatly adjust a large number of routes Nodes, you only need to deploy individual edge nodes as devices that support computing power network services, which can reduce the development workload, simplify the deployment process, and reduce deployment costs.

当然,实施本发明的任一产品或方法并不一定需要同时达到以上所述的所有优点。Of course, implementing any product or method of the present invention does not necessarily need to achieve all the above-mentioned advantages at the same time.

附图说明Description of drawings

为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,还可以根据这些附图获得其他的实施例。In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and those skilled in the art can also obtain other embodiments according to these drawings.

图1为相关技术中提供的一种集中式部署的算力网络的结构示意图;FIG. 1 is a schematic structural diagram of a centralized deployment computing power network provided in the related art;

图2为相关技术中提供的一种分布式部署的算力网络的结构示意图;FIG. 2 is a schematic structural diagram of a distributed computing power network provided in the related art;

图3为相关技术中提供的一种混合式部署的算力网络的结构示意图;FIG. 3 is a schematic structural diagram of a hybrid deployment computing power network provided in the related art;

图4为本申请实施例提供的一种算力服务处理系统的结构示意图;FIG. 4 is a schematic structural diagram of a computing power service processing system provided by an embodiment of the present application;

图5为本申请实施例提供的一种算力服务处理系统的示例性示意图;Fig. 5 is an exemplary schematic diagram of a computing power service processing system provided by an embodiment of the present application;

图6为本申请实施例提供的一种算力服务处理方法的流程图;Fig. 6 is a flow chart of a computing power service processing method provided by an embodiment of the present application;

图7为本申请实施例提供的另一种算力服务处理方法的流程图;Fig. 7 is a flow chart of another computing power service processing method provided by the embodiment of the present application;

图8为本申请实施例提供的一种算力服务处理装置的结构示意图;FIG. 8 is a schematic structural diagram of a computing power service processing device provided by an embodiment of the present application;

图9为本申请实施例提供的另一种算力服务处理装置的结构示意图;Fig. 9 is a schematic structural diagram of another computing power service processing device provided by the embodiment of the present application;

图10为本申请实施例提供的一种网络设备的结构示意图。FIG. 10 is a schematic structural diagram of a network device provided by an embodiment of the present application.

具体实施方式detailed description

下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员基于本申请所获得的所有其他实施例,都属于本发明保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art based on the present application belong to the protection scope of the present invention.

为方便理解,首先对本申请实施例中涉及到的相关概念进行解释。For ease of understanding, firstly, related concepts involved in the embodiments of the present application are explained.

算力网络,是指网络基础设施具备对算力资源的感知、调度和编排能力,可以在网络层提供网络、计算、存储的新型信息与通信技术(information and communicationtechnology,ICT)融合服务,是以网络为基础的新一代技术架构和业务运营体系。Computing power network means that the network infrastructure has the ability to perceive, schedule and orchestrate computing power resources, and can provide new information and communication technology (ICT) integrated services of network, computing and storage at the network layer. Network-based new generation technology architecture and business operation system.

在5G以及后5G时代,为了更迅捷高速地响应业务的计算需求,算力资源逐渐被下沉至靠近用户的边缘,形成异构多样、分布式的算力部署新态势。运营商可以借助网络基础设施,将多级分布的算力资源进行统一的动态纳管、调度和编排,实现全网资源的虚拟算力池化优势。In the 5G and post-5G era, in order to respond to business computing needs more quickly and quickly, computing power resources are gradually sinking to the edge close to users, forming a new situation of heterogeneous, diverse and distributed computing power deployment. With the help of network infrastructure, operators can conduct unified dynamic management, scheduling, and orchestration of multi-level distributed computing power resources to realize the advantages of virtual computing power pooling of network-wide resources.

如图1所示,图1为相关技术中的一种集中式控制面部署的算力网络,图1中示例性地示出了三个地区的网络架构,在纵向上,每个地区的网络均包括算力资源层、网络资源层和IP层(IP layer),且还包括应用(Application,APP)、计算机资源(computer resource),可选地,还可以包括一些特定的服务功能,例如图1中示出的功能1(function1)。As shown in Figure 1, Figure 1 is a computing power network deployed by a centralized control plane in the related art. Figure 1 schematically shows the network architecture of three regions. All include the computing power resource layer, the network resource layer and the IP layer (IP layer), and also include applications (Application, APP), computer resources (computer resources), and optionally, some specific service functions, such as graph Function 1 (function1) shown in 1.

在图1中,算力资源层和网络资源层需向控制器上报算力资源信息和网络状态信息,IP层需向控制器上报路由信息,进而控制器和编排器可以按照应用需求,结合全网的算力资源信息、网络状态信息以及路由信息,编排转发路径。In Figure 1, the computing power resource layer and the network resource layer need to report computing power resource information and network status information to the controller, and the IP layer needs to report routing information to the controller, and then the controller and orchestrator can combine the whole Network computing resource information, network status information and routing information, and arrange forwarding paths.

如图2所示,图2为相关技术中的一种分布式控制面部署的算力网络,图2中示例性地示出了三个地区的网络结构,与图1的区别为该算力网络中无需部署集中编排器和控制器,算力资源层之间可以互相泛洪发布算力资源,网络资源层之间可以互相泛洪发布网络状态信息,IP层之间可以互相泛洪发布路由信息。具体可以在网络域内通过IGP协议或在网络域间通过BGP协议进行发布。As shown in Figure 2, Figure 2 is a computing power network deployed in a distributed control plane in the related art. Figure 2 exemplarily shows the network structure of three regions. There is no need to deploy centralized orchestrators and controllers in the network. Computing resource layers can flood each other to release computing resources, network resource layers can flood each other to release network status information, and IP layers can flood each other to release routes information. Specifically, it can be advertised through an IGP protocol within a network domain or through a BGP protocol between network domains.

进而分布式算力网络中的路由设备可以基于学习到的全网的算力资源信息、网络状态信息和路由信息计算转发路径。Furthermore, the routing device in the distributed computing power network can calculate the forwarding path based on the learned computing power resource information, network status information and routing information of the entire network.

如图3所示,图3为相关技术中的一种混合式控制面部署的算力网络,图3为在图1和图2基础上的改进,其中左侧三列为一个网络域,右侧三列为另一个网络域,各节点可以在网络域内进行泛洪发布,通过控制器和编排器进行网络域间的信息传递。As shown in Figure 3, Figure 3 is a computing power network deployed in a hybrid control plane in the related art. The third column on the side is another network domain, and each node can perform flood publishing in the network domain, and transmit information between network domains through the controller and orchestrator.

上述三种算力网络均存在部署复杂的问题,例如在集中式算力网络中,需要对控制器进行大量的开发工作,在分布式算力网络中,需要所有节点均支持算力网络服务,且需为所有节点配置用于泛洪发布信息的交互协议,开发工作量巨大。分布式算力网络缓解了上述问题,但仍未完全解决。The above three computing power networks all have complex deployment problems. For example, in a centralized computing power network, a lot of development work needs to be done on the controller. In a distributed computing power network, all nodes need to support computing power network services. And it is necessary to configure an interactive protocol for flooding and publishing information for all nodes, and the development workload is huge. Distributed computing power network alleviates the above problems, but it is still not completely solved.

为了解决上述问题,本申请实施例提供了一种算力服务处理系统,如图4所示,该系统包括至少一个算力服务节点41、至少一个算力边缘节点42以及至少一个路由节点43,每个算力服务节点41连接于至少一个算力服务器44。需要说明的是,图4仅为示例,实际实现中各设备的数量不限于此。In order to solve the above problems, an embodiment of the present application provides a computing power service processing system, as shown in FIG. 4 , the system includes at least one computing power service node 41, at least one computing power edge node 42, and at least one routing node 43, Each computing power service node 41 is connected to at least one computing power server 44 . It should be noted that FIG. 4 is only an example, and the number of devices in actual implementation is not limited thereto.

其中,算力服务节点41可以为服务侧的边缘路由设备,算力边缘节点42为终端侧的边缘路由设备,算力服务节点41和算力边缘节点42均为支持算力网络服务的路由设备,路由节点43为普通的路由设备,可以不支持算力网络服务。算力服务器44可以为实体的服务器,也可以为分布式部署的算力服务单元。Among them, the computing power service node 41 can be an edge routing device on the service side, and the computing power edge node 42 is an edge routing device on the terminal side, and both the computing power service node 41 and the computing power edge node 42 are routing devices that support computing power network services , the routing node 43 is an ordinary routing device, which may not support computing power network services. The computing power server 44 can be a physical server, or a distributed computing power service unit.

