A computing power scheduling method and device of a computing power network

Abstract

Embodiments of the present application provide a computing power scheduling method and device of a computing power network, applied to a distributed database, the method comprising: receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device (PE); publishing the node computing power state information and the link communication state information to an edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user device. By applying the method, modification of a routing protocol is avoided, and the applicability of the computing power scheduling method and the stability of the computing power network are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a computing power scheduling method and device of computing power network. BACKGROUND

[0002] At present, 5G (5th Generation Mobile Communication Technology), artificial intelligence, big data and other new generation information technologies are reshaping the world. Facing the huge market of industry digital transformation, various industries have put forward more urgent demands for computing power and network. The next generation network represented by computing power network is gradually becoming the focus of attention of all parties in the industry.

[0003] Computing power network is a new type of information infrastructure that allocates and flexibly schedules computing resources, storage resources and network resources according to business needs among cloud, network and edge.

[0004] For distributed computing power network, in the existing computing power scheduling scheme, the provider edge device PE diffuses node computing power state information and link communication state information to the neighbor provider edge device PE (Provider Edge) through the Border Gateway Protocol (BGP), and the provider edge device PE comprehensively considers the node computing power state information and the link communication state information to select appropriate computing power resource pool for the user and determine the corresponding message transmission path.

[0005] However, by diffusing node computing power state information and link communication state information through the protocol, the current routing protocol needs to be modified, but the underlying physical network is a stock network in many cases, rather than a newly built network, therefore, the method of modifying the routing protocol to realize computing power scheduling will cause great impact on the existing network. SUMMARY

[0006] The purpose of the embodiment of the present application is to provide a computing power scheduling method and device of computing power network, so as to realize computing power scheduling without modifying the underlying routing protocol, and improve the applicability of the computing power scheduling method and the stability of the computing power network.

[0007] The specific technical solutions are as follows:

[0008] The embodiment of the present application provides a computing power scheduling method of computing power network, applied to a distributed database, and the method comprises:

[0009] Receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE;

[0010] publish the node computing power state information and the link communication state information to the edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user equipment.

[0011] Optionally, the edge device includes a provider edge device PE and / or a user edge device CE.

[0012] Optionally, the distributed database pre-establishes a communication mode based on publication and subscription with at least one edge device in the computing power network.

[0013] After receiving the node computing power state information announced by each computing power node in the computing power network and the link communication state information announced by each provider edge device PE, the method further includes:

[0014] integrating the node computing power state information to obtain network computing power state information of the computing power network, and integrating the link communication state information to obtain network communication state information of the computing power network;

[0015] The publishing of the node computing power state information and the link communication state information to the edge device specifically includes:

[0016] When a change in the network computing power state information and / or the network communication state information is detected, the changed network computing power state information and / or network communication state information is published to a corresponding information channel, so that the edge device pulls pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0017] Optionally, each database node of the distributed database is respectively deployed on a different server.

[0018] Or,

[0019] Each database node of the distributed database is respectively deployed on a different provider edge device PE.

[0020] Optionally, the storage node in the distributed database stores the node computing power state information and / or the link communication state information in the form of a data table.

[0021] The data table includes a key-value pair, wherein the key is a field representing a device category, a device identifier, and / or a device computing power state attribute and a device-to-device communication state attribute of a device in the computing power network, and the value is a value corresponding to the key.

[0022] Optionally, the distributed database includes a non-primary storage node and a primary storage node determined through an election mechanism.

[0023] The master storage node is configured to receive node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device (PE) and synchronize the received data to the non-master storage node.

[0024] When a node in the distributed database fails, a master storage node is determined again through an election mechanism, and the determined master storage node contains full data.

[0025] When the failed node re-joins the distributed database, the master storage node synchronizes the full data stored by itself to the re-joined node.

[0026] Embodiments of the present application also provide a computing power scheduling device of a computing power network, applied to a distributed database, and the device comprises:

[0027] A receiving module is configured to receive node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device (PE).