每个算力边缘节点42,用于接收终端设备发送的算力服务请求报文,并向自身作为组播源的组播组中的算力服务节点41发送组播查询报文,组播组对应的组播树中的叶子节点均为算力服务节点41,非叶子节点为路由节点43,根节点为算力边缘节点42。Each computing power edge node 42 is used to receive the computing power service request message sent by the terminal device, and send a multicast query message to the computing power service node 41 in the multicast group that serves as the multicast source, and the multicast group The leaf nodes in the corresponding multicast tree are computing power service nodes 41 , the non-leaf nodes are routing nodes 43 , and the root nodes are computing power edge nodes 42 .

组播树也可称为组播转发树,用于表示组播源至组播组中各组播组成员之间的拓扑关系,组播组中的组播成员为算力服务节点41,所以组播树中的叶子节点均为算力服务节点41。可以理解的是,算力边缘节点42发出的组播查询报文,可沿着组播树经过路由节点43被传输至算力服务节点41。The multicast tree can also be called a multicast forwarding tree, which is used to represent the topological relationship between the multicast source and the multicast group members in the multicast group. The multicast members in the multicast group are computing power service nodes 41, so The leaf nodes in the multicast tree are computing power service nodes 41 . It can be understood that the multicast query message sent by the computing power edge node 42 can be transmitted to the computing power service node 41 through the routing node 43 along the multicast tree.

其中,算力服务请求报文为终端设备发送的用于请求提供算力服务的报文,作为示例,该算力服务可以为游戏服务。组播查询报文用于请求查询组播组中各算力服务节点41连接的算力服务器44的算力资源信息。Wherein, the computing power service request message is a message sent by the terminal device for requesting provision of computing power service. As an example, the computing power service may be a game service. The multicast query message is used to request to query the computing power resource information of the computing power server 44 connected to each computing power service node 41 in the multicast group.

每个算力服务节点41,用于响应于来自目标算力边缘节点的目标组播查询报文,向目标算力边缘节点发送查询响应报文,查询响应报文中包括该算力服务节点41连接的算力服务器44的算力资源信息。Each computing power service node 41 is configured to send a query response message to the target computing power edge node in response to the target multicast query message from the target computing power edge node, and the query response message includes the computing power service node 41 The computing power resource information of the connected computing power server 44.

其中,算力资源信息包括算力服务器的CPU计算能力信息、CPU占用率以及存储容量中的任意一项或多项,也可包括其他用于表示算力能力的信息,本申请实施例对此不作限定。Wherein, the computing power resource information includes any one or more items of CPU computing capability information, CPU occupancy rate, and storage capacity of the computing power server, and may also include other information used to represent computing power capabilities. Not limited.

本申请实施例中,目标算力边缘节点为该算力服务处理系统中发送该目标组播查询报文的算力边缘节点42。组播组中可能存在多个算力服务节点41接收到组播查询报文,每个接收到组播查询报文的算力服务节点41均可针对接收到的组播查询报文向算力边缘节点42发送查询响应报文。查询响应报文为单播报文。In the embodiment of the present application, the target computing power edge node is the computing power edge node 42 that sends the target multicast query message in the computing power service processing system. There may be multiple computing power service nodes 41 in the multicast group that receive the multicast query message, and each computing power service node 41 that receives the multicast query message can send the computing power service node 41 for the received multicast query message. The edge node 42 sends a query response message. The query response message is a unicast message.

每个算力边缘节点42,还用于基于接收到的各算力服务器44的算力资源信息,选择一个目标算力服务器,向目标算力服务器转发算力服务请求报文。Each computing power edge node 42 is further configured to select a target computing power server based on the received computing power resource information of each computing power server 44, and forward the computing power service request message to the target computing power server.

可以理解的是,算力边缘节点42可以沿着从算力边缘节点42至目标算力服务器连接的算力服务节点41的转发路径,转发已接收到的算力服务请求报文,进而算力服务节点41可将算力服务请求报文转发给目标算力服务器。目标算力服务器即可基于算力服务请求报文为终端设备提供算力服务。It can be understood that the computing power edge node 42 can forward the received computing power service request message along the forwarding path from the computing power edge node 42 to the computing power service node 41 connected to the target computing power server, and then the computing power The service node 41 may forward the computing power service request message to the target computing power server. The target computing power server can provide computing power service for the terminal device based on the computing power service request message.

采用本申请实施例,在该算力服务处理系统中,算力边缘节点接收到终端设备发送的算力服务请求报文后,可向自身作为组播源的组播组中的算力服务节点发送组播查询报文,以便于获取各算力服务节点连接的算力服务器的算力资源信息,进而就无需控制器统一接收各算力服务节点的算力资源,降低了对控制器的要求。且算力服务处理系统中的各节点也无需互相泛洪通告算力资源,如此,就无需要求算力服务处理系统中的所有节点均支持算力网络服务。本申请实施例中,只需算力边缘节点和算力服务节点支持算力网络服务,其余路由节点只是进行普通的报文转发,所以在部署算力服务处理系统时,无需大幅调整大量的路由节点,只需将个别边缘节点部署为支持算力网络服务的设备,可以减少开发工作量,简化部署过程,降低部署成本。Using the embodiment of the present application, in the computing power service processing system, after receiving the computing power service request message sent by the terminal device, the computing power edge node can send the computing power service node in the multicast group as the multicast source to itself Send a multicast query message to obtain the computing power resource information of the computing power server connected to each computing power service node, and then there is no need for the controller to uniformly receive the computing power resources of each computing power service node, reducing the requirements for the controller . Moreover, each node in the computing power service processing system does not need to flood and notify each other of computing power resources. In this way, there is no need to require all nodes in the computing power service processing system to support computing power network services. In the embodiment of this application, only computing power edge nodes and computing power service nodes need to support computing power network services, and the rest of the routing nodes only perform ordinary message forwarding, so when deploying computing power service processing systems, there is no need to greatly adjust a large number of routes Nodes, you only need to deploy individual edge nodes as devices that support computing power network services, which can reduce the development workload, simplify the deployment process, and reduce deployment costs.

可选地,终端设备发送的算力服务请求报文用于请求指定服务类型的算力服务。Optionally, the computing power service request message sent by the terminal device is used to request computing power service of a specified service type.

每个算力边缘节点42,具体用于在接收到算力服务请求报文后,从自身作为组播源的多个组播组中,查找支持指定服务类型的指定组播组,向指定组播组中的算力服务节点41发送组播查询报文。Each computing power edge node 42 is specifically used to search for a designated multicast group that supports a designated service type from a plurality of multicast groups that it serves as a multicast source after receiving a computing power service request message, and send a message to the designated group The computing power service node 41 in the broadcast group sends a multicast query message.

其中,上述多个组播组的组播源均为该算力边缘节点42,算力边缘节点42可预先存储各组播组支持的服务类型,每个组播组支持一种服务类型,组播组的组播组成员为支持该服务类型的算力服务节点41。例如,服务类型可以为内容分发网络(Content DeliveryNetwork,CDN)、人工智能(Artificial Intelligence,AI)和物联网(Internet of things,IoT)等。Among them, the multicast source of the above-mentioned multiple multicast groups is the computing power edge node 42, and the computing power edge node 42 can pre-store the service types supported by each multicast group, and each multicast group supports a service type, and the group The multicast group members of the broadcast group are computing power service nodes 41 that support the service type. For example, the service type may be content delivery network (Content Delivery Network, CDN), artificial intelligence (Artificial Intelligence, AI), Internet of things (Internet of things, IoT) and so on.

假设一个算力边缘节点42分别是组播组A、组播组B和组播组C的组播源,组播组A支持的服务类型为CDN、组播组B支持的服务类型为AI,组播组C支持的服务类型为IoT,若上述指定服务类型为AI,即算力服务请求报文用于请求AI服务,则该算力边缘节点42可向组播组B发送组播查询报文。Assume that a computing power edge node 42 is the multicast source of multicast group A, multicast group B, and multicast group C respectively, the service type supported by multicast group A is CDN, and the service type supported by multicast group B is AI. The service type supported by multicast group C is IoT. If the specified service type is AI, that is, the computing power service request message is used to request AI service, then the computing power edge node 42 can send a multicast query message to multicast group B. arts.

采用本申请实施例,算力边缘节点可向支持指定服务类型的组播组中的算力服务节点发送组播查询报文,可以避免不支持指定服务类型的算力服务节点接收到组播查询报文,可以降低查询开销,提高查询效率。By adopting the embodiment of this application, the computing power edge node can send a multicast query message to the computing power service node in the multicast group that supports the specified service type, which can prevent the computing power service node that does not support the specified service type from receiving the multicast query message, which can reduce query overhead and improve query efficiency.

在上述任一实施例的基础上,目标组播查询报文用于请求查询指定服务类型的算力资源信息,该指定服务类型为客户端发送的算力服务请求报文请求的算力服务的类型。On the basis of any of the above embodiments, the target multicast query message is used to request to query the computing power resource information of the specified service type, and the specified service type is the computing power service requested by the computing power service request message sent by the client. type.

可以理解为,算力边缘节点42接收到算力服务请求报文后,可获取该算力服务请求报文所请求的算力服务的指定服务类型,进而向算力服务节点41发送用于查询该指定服务类型的算力资源的组播查询报文。It can be understood that, after receiving the computing power service request message, the computing power edge node 42 can obtain the specified service type of the computing power service requested by the computing power service request message, and then send a query to the computing power service node 41 A multicast query message for computing resources of the specified service type.

每个算力服务节点41,具体用于响应于接收到的目标组播查询报文,从连接的算力服务器44中,筛选出支持指定服务类型的指定算力服务器,向目标算力边缘节点发送查询响应报文,查询响应报文中包括指定算力服务器的算力资源信息。Each computing power service node 41 is specifically used to respond to the received target multicast query message, select the designated computing power server that supports the specified service type from the connected computing power servers 44, and send the target computing power to the target computing power edge node. Send a query response message, and the query response message includes the computing power resource information of the specified computing power server.

可选地,每个算力服务节点41,还用于周期性获取与自身连接的各算力服务器44的算力资源信息。比如,可以每隔5秒获取一次,如此可以使得算力服务节点41及时获知各算力服务器的算力资源信息,并缓存在本地,进而可以准确快速地向算力边缘节点42反馈各算力服务器的算力资源信息。Optionally, each computing power service node 41 is also configured to periodically obtain computing power resource information of each computing power server 44 connected to itself. For example, it can be obtained once every 5 seconds, so that the computing power service node 41 can know the computing power resource information of each computing power server in time, and cache it locally, and then can accurately and quickly feed back each computing power to the computing power edge node 42 Information about computing power resources of the server.

其中,算力服务节点41接收到的目标组播查询报文携带指定服务类型的标识,算力服务节点41连接的不同算力服务器可以支持不同类型的算力服务,算力服务节点可以筛选出支持指定服务类型的算力服务器,并获取指定服务类型的算力服务器的算力资源信息。如此,算力边缘节点42接收到的算力资源信息均为支持指定服务类型的算力服务器的算力资源信息,可以避免向算力边缘节点42传输不支持指定服务类型的算力服务器的算力资源,可以节省传输开销,进而节省网络资源,提升查询效率。Among them, the target multicast query message received by the computing power service node 41 carries the identifier of the specified service type, different computing power servers connected to the computing power service node 41 can support different types of computing power services, and the computing power service node can filter out Support computing power servers of the specified service type, and obtain the computing power resource information of the computing power servers of the specified service type. In this way, the computing power resource information received by the computing power edge node 42 is the computing power resource information of the computing power server that supports the specified service type, which can avoid transmitting the computing power server that does not support the specified service type to the computing power edge node 42. It can save transmission overhead, thereby saving network resources and improving query efficiency.

另一实施方式中,每个算力服务节点41,具体用于响应于接收到的目标组播查询报文,从连接的算力服务器44中,筛选出支持指定服务类型且剩余算力资源满足预设条件的指定算力服务器,向目标算力边缘节点发送查询响应报文,查询响应报文中包括指定算力服务器的算力资源信息。In another embodiment, each computing power service node 41 is specifically used to respond to the received target multicast query message, and select from the connected computing power servers 44 that support the specified service type and the remaining computing power resources satisfy The designated computing power server with preset conditions sends a query response message to the target computing power edge node, and the query response message includes the computing power resource information of the designated computing power server.

即算力服务节点41不仅需要筛选支持指定服务类型的算力服务器,还需进一步判断筛选出的算力服务器的剩余算力资源是否满足预设条件,进而向目标算力边缘节点发送满足预设条件的算力服务器的算力资源信息。That is, the computing power service node 41 not only needs to screen the computing power servers that support the specified service type, but also needs to further judge whether the remaining computing power resources of the selected computing power servers meet the preset conditions, and then send the target computing power edge node that satisfies the preset conditions. Computing power resource information of the computing power server of the condition.

其中,预设条件可以基于实际需求设置,例如,预设条件可以为剩余算力资源中的CPU利用率小于预设的CPU利用率阈值,且剩余内存空间大于内存空间阈值。Wherein, the preset condition can be set based on actual needs, for example, the preset condition can be that the CPU utilization rate in the remaining computing resources is less than a preset CPU utilization rate threshold, and the remaining memory space is greater than the memory space threshold value.

也就是说,筛选出的指定算力服务器有足够的算力资源,可以避免向目标算力边缘节点反馈剩余算力资源不足的算力服务器的资源信息,可以节省传输开销,且可提高算力边缘节点选择目标服务器的效率,使得算力边缘节点可以更快速地处理算力服务请求报文,降低算力服务的时延。That is to say, the selected designated computing power server has sufficient computing power resources, which can avoid feeding back the resource information of computing power servers with insufficient remaining computing power resources to the target computing power edge node, can save transmission costs, and can increase computing power The efficiency of edge nodes in selecting target servers enables computing power edge nodes to process computing power service request messages more quickly and reduce the delay of computing power services.

在本申请另一实施例中,该系统中还包括汇聚节点(Rendezvous Point,RP),汇聚节点也是一台路由器,为组播网络中的关键角色。其中,在组播网络中,可以预先以静态配置的方式将一台路由器配置为RP,或者也可动态选择RP,确定RP的方法可参考标准组播协议。In another embodiment of the present application, the system further includes a rendezvous point (Rendezvous Point, RP), and the rendezvous point is also a router, which is a key role in the multicast network. Wherein, in a multicast network, a router may be statically configured in advance as an RP, or an RP may be dynamically selected, and the method for determining an RP may refer to a standard multicast protocol.

在算力服务处理系统中,运行组播协议的路由器可以通过组播路由协议获知RP。In the computing power service processing system, the router running the multicast protocol can learn the RP through the multicast routing protocol.

每个算力服务节点41,还用于在接入到算力服务处理系统中时,向汇聚节点发送注册信息,该注册信息包括该算力服务节点41连接的算力服务器的服务类型。Each computing power service node 41 is also configured to send registration information to the aggregation node when accessing the computing power service processing system, and the registration information includes the service type of the computing power server to which the computing power service node 41 is connected.

汇聚节点,用于响应于来自目标算力服务节点的目标注册信息,基于目标注册信息携带的服务类型,将目标算力服务节点加入到与注册信息携带的服务类型匹配的组播组中,并更新该组播组对应的组播树。The converging node is configured to, in response to the target registration information from the target computing power service node, add the target computing power service node to the multicast group matching the service type carried in the registration information based on the service type carried in the target registration information, and Update the multicast tree corresponding to the multicast group.

可以理解的是,每个算力服务节点41在接入算力服务处理系统中时,需先向汇聚节点进行注册,进而才可以作为组播成员加入组播组。目标算力服务节点为向汇聚节点发送注册信息的算力服务节点41。It can be understood that, when accessing the computing power service processing system, each computing power service node 41 needs to register with the aggregation node first, and then can join the multicast group as a multicast member. The target computing power service node is the computing power service node 41 that sends registration information to the aggregation node.

本申请实施例中,每个算力服务节点41可以连接多个算力服务器44,且每个算力服务器44的服务类型可以不同,并且单个算力服务器44也可以同时提供多种类型的服务。相应地,算力服务节点41向汇聚节点发送的注册信息可以携带多种服务类型,汇聚节点可以基于多种服务类型将该算力服务节点41加入多个组播组中。In this embodiment of the application, each computing power service node 41 can be connected to multiple computing power servers 44, and the service types of each computing power server 44 can be different, and a single computing power server 44 can also provide multiple types of services at the same time . Correspondingly, the registration information sent by the computing power service node 41 to the convergence node may carry multiple service types, and the convergence node may add the computing power service node 41 to multiple multicast groups based on multiple service types.

采用该方法,当有算力服务节点加入算力服务处理系统时,汇聚节点可以实时将该算力服务节点加入组播组并更新组播组,无需该算力服务节点向算力服务处理系统中的其他节点进行信息通告,可以实现算力网络的快速部署。With this method, when a computing power service node joins the computing power service processing system, the aggregation node can add the computing power service node to the multicast group and update the multicast group in real time, without the computing power service node sending the computing power service processing system The other nodes in the network are notified of information, which can realize the rapid deployment of the computing power network.