[0028] A publishing module is configured to publish the node computing power state information and the link communication state information to an edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user equipment.

[0029] Optionally, the edge device comprises a provider edge device (PE) and / or a user edge device (CE).

[0030] Optionally, the distributed database pre-establishes a communication mode based on publishing and subscribing with at least one edge device in the computing power network.

[0031] The device further comprises:

[0032] An integrating module is configured to integrate the node computing power state information to obtain network computing power state information of the computing power network and integrate the link communication state information to obtain network communication state information of the computing power network.

[0033] The publishing module is specifically configured to:

[0034] When it is detected that the network computing power state information and / or the network communication state information changes, the changed network computing power state information and / or network communication state information is published to a corresponding information channel, so that the edge device pulls pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0035] Optionally, each database node of the distributed database is respectively deployed on a different server.

[0036] Or,

[0037] Each database node of the distributed database is respectively deployed on a different provider edge device PE.

[0038] Optionally, the storage node in the distributed database stores the node computing power state information and / or the link communication state information in the form of a data table.

[0039] The data table includes key-value pairs, wherein key is a field representing device category, device identification, device computing power state attribute, and / or device intercommunication state attribute in the computing power network; and value is a value corresponding to key.

[0040] Optionally, the distributed database includes a non-primary storage node and a primary storage node determined through an election mechanism.

[0041] The primary storage node is configured to receive node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE, and synchronize the received data to the non-primary storage node.

[0042] The apparatus further includes:

[0043] The selection module is configured to determine a primary storage node through an election mechanism again when a node in the distributed database fails, and the determined primary storage node includes full data.

[0044] The synchronization module is configured to synchronize the full data stored in the primary storage node to the re-joined node when the failed node re-joins the distributed database.

[0045] The embodiment of the application has the following advantages:

[0046] The computing power scheduling method and apparatus for the computing power network provided by the embodiment of the application introduce a distributed database on the basis of the original computing power network architecture, receive node computing power state information and link communication state information in the computing power network by the distributed database, and publish the information to edge devices, so that the edge devices determine target computing power nodes and packet forwarding paths based on the node computing power state information and the link communication state information, and implement computing power scheduling. Therefore, the node computing power state information and the link communication state information do not need to be transmitted between provider edge devices PE, and the underlying routing protocol between the provider edge devices PE does not need to be modified, that is, the existing network is not impacted, and the applicability of the computing power scheduling method and the stability of the computing power network are improved.

[0047] Of course, practicing any of the products or methods of the present application does not necessarily require that all of the advantages described above be achieved simultaneously. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other embodiments can also be obtained by those skilled in the art based on these drawings.

[0049] Figure 1 is a schematic diagram of a computing power scheduling method based on a BGP protocol in the related art;

[0050] Figure 2 is a flowchart of a first embodiment of a computing power scheduling method of a computing power network provided by the embodiments of the present application;

[0051] Figure 3 is a flowchart of a second embodiment of a computing power sensing method of a computing power network provided by the embodiments of the present application;

[0052] Figure 4 is an example diagram of a computing power routing table provided by the embodiments of the present application;

[0053] Figure 5 is an example diagram of a data table provided by the embodiments of the present application;

[0054] Figure 6 is an example diagram of a computing power sensing method of a computing power network provided by the embodiments of the present application;

[0055] Figure 7 is a flowchart of a third embodiment of a computing power sensing method of a computing power network provided by the embodiments of the present application;

[0056] Figure 8 is another example diagram of a computing power sensing method of a computing power network provided by the embodiments of the present application;

[0057] Figure 9 is an example diagram of etcd provided by the embodiments of the present application;

[0058] Figure 10 is an example diagram of database node deployment provided by the embodiments of the present application;

[0059] Figure 11 is a structural schematic diagram of a computing power scheduling device of a computing power network provided by the embodiments of the present application;

[0060] Figure 12 is a structural schematic diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art based on the present application belong to the scope of protection of the present application.

[0062] For the convenience of description, the nouns and terms involved in one or more embodiments of the present application are explained first.