此外,每个算力边缘节点42也可根据自身接入的终端所需的服务类型,向汇聚节点请求注册为组播源。即算力边缘节点42也可以向汇聚节点发送注册信息,该注册信息包括算力边缘节点42能够为终端设备提供的服务类型。汇聚节点基于注册信息携带的服务类型,将算力边缘节点42作为该服务类型对应的组播组的组播源。In addition, each computing power edge node 42 can also request the aggregation node to register as a multicast source according to the type of service required by the terminal it accesses. That is, the computing power edge node 42 may also send registration information to the aggregation node, and the registration information includes the type of service that the computing power edge node 42 can provide for the terminal device. Based on the service type carried in the registration information, the convergence node uses the computing power edge node 42 as the multicast source of the multicast group corresponding to the service type.

以下结合具体例子对本申请实施例提供的算力服务处理系统进行介绍,如图5所示,图5中的节点A至H均为算力服务处理系统中的网络节点,这些网络节点均可以为路由设备。The following is an introduction to the computing power service processing system provided by the embodiment of the present application in conjunction with specific examples. As shown in Figure 5, the nodes A to H in Figure 5 are all network nodes in the computing power service processing system, and these network nodes can be routing device.

其中,节点A连接于算力服务器1,节点B连接于算力服务器2。节点A和节点B均为算力服务节点。图5中示例性的示出了节点A连接的一个算力服务器1以及节点B连接的一个算力服务器2,在实际部署中,节点A和节点B也可连接多个算力服务器。Among them, node A is connected to computing power server 1, and node B is connected to computing power server 2. Node A and Node B are computing power service nodes. Figure 5 exemplarily shows a computing power server 1 connected to node A and a computing power server 2 connected to node B. In actual deployment, node A and node B may also be connected to multiple computing power servers.

节点C、D、E、H为传统网络路由节点,支持IPv6和/或IPv4组播和单播转发。Nodes C, D, E, and H are traditional network routing nodes, supporting IPv6 and/or IPv4 multicast and unicast forwarding.

节点G和F为算力边缘节点,即新型的支持算力网络服务的边缘节点,如果终端设备需要算力网络服务,则终端设备可接入算力边缘节点。Nodes G and F are computing power edge nodes, which are new edge nodes that support computing power network services. If terminal devices need computing power network services, terminal devices can access computing power edge nodes.

以终端设备为个人计算机(Personal Computer,PC)为例,图5中还示出了两个终端设备,分别为PC-A和PC-B。其中,PC-A上联的节点G为支持算力网络服务的算力边缘节点,因此PC-A可以使用算力网络服务。PC-B上联的节点H为传统路由器,因此PC-B不能使用算力网络服务。Taking the terminal device as an example of a personal computer (Personal Computer, PC), FIG. 5 also shows two terminal devices, PC-A and PC-B. Among them, the node G connected to PC-A is a computing power edge node that supports computing power network services, so PC-A can use computing power network services. The node H connected to PC-B is a traditional router, so PC-B cannot use computing power network services.

可选地,图5中的各节点还可以与SDN控制器(图5中未示出)连接。Optionally, each node in FIG. 5 may also be connected to an SDN controller (not shown in FIG. 5 ).

在网络部署时,如果需要实现算力网络服务,则可以在有算力网络服务需求的终端设备接入的边缘节点,替换为本申请实施例中的算力边缘节点,并配置算力边缘节点支持的算力网络服务类型。During network deployment, if computing power network services need to be implemented, the edge nodes accessed by terminal devices with computing power network service requirements can be replaced with computing power edge nodes in the embodiment of this application, and computing power edge nodes can be configured Supported computing power network service types.

例如,在PC-A无需使用算力网络服务时,PC-A接入的边缘节点可以为普通的路由节点,在需要使用算力网络服务时,可将PC-A接入的路由节点替换为支持算力服务的算力边缘节点。For example, when PC-A does not need to use computing power network services, the edge nodes accessed by PC-A can be ordinary routing nodes. When PC-A needs to use computing power network services, the routing nodes accessed by PC-A can be replaced by A computing power edge node that supports computing power services.

节点G和节点F均可作为组播组的组播源,节点A和节点B均可作为组播组中的组播成员。Both node G and node F can be used as the multicast source of the multicast group, and both node A and node B can be used as multicast members in the multicast group.

作为示例,算力服务器1为游戏a提供算力服务,算力服务器2为游戏a和游戏b提供算力服务。As an example, computing power server 1 provides computing power service for game a, and computing power server 2 provides computing power service for game a and game b.

游戏a对应的组播组为(*,游戏a),游戏b对应的组播组为(*,游戏b)。The multicast group corresponding to game a is (*, game a), and the multicast group corresponding to game b is (*, game b).

在节点A接入该算力服务处理系统时,节点A可向汇聚节点(图5中未示出)发送注册信息,注册信息中包括的服务类型为游戏a,进而汇聚节点将节点A加入到组播组(*,游戏a)中。When node A accesses the computing power service processing system, node A can send registration information to the sink node (not shown in Figure 5), the service type included in the registration information is game a, and then the sink node adds node A to In the multicast group (*, game a).

在节点B接入该算力服务处理系统时,节点B向汇聚节点发送注册信息,注册信息中包括的服务类型为游戏a和游戏b,进而汇聚节点将节点B加入到组播组(*,游戏a)和(*,游戏b)中。When Node B accesses the computing power service processing system, Node B sends registration information to the sink node. The service types included in the registration information are game a and game b, and then the sink node adds Node B to the multicast group (*, game a) and (*, game b).

假设节点G连接的PC-A需要游戏a的算力服务,节点G可向汇聚节点申请注册为组播组(*,游戏a)的组播源。Assuming that the PC-A connected to node G needs the computing power service of game a, node G can apply to the sink node for registration as the multicast source of the multicast group (*, game a).

当组播组中的组播成员发生更新时,汇聚节点可根据组播源与组播成员之间的拓扑位置关系,更新组播树。或者,在创建一个新的组播组后,汇聚节点可根据组播源与组播成员之间的拓扑位置关系构建组播树。When the multicast members in the multicast group are updated, the sink node can update the multicast tree according to the topology position relationship between the multicast source and the multicast members. Alternatively, after creating a new multicast group, the sink node can build a multicast tree according to the topological positional relationship between the multicast source and the multicast members.

例如,组播组(*,游戏a)对应的组播树包括两个分支,分别为节点G-节点D-节点A,和节点G-节点D-节点E-节点B,具体为图5中的组播树①。可以看出,组播树①对应的组播源为节点G,组播成员为节点A和节点B。For example, the multicast tree corresponding to the multicast group (*, game a) includes two branches, which are respectively node G-node D-node A, and node G-node D-node E-node B, specifically in Figure 5 The multicast tree①. It can be seen that the multicast source corresponding to the multicast tree ① is node G, and the multicast members are node A and node B.

组播组(*,游戏b)对应的组播树包括一个分支,为节点F-节点E-节点B,具体为图5中的组播树②。可以看出,组播树②对应的组播源为节点F,组播成员为节点B。The multicast tree corresponding to the multicast group (*, game b) includes a branch, node F-node E-node B, specifically the multicast tree ② in FIG. 5 . It can be seen that the multicast source corresponding to the multicast tree ② is node F, and the multicast member is node B.

当PC-A需要网络服务时,可向节点G发送算力服务请求报文。节点G接收到算力服务请求报文后,可以基于算力服务请求报文识别PC-A请求的服务类型为游戏a。进而,节点G向RP发送目的组播地址为(*,游戏a)的组播请求报文,RP接收到该组播请求报文后,向组播组(*,游戏a)的所有组播成员转发该组播请求报文,也就是沿着组播树①向节点A和节点B转发该组播请求报文。When PC-A needs network services, it can send a computing power service request message to node G. After node G receives the computing power service request message, it can identify the service type requested by PC-A as game a based on the computing power service request message. Furthermore, node G sends a multicast request message with the destination multicast address as (*, game a) to the RP, and after receiving the multicast request message, the RP sends all multicast messages to the multicast group (*, game a) Members forward the multicast request message, that is, forward the multicast request message to node A and node B along the multicast tree ①.

节点A接收到组播查询报文后,可沿路径:节点A-节点D-节点G,向节点G回复单播的业务查询响应报文,该业务查询响应报文中包括算力服务器1的算力资源信息。After node A receives the multicast query message, it can reply a unicast service query response message to node G along the path: node A-node D-node G, and the service query response message includes computing power server 1 Computing resource information.

同理,节点B接收到组播查询报文后,可沿路径:节点B-节点E-节点D-节点G,向节点G回复单播的业务查询响应报文,该业务查询响应报文中包括算力服务器2的算力资源信息。Similarly, after node B receives the multicast query message, it can reply a unicast service query response message to node G along the path: node B-node E-node D-node G, and the service query response message contains Including computing power resource information of computing power server 2.

进而节点G可从算力服务器1和算力服务器2的算力资源信息,选择合适的算力服务器。例如若算力服务器1的算力资源更充足,且距离节点G较近,则可选择算力服务器1,进而节点G可向算力服务器1转发算力服务请求报文,后续算力服务器1可为PC-A提供算力服务。Furthermore, node G can select an appropriate computing power server from computing power resource information of computing power server 1 and computing power server 2 . For example, if computing power server 1 has more sufficient computing resources and is closer to node G, computing power server 1 can be selected, and then node G can forward computing power service request messages to computing power server 1, and subsequent computing power server 1 Computing service can be provided for PC-A.

可见,采用该系统,相比于分布式的部署方式,本申请实施例无需额外部署复杂的路由协议,可以极大减少算力路由协议的复杂度和开发工作量。并且通过在现有网络的基础上更换算力边缘节点和算力服务节点,即可实现算力网络设备的快速部署,无需对现有网络中的其他设备进行开发,通过简单的组播路由协议即可实现算力路由和算力状态的同步,极大减少了开发工作量和部署复杂度。It can be seen that using this system, compared with the distributed deployment method, the embodiment of the present application does not require additional deployment of complex routing protocols, which can greatly reduce the complexity and development workload of computing power routing protocols. And by replacing the computing power edge nodes and computing power service nodes on the basis of the existing network, the rapid deployment of computing power network equipment can be realized, without the need to develop other devices in the existing network, through a simple multicast routing protocol The synchronization of computing power routing and computing power status can be realized, which greatly reduces the development workload and deployment complexity.

对应于上述系统实施例,本申请实施例还提供一种算力服务处理方法,该方法可以应用于上述算力服务处理系统中的任一算力边缘节点,如图6所示,该方法包括:Corresponding to the above system embodiment, the embodiment of the present application also provides a computing power service processing method, which can be applied to any computing power edge node in the above computing power service processing system, as shown in Figure 6, the method includes :

S601、接收终端设备发送的算力服务请求报文。S601. Receive a computing power service request message sent by a terminal device.

其中,算力服务请求报文用于请求指定服务类型的算力服务。Among them, the computing power service request message is used to request the computing power service of the specified service type.

S602、向自身作为组播源的组播组中的算力服务节点发送组播查询报文。S602. Send a multicast query message to the computing service node in the multicast group that serves as the multicast source.

其中,组播组对应的组播树中的叶子节点均为算力服务节点,非叶子节点为路由节点,根节点为算力边缘节点。Among them, the leaf nodes in the multicast tree corresponding to the multicast group are computing power service nodes, the non-leaf nodes are routing nodes, and the root nodes are computing power edge nodes.

可选地,算力边缘节点可向自身作为组播源的所有组播组发送组播查询报文,进而接收到该组播查询报文的算力服务节点可判断是否支持指定服务类型,若支持,则向算力边缘节点发送查询响应报文。或者,算力边缘节点可从多个组播组中,查找支持指定服务类型的指定组播组,并向该指定组播组中的算力服务节点发送组播查询报文。Optionally, the computing power edge node can send a multicast query message to all the multicast groups that it serves as the multicast source, and then the computing power service node that receives the multicast query message can determine whether to support the specified service type, if If supported, send a query response message to the computing power edge node. Alternatively, the computing power edge node may search for a specified multicast group supporting a specified service type from multiple multicast groups, and send a multicast query message to the computing power service nodes in the specified multicast group.

S603、接收组播组中的算力服务节点发送的查询响应报文,查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息。S603. Receive a query response message sent by the computing power service node in the multicast group, where the query response message includes computing power resource information of the computing power server connected to the computing power service node.

S604、基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向目标算力服务器转发算力服务请求报文。S604. Based on the received computing power resource information of each computing power server, select a target computing power server, and forward the computing power service request message to the target computing power server.

其中,算力边缘节点可以根据预设的判断条件选择目标算力服务器。Among them, the computing power edge node can select the target computing power server according to the preset judgment conditions.

例如,可以选择与算力边缘节点的拓扑距离最近的算力服务器作为目标算力服务器,或者选择CPU占用率低于预设阈值的算力服务器作为目标算力服务器,或者对拓扑距离远近、CPU占用率以及存储容量等算力资源信息进行综合加权,选择一个加权值最高的算力服务器作为目标算力服务器。For example, the computing power server with the closest topological distance to the computing power edge node can be selected as the target computing power server, or the computing power server whose CPU usage is lower than the preset threshold can be selected as the target computing power server, or the distance between the topology, CPU Computing power resource information such as occupancy rate and storage capacity is comprehensively weighted, and a computing power server with the highest weighted value is selected as the target computing power server.

采用本申请实施例,算力边缘节点接收到终端设备发送的算力服务请求报文后,可向组播组中的算力服务节点发送组播查询报文,以便于获取各算力服务节点连接的算力服务器的算力资源信息,进而就无需控制器统一接收各算力服务节点的算力资源,降低了对控制器的要求。且算力服务处理系统中的各节点也无需互相泛洪通告算力资源,如此,就无需要求算力服务处理系统中的所有节点均支持算力网络服务。本申请实施例中,只需算力边缘节点和算力服务节点支持算力网络服务,其余路由节点只是进行普通的报文转发,所以在部署算力服务处理系统时,无需大幅调整大量的路由节点,只需将个别边缘节点部署为支持算力网络服务的设备,可以减少开发工作量,简化部署过程,降低部署成本。According to the embodiment of this application, after receiving the computing power service request message sent by the terminal device, the computing power edge node can send a multicast query message to the computing power service node in the multicast group, so as to obtain each computing power service node The computing power resource information of the connected computing power server does not require the controller to uniformly receive the computing power resources of each computing power service node, which reduces the requirements for the controller. Moreover, each node in the computing power service processing system does not need to flood and notify each other of computing power resources. In this way, there is no need to require all nodes in the computing power service processing system to support computing power network services. In the embodiment of this application, only computing power edge nodes and computing power service nodes need to support computing power network services, and the rest of the routing nodes only perform ordinary message forwarding, so when deploying computing power service processing systems, there is no need to greatly adjust a large number of routes Nodes, you only need to deploy individual edge nodes as devices that support computing power network services, which can reduce the development workload, simplify the deployment process, and reduce deployment costs.

对应于上述系统实施例,本申请实施例还提供一种算力服务处理方法,该方法应用于算力服务处理系统中的任一算力服务节点,如图7所示,该方法包括:Corresponding to the above system embodiment, the embodiment of the present application also provides a computing power service processing method, which is applied to any computing power service node in the computing power service processing system, as shown in Figure 7, the method includes:

S701、接收目标算力边缘节点发送的组播查询报文,组播查询报文为目标算力边缘节点接收到终端设备的算力服务请求报文后发送的。S701. Receive a multicast query message sent by a target computing power edge node. The multicast query message is sent after the target computing power edge node receives a computing power service request message from a terminal device.

其中,目标算力边缘节点可以为算力服务处理系统中的任一算力边缘节点,算力服务请求报文用于请求指定服务类型的算力服务,组播查询报文可以携带该指定服务类型的标识,组播查询报文用于请求查询指定服务类型的算力资源信息。Among them, the target computing power edge node can be any computing power edge node in the computing power service processing system, the computing power service request message is used to request the computing power service of the specified service type, and the multicast query message can carry the specified service The identification of the type, the multicast query message is used to request to query the computing resource information of the specified service type.

S702、响应于组播查询报文,向目标算力边缘节点发送查询响应报文,查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息,以使得目标算力边缘节点基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向目标算力服务器转发算力服务请求报文。S702. In response to the multicast query message, send a query response message to the target computing power edge node. The query response message includes the computing power resource information of the computing power server connected to the computing power service node, so that the target computing power edge node Based on the received computing power resource information of each computing power server, a target computing power server is selected, and the computing power service request message is forwarded to the target computing power server.

采用本申请实施例,算力服务节点可接收目标算力边缘节点发送的组播查询报文,并获取各算力服务节点连接的算力服务器的算力资源信息,进而就无需控制器统一接收各算力服务节点的算力资源,降低了对控制器的要求。且算力服务处理系统中的各节点也无需互相泛洪通告算力资源,如此,就无需要求算力服务处理系统中的所有节点均支持算力网络服务。本申请实施例中,只需算力边缘节点和算力服务节点支持算力网络服务,其余路由节点只是进行普通的报文转发,所以在部署算力服务处理系统时,无需大幅调整大量的路由节点,只需将个别边缘节点部署为支持算力网络服务的设备,可以减少开发工作量,简化部署过程,降低部署成本。Using the embodiment of this application, the computing power service node can receive the multicast query message sent by the target computing power edge node, and obtain the computing power resource information of the computing power server connected to each computing power service node, and then there is no need for the controller to receive The computing power resources of each computing power service node reduce the requirements for the controller. Moreover, each node in the computing power service processing system does not need to flood and notify each other of computing power resources. In this way, there is no need to require all nodes in the computing power service processing system to support computing power network services. In the embodiment of this application, only computing power edge nodes and computing power service nodes need to support computing power network services, and the rest of the routing nodes only perform ordinary message forwarding, so when deploying computing power service processing systems, there is no need to greatly adjust a large number of routes Nodes, you only need to deploy individual edge nodes as devices that support computing power network services, which can reduce the development workload, simplify the deployment process, and reduce deployment costs.