[0063] Computing power: the computing power of a device.

[0064] Computing power network: a new type of information infrastructure that allocates and flexibly schedules computing resources, storage resources and network resources according to business needs among cloud, edge and end.

[0065] Cloud: cloud computing, a supercomputing mode based on a network "cloud". In a remote data center, thousands of computers and servers are connected to form a computing cloud.

[0066] Edge: edge, referring to Multi-access Edge Computing (MEC), a network architecture that provides cloud computing functions and Information Technology (IT) service environment at the edge of the network.

[0067] End: end side, referring to terminal, i.e. all devices with Internet and computing capabilities.

[0068] With the digital transformation of various industries, the intelligent world of "Internet of Everything, Everything Sensing and Everything Intelligent" is accelerating. In the era of everything intelligent, data is a production factor and computing power is productivity.

[0069] The emergence of emerging applications such as autonomous driving, cloud gaming, facial recognition, Virtual Reality (VR) and Augmented Reality (AR) has led to rapid growth in computing power demand. The computing power of traditional single-terminal devices can no longer meet business needs, so the computing power network has emerged to meet the computing power needs brought about by digital development.

[0070] In simple terms, when a user makes a business request, the computing power network can perceive the network state and the computing power state of the cloud to provide integrated services for the user.

[0071] In the distributed computing power network, each router node in the computing power network maintains node computing power state information and link communication state information, and each router node comprehensively considers the node computing power state information and the link communication state information to perform routing, so as to provide suitable computing power services for users.

[0072] However, the current computing power scheduling method generally publishes the node computing power state information and the link communication state information by extending the existing protocol. Figure 1 is a schematic diagram of a computing power scheduling method based on a BGP protocol in the related art, Figure 1 In the MEC edge cloud shown in the figure, after the user initiates a computing power request, the computing power network determines a computing power node corresponding to the user demand for the user, and determines a path for forwarding a message to the node. Figure 1 In the MEC edge cloud shown in the figure, after the user initiates a computing power request, the computing power network determines a computing power node corresponding to the user demand for the user, and determines a path for forwarding a message to the node.

[0073] In combination with Figure 1 The process of the method is as follows:

[0074] The computing power node advertises its own node computing power state information to the accessed provider edge device PE;

[0075] The provider edge device PE diffuses the node computing power state information and the link communication state information to its BGP neighbors;

[0076] The head node router receives the downstream advertised node computing power state information and link communication state information;

[0077] The user initiates a computing power request;

[0078] The edge router selects a corresponding computing power node and a corresponding message forwarding path for the user according to the node computing power state information and the link communication state information.

[0079] However, the above computing power scheduling method based on the BGP protocol has the following disadvantages:

[0080] In order to realize the diffusion of the node computing power state information and the link communication state information, the current routing protocol needs to be modified, which has a great impact on the existing network. Although the computing power network is a new type of information infrastructure, the underlying physical network is still a stock network rather than a newly built network in many cases, so this scheme will have a great impact on the existing routing protocol of the existing network.

[0081] To solve the above technical problems, an embodiment of the present application provides a computing power scheduling method of a computing power network, as shown in Figure 2 , Figure 2is a flowchart of a first embodiment of a computing power scheduling method of a computing power network provided by the embodiment of the application, and the computing power scheduling method provided by the embodiment of the application is specifically executed by a distributed database, and reference is made to Figure 2 The method comprises the following steps.

[0082] Step S201: receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE.

[0083] Step S202: publishing the node computing power state information and the link communication state information to the edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user device.

[0084] The computing power scheduling method of the computing power network provided by the embodiment of the application introduces a distributed database on the basis of the original computing power network architecture, receives node computing power state information and link communication state information in the computing power network by the distributed database, and publishes the node computing power state information and the link communication state information to the edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information, and the computing power scheduling is realized. Therefore, the node computing power state information and the link communication state information do not need to be transmitted between the provider edge devices PE, and the underlying routing protocol between the