可选的,算力服务节点可以周期性获取与自身连接的各算力服务器的算力资源信息,并将算力服务器的算力资源信息缓存在本地,当接收到组播查询报文后,可以快速的向算力边缘节点反馈自身连接的算力服务器的算力资源信息。Optionally, the computing power service node can periodically obtain the computing power resource information of each computing power server connected to itself, and cache the computing power resource information of the computing power server locally. After receiving the multicast query message, It can quickly feed back the computing power resource information of the computing power server connected to itself to the computing power edge node.

在一种实施方式中,组播查询报文用于请求查询指定服务类型的算力资源信息,相应地,在向目标算力边缘节点发送查询响应报文之前,该方法还包括:In one embodiment, the multicast query message is used to request to query computing power resource information of a specified service type, and accordingly, before sending the query response message to the target computing power edge node, the method further includes:

从连接的算力服务器中,筛选出支持指定服务类型的指定算力服务器,获取指定算力服务器的算力资源信息。进而,算力服务节点发送的查询响应报文中包括该指定算力服务器的资源信息。From the connected computing power servers, filter out the designated computing power servers that support the specified service type, and obtain the computing power resource information of the designated computing power server. Furthermore, the query response message sent by the computing power service node includes the resource information of the designated computing power server.

在另一实施方式中,在向目标算力边缘节点发送查询响应报文之前,算力服务节点可以从连接的服务器中,筛选出支持指定服务类型且剩余算力资源满足预设条件的指定算力服务器,获取该指定算力服务器的算力资源信息。进而,算力服务节点发送的查询响应报文中包括支持指定服务类型,且剩余算力资源满足预设条件的指定算力服务器的资源信息。In another embodiment, before sending the query response message to the target computing power edge node, the computing power service node can filter out the specified computing power resources that support the specified service type and the remaining computing power resources meet the preset conditions from the connected servers. power server to obtain the computing power resource information of the specified computing power server. Furthermore, the query response message sent by the computing power service node includes the resource information of the designated computing power server that supports the specified service type and whose remaining computing power resources meet the preset conditions.

由于一个算力服务节点可能连接多个算力服务器,所以算力服务节点不仅需要从多个算力服务器中筛选支持指定服务类型的算力服务器,还需进一步判断筛选出的算力服务器的剩余算力资源是否满足预设条件。该预设条件基于实际需求设置,例如,预设条件可以为剩余算力资源中的CPU利用率小于预设的CPU利用率阈值,且剩余内存空间大于内存空间阈值。Since a computing power service node may be connected to multiple computing power servers, the computing power service node not only needs to select the computing power servers that support the specified service type from multiple computing power servers, but also needs to further judge the remaining power of the selected computing power servers. Whether the computing resources meet the preset conditions. The preset condition is set based on actual requirements. For example, the preset condition may be that the CPU utilization rate in the remaining computing resources is less than a preset CPU utilization rate threshold, and the remaining memory space is greater than the memory space threshold value.

采用上述方法,筛选出的算力服务器有足够的算力资源,可以避免向算力边缘节点反馈剩余算力资源不足的算力服务器的资源信息,可以节省传输开销,且可提高算力边缘节点选择目标服务器的效率,使得算力边缘节点可以更快速地处理算力服务请求报文,降低算力服务的时延。Using the above method, the selected computing power servers have sufficient computing power resources, which can avoid feeding back resource information of computing power servers with insufficient remaining computing power resources to computing power edge nodes, can save transmission costs, and can improve computing power edge nodes. Selecting the efficiency of the target server enables computing power edge nodes to process computing power service request messages more quickly and reduce the latency of computing power services.

基于相同的发明构思,本申请实施例还提供一种算力服务处理装置,该装置应用于算力服务处理系统中的任一算力边缘节点,算力服务处理系统中还包括至少一个算力服务节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器。如图8所示,该装置包括:Based on the same inventive concept, the embodiment of the present application also provides a computing power service processing device, which is applied to any computing power edge node in the computing power service processing system, and the computing power service processing system also includes at least one computing power Service nodes and at least one routing node, each computing power service node is connected to at least one computing power server. As shown in Figure 8, the device includes:

接收模块801,用于接收终端设备发送的算力服务请求报文;A receiving module 801, configured to receive a computing power service request message sent by a terminal device;

发送模块802,用于向算力边缘节点作为组播源的组播组中的算力服务节点发送组播查询报文,组播组对应的组播树中的叶子节点均为算力服务节点,非叶子节点为路由节点,根节点为算力边缘节点;The sending module 802 is configured to send a multicast query message to a computing power service node in a multicast group whose computing power edge node is the multicast source, and the leaf nodes in the multicast tree corresponding to the multicast group are all computing power service nodes , the non-leaf nodes are routing nodes, and the root node is a computing power edge node;

接收模块801,还用于接收组播组中的算力服务节点发送的查询响应报文,查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息;The receiving module 801 is also configured to receive a query response message sent by a computing power service node in the multicast group, and the query response message includes computing power resource information of a computing power server connected to the computing power service node;

选择模块803,还用于基于接收模块801接收到的各算力服务器的算力资源信息,选择一个目标算力服务器;The selection module 803 is further configured to select a target computing power server based on the computing power resource information of each computing power server received by the receiving module 801;

发送模块802,还用于向选择模块803选择的目标算力服务器转发算力服务请求报文。The sending module 802 is further configured to forward the computing power service request message to the target computing power server selected by the selecting module 803 .

可选的,发送模块802,具体用于:Optionally, the sending module 802 is specifically used for:

从多个组播组中,查找支持指定服务类型的指定组播组;Find a specified multicast group that supports a specified service type from multiple multicast groups;

向指定组播组中的算力服务节点发送组播查询报文。Send a multicast query message to the computing service nodes in the specified multicast group.

基于相同的发明构思,本申请实施例还提供一种算力服务处理装置,应用于算力服务处理系统中的任一算力服务节点,算力服务处理系统中还包括至少一个算力边缘节点和至少一个路由节点,每个算力服务节点连接于至少一个算力服务器;算力服务节点为组播组对应的组播树的叶子节点,组播树的非叶子节点为路由节点,根节点为算力边缘节点。如图9所示,该装置包括:Based on the same inventive concept, the embodiment of the present application also provides a computing power service processing device, which is applied to any computing power service node in the computing power service processing system, and the computing power service processing system also includes at least one computing power edge node and at least one routing node, each computing power service node is connected to at least one computing power server; the computing power service node is the leaf node of the multicast tree corresponding to the multicast group, the non-leaf node of the multicast tree is the routing node, and the root node It is the computing power edge node. As shown in Figure 9, the device includes:

接收模块901,用于接收目标算力边缘节点发送的组播查询报文,组播查询报文为目标算力边缘节点接收到终端设备的算力服务请求报文后发送的;The receiving module 901 is configured to receive a multicast query message sent by the target computing power edge node, and the multicast query message is sent after the target computing power edge node receives the computing power service request message of the terminal device;

发送模块902,用于响应于组播查询报文,向目标算力边缘节点发送查询响应报文,查询响应报文中包括算力服务节点连接的算力服务器的算力资源信息,以使得目标算力边缘节点基于接收到的各算力服务器的算力资源信息,选择一个目标算力服务器,向目标算力服务器转发算力服务请求报文。The sending module 902 is configured to send a query response message to the target computing power edge node in response to the multicast query message, and the query response message includes the computing power resource information of the computing power server connected to the computing power service node, so that the target The computing power edge node selects a target computing power server based on the received computing power resource information of each computing power server, and forwards the computing power service request message to the target computing power server.

可选的,组播查询报文用于请求查询指定服务类型的算力资源信息,查询响应报文中包括支持指定服务类型的指定算力服务器的算力资源信息;Optionally, the multicast query message is used to request to query the computing power resource information of the specified service type, and the query response message includes the computing power resource information of the specified computing power server supporting the specified service type;

该装置还包括:The unit also includes:

获取模块,用于从连接的算力服务器中,筛选出支持指定服务类型的指定算力服务器,获取指定算力服务器的算力资源信息。The obtaining module is used to filter out the designated computing power server that supports the specified service type from the connected computing power servers, and obtain the computing power resource information of the designated computing power server.

可选的,组播查询报文用于请求查询指定服务类型的算力资源信息,查询响应报文中包括支持指定服务类型,且剩余算力资源满足预设条件的指定算力服务器的资源信息;Optionally, the multicast query message is used to request to query the computing power resource information of the specified service type, and the query response message includes the resource information of the specified computing power server that supports the specified service type and the remaining computing power resources meet the preset conditions ;

该装置还包括:The unit also includes:

获取模块,用于从自身连接的服务器中,筛选出支持指定服务类型且剩余算力资源满足预设条件的指定算力服务器,获取指定算力服务器的算力资源信息。The acquisition module is used to filter out the specified computing power server that supports the specified service type and the remaining computing power resources meet the preset conditions from the servers connected to itself, and obtain the computing power resource information of the specified computing power server.

本申请实施例还提供了一种网络设备,该网络设备可以为上述实施例中的算力边缘节点或算力服务节点,如图10所示,包括处理器1001、通信接口1002、存储器1003和通信总线1004,其中,处理器1001,通信接口1002,存储器1003通过通信总线1004完成相互间的通信,The embodiment of the present application also provides a network device, which can be the computing power edge node or computing power service node in the above embodiments, as shown in FIG. 10 , including a processor 1001, a communication interface 1002, a memory 1003 and A communication bus 1004, wherein, the processor 1001, the communication interface 1002, and the memory 1003 complete mutual communication through the communication bus 1004,

存储器1003,用于存放计算机程序;Memory 1003, for storing computer programs;

处理器1001,用于执行存储器1003上所存放的程序时,实现上述方法实施例中由算力服务节点执行的方法步骤,或者实现上述方法实施例中由算力边缘节点执行的方法步骤。The processor 1001 is configured to implement the method steps performed by the computing power service node in the above method embodiments when executing the program stored in the memory 1003, or realize the method steps performed by the computing power edge node in the above method embodiments.

上述网络设备提到的通信总线可以是外设部件互连标准(Peripheral ComponentInterconnect,PCI)总线或扩展工业标准结构(Extended Industry StandardArchitecture,EISA)总线等。该通信总线可以分为地址总线、数据总线、控制总线等。为便于表示,图中仅用一条粗线表示,但并不表示仅有一根总线或一种类型的总线。The communication bus mentioned in the above network device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, EISA) bus or the like. The communication bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

通信接口用于上述电子设备与其他设备之间的通信。The communication interface is used for communication between the electronic device and other devices.

存储器可以包括随机存取存储器(Random Access Memory,RAM),也可以包括非易失性存储器(Non-Volatile Memory,NVM),例如至少一个磁盘存储器。可选的,存储器还可以是至少一个位于远离前述处理器的存储装置。The memory may include a random access memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-Volatile Memory, NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located far away from the aforementioned processor.

上述的处理器可以是通用处理器,包括中央处理器(Central Processing Unit,CPU)、网络处理器(Network Processor,NP)等;还可以是数字信号处理器(Digital SignalProcessor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现场可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。Above-mentioned processor can be general-purpose processor, comprises central processing unit (Central Processing Unit, CPU), network processor (Network Processor, NP) etc.; It can also be Digital Signal Processor (Digital Signal Processor, DSP), ASIC (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

在本发明提供的又一实施例中,还提供了一种计算机可读存储介质,该计算机可读存储介质内存储有计算机程序,计算机程序被处理器执行时实现上述任一算力服务处理方法的步骤。In yet another embodiment provided by the present invention, a computer-readable storage medium is also provided. A computer program is stored in the computer-readable storage medium. When the computer program is executed by a processor, any of the above-mentioned computing power service processing methods is implemented. A step of.

在本发明提供的又一实施例中,还提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述实施例中任一算力服务处理方法。In yet another embodiment provided by the present invention, a computer program product including instructions is also provided, which, when run on a computer, causes the computer to execute any computing power service processing method in the above embodiments.

在上述实施例中,可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时,可以全部或部分地以计算机程序产品的形式实现。计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行计算机程序指令时,全部或部分地产生按照本发明实施例的流程或功能。计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。计算机指令可以存储在计算机可读存储介质中,或者从一个计算机可读存储介质向另一个计算机可读存储介质传输,例如,计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。可用介质可以是磁性介质,(例如,软盘、硬盘、磁带)、光介质(例如,DVD)、或者半导体介质(例如固态硬盘Solid State Disk(SSD))等。In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present invention are produced in whole or in part. A computer can be a general purpose computer, special purpose computer, computer network, or other programmable device. Computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, e.g. Coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server, a data center, etc. integrated with one or more available media. The available media may be magnetic media (eg, floppy disk, hard disk, magnetic tape), optical media (eg, DVD), or semiconductor media (eg, Solid State Disk (SSD)).

需要说明的是,在本文中,诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来,而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括要素的过程、方法、物品或者设备中还存在另外的相同要素。It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is a relationship between these entities or operations. There is no such actual relationship or order between them. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements of or also include elements inherent in such a process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

本说明书中的各个实施例均采用相关的方式描述,各个实施例之间相同相似的部分互相参见即可,每个实施例重点说明的都是与其他实施例的不同之处。尤其,对于方法实施例而言,由于其基本相似于系统实施例,所以描述的比较简单,相关之处参见系统实施例的部分说明即可。Each embodiment in this specification is described in a related manner, the same and similar parts of each embodiment can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, as for the method embodiment, since it is basically similar to the system embodiment, the description is relatively simple, and for the related parts, please refer to the part of the description of the system embodiment.

以上仅为本发明的较佳实施例,并非用于限定本发明的保护范围。凡在本发明的精神和原则之内所作的任何修改、等同替换、改进等,均包含在本发明的保护范围内。The above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims (13)

1. A computing power service processing system, comprising at least one computing power service node, at least one computing power edge node, and at least one routing node, each computing power service node connected to at least one computing power server;
each computational power edge node is used for receiving a computational power service request message sent by terminal equipment and sending a multicast query message to the computational power service nodes in multicast groups which are taken as multicast sources, wherein leaf nodes in a multicast tree corresponding to each multicast group are computational power service nodes, non-leaf nodes are routing nodes, and root nodes are computational power edge nodes;
each computing power service node is used for responding to a target multicast query message from a target computing power edge node and sending a query response message to the target computing power edge node, wherein the query response message comprises computing power resource information of a computing power server connected with the computing power service node;
each computing force edge node is also used for selecting a target computing force server based on the received computing force resource information of each computing force server and forwarding a computing force service request message to the target computing force server.
2. The system according to claim 1, wherein the computing power service request message is used for requesting a computing power service of a specified service type;
each computing force edge node is specifically configured to, after receiving a computing force service request message, search for an assigned multicast group supporting the assigned service type from multiple multicast groups that serve as multicast sources, and send a multicast query message to a computing force service node in the assigned multicast group.
3. The system according to claim 1 or 2, wherein the target multicast query message is used for requesting to query computing resource information of a specified service type;
each computational power service node is specifically configured to screen out, in response to a received target multicast query message, a designated computational power server supporting the designated service type from connected computational power servers, and send the query response message to the target computational power edge node, where the query response message includes computational power resource information of the designated computational power server.
4. The system according to claim 1 or 2, wherein the multicast query message is used for requesting to query computing resource information of a specified service type;
each computational power service node is specifically configured to screen out, in response to a received target multicast query message, a specified computational power server that supports the specified service type and for which the remaining computational power resources meet preset conditions from connected computational power servers, and send the query response message to the target computational power edge node, where the query response message includes computational power resource information of the specified computational power server.
5. The system of claim 1,
each computing power service node is also used for periodically acquiring computing power resource information of each computing power server connected with the computing power service node.
6. The system of claim 1, further comprising an aggregation node;
each computing service node is also used for sending registration information to the aggregation node when the computing service node is accessed into the computing service processing system, wherein the registration information comprises the service type of a computing server connected with the computing service node;
and the sink node is used for responding to target registration information from a target computing power service node, adding the target computing power service node into a target multicast group matched with the service type carried by the target registration information based on the service type carried by the target registration information, and updating a multicast tree corresponding to the target multicast group.
7. A computing power service processing method is characterized in that the method is applied to any computing power edge node in a computing power service processing system, the computing power service processing system also comprises at least one computing power service node and at least one routing node, and each computing power service node is connected with at least one computing power server; the method comprises the following steps:
receiving a computing power service request message sent by terminal equipment;
sending a multicast query message to a computational power service node in a multicast group which is taken as a multicast source, wherein leaf nodes in a multicast tree corresponding to the multicast group are computational power service nodes, non-leaf nodes are routing nodes, and a root node is a computational power edge node;
receiving an inquiry response message sent by a computing service node in the multicast group, wherein the inquiry response message comprises computing resource information of a computing server connected with the computing service node;
and selecting a target computing power server based on the received computing power resource information of each computing power server, and forwarding the computing power service request message to the target computing power server.
8. The method of claim 7, wherein sending the multicast query message to the computation force service node in the multicast group comprises:
searching a specified multicast group supporting the specified service type from a plurality of multicast groups;
and sending the multicast query message to the computing service node in the specified multicast group.
9. A computing power service processing method is characterized in that the method is applied to any computing power service node in a computing power service processing system, the computing power service processing system further comprises at least one computing power edge node and at least one routing node, and each computing power service node is connected to at least one computing power server; the calculation force service node is a leaf node of a multicast tree corresponding to a multicast group, a non-leaf node of the multicast tree is a routing node, and a root node is a calculation force edge node; the method comprises the following steps:
receiving a multicast query message sent by a target computing power edge node, wherein the multicast query message is sent by the target computing power edge node after receiving a computing power service request message of a terminal device;
responding to the multicast query message, sending a query response message to the target computing power edge node, wherein the query response message comprises computing power resource information of the computing power servers connected with the computing power service node, so that the target computing power edge node selects one target computing power server based on the received computing power resource information of each computing power server, and forwards the computing power service request message to the target computing power server.
10. The method according to claim 9, wherein the multicast query message is used to request query for computing resource information of a specified service type, and the query response message includes computing resource information of a specified computing server supporting the specified service type;
before the sending of the query response message to the target computing force edge node, the method further includes:
and screening out the appointed force calculation server supporting the appointed service type from the connected force calculation servers, and acquiring the force calculation resource information of the appointed force calculation server.
11. The method according to claim 9, wherein the multicast query message is used to request query for computing power resource information of a specified service type, and the query response message includes resource information of a specified computing power server that supports the specified service type and for which the remaining computing power resources meet a preset condition;
before the sending of the query response message to the target computing force edge node, the method further includes:
and screening out a designated computing power server which supports the designated service type and has residual computing power resources meeting preset conditions from the server connected with the designated computing power server, and acquiring computing power resource information of the designated computing power server.
12. The computing power service processing device is applied to any computing power edge node in a computing power service processing system, the computing power service processing system also comprises at least one computing power service node and at least one routing node, and each computing power service node is connected with at least one computing power server; the device comprises:
the receiving module is used for receiving a computing power service request message sent by the terminal equipment;
a sending module, configured to send a multicast query packet to a computation force service node in a multicast group in which the computation force edge node serves as a multicast source, where leaf nodes in a multicast tree corresponding to the multicast group are all computation force service nodes, non-leaf nodes are routing nodes, and a root node is the computation force edge node;
the receiving module is further configured to receive a query response packet sent by a computational power service node in the multicast group, where the query response packet includes computational power resource information of a computational power server connected to the computational power service node;
the selection module is used for selecting a target calculation force server based on the calculation force resource information of each calculation force server received by the receiving module;
the sending module is further configured to forward the computing power service request message to the target computing power server selected by the selecting module.
13. The computing power service processing device is applied to any computing power service node in a computing power service processing system, the computing power service processing system also comprises at least one computing power edge node and at least one routing node, and each computing power service node is connected with at least one computing power server; the calculation force service node is a leaf node of a multicast tree corresponding to a multicast group, a non-leaf node of the multicast tree is a routing node, and a root node is a calculation force edge node; the device comprises:
the receiving module is used for receiving a multicast query message sent by a target computing power edge node, wherein the multicast query message is sent after the target computing power edge node receives a computing power service request message of a terminal device;
and the sending module is used for responding to the multicast query message and sending a query response message to the target computing power edge node, wherein the query response message comprises computing power resource information of the computing power servers connected with the computing power service node, so that the target computing power edge node selects one target computing power server based on the received computing power resource information of each computing power server and forwards the computing power service request message to the target computing power server.
CN202211349514.9A 2022-10-31 2022-10-31 Computing service processing method, device and system Active CN115622938B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211349514.9A CN115622938B (en) 2022-10-31 2022-10-31 Computing service processing method, device and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211349514.9A CN115622938B (en) 2022-10-31 2022-10-31 Computing service processing method, device and system

Publications (2)

Publication Number Publication Date
CN115622938A true CN115622938A (en) 2023-01-17
CN115622938B CN115622938B (en) 2025-02-21

Family

ID=84875644

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211349514.9A Active CN115622938B (en) 2022-10-31 2022-10-31 Computing service processing method, device and system

Country Status (1)

Country Link
CN (1) CN115622938B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118264762A (en) * 2024-05-29 2024-06-28 浙江宇视科技有限公司 GPU cluster scheduling method, system and equipment for intelligent service
WO2024188315A1 (en) * 2023-03-16 2024-09-19 中国移动通信有限公司研究院 Computing power information notification method and apparatus, and network edge node and storage medium
WO2024222006A1 (en) * 2023-04-24 2024-10-31 中兴通讯股份有限公司 Service request response method, service data acquisition method and electronic device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110532092A (en) * 2019-08-21 2019-12-03 云湾科技(嘉兴)有限公司 Reso urce matching method, device and equipment
WO2021208914A1 (en) * 2020-04-15 2021-10-21 展讯半导体(南京)有限公司 Network-scheduling-based computing power sharing method, and related product
CN114095577A (en) * 2020-07-31 2022-02-25 中国移动通信有限公司研究院 Resource request method, device, computing power network element node and computing power application device
CN114978978A (en) * 2022-06-07 2022-08-30 中国电信股份有限公司 Computing resource scheduling method and device, electronic equipment and medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110532092A (en) * 2019-08-21 2019-12-03 云湾科技(嘉兴)有限公司 Reso urce matching method, device and equipment
WO2021208914A1 (en) * 2020-04-15 2021-10-21 展讯半导体(南京)有限公司 Network-scheduling-based computing power sharing method, and related product
CN114095577A (en) * 2020-07-31 2022-02-25 中国移动通信有限公司研究院 Resource request method, device, computing power network element node and computing power application device
CN114978978A (en) * 2022-06-07 2022-08-30 中国电信股份有限公司 Computing resource scheduling method and device, electronic equipment and medium

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024188315A1 (en) * 2023-03-16 2024-09-19 中国移动通信有限公司研究院 Computing power information notification method and apparatus, and network edge node and storage medium
WO2024222006A1 (en) * 2023-04-24 2024-10-31 中兴通讯股份有限公司 Service request response method, service data acquisition method and electronic device
CN118264762A (en) * 2024-05-29 2024-06-28 浙江宇视科技有限公司 GPU cluster scheduling method, system and equipment for intelligent service

Also Published As

Publication number Publication date
CN115622938B (en) 2025-02-21

Similar Documents

Publication Publication Date Title
CN115622938A (en) Computing power service processing method, device and system
CN108833166B (en) Edge cloud message forwarding method and system, network message forwarding method and system
CN104335537B (en) Systems and methods for layer 2 multicast multipath delivery
Li et al. ESM: Efficient and scalable data center multicast routing
US20160212066A1 (en) Software-Defined Information Centric Network (ICN)
WO2021218718A1 (en) Method and device for determining function and information provision of user plane, and medium
CN104509073A (en) Discovering ip multicast group memberships in software defined networks
CN1475063A (en) Subgroup Multicast in Communication Networks
US8966107B2 (en) System and method of streaming data over a distributed infrastructure
WO2010127618A1 (en) System and method for implementing streaming media content service
WO2021237827A1 (en) Method, apparatus and device for pushing video stream, and storage medium
US20190280968A1 (en) Multicasting system
US20120185487A1 (en) Method, device and system for publication and acquisition of content
CN109495526A (en) A kind of file transmitting method, device, system, electronic equipment and storage medium
CN106899641A (en) The acquisition methods of data resource, server, control device and communication system
US11223697B2 (en) Scaling microservices communication over information centric networks
CN115665262A (en) A request processing method, device, electronic device and storage medium
CN101267449A (en) A tree structure P2P system resource transmission method based on mobile agent mechanism
US8102846B2 (en) Method and apparatus for managing a multicast tree using a multicast tree manager and a content server
US10938591B2 (en) Multicast system
CN114124890B (en) A determination method, virtual router, control device and domain name resolution system
CN113545130B (en) Fast roaming and unified policies for wireless clients using distributed hashing
Wang et al. Solving multicast problem in cloud networks using overlay routing
Zarrin et al. A self-organizing and self-configuration algorithm for resource management in service-oriented systems
CN111866759A (en) Group communication method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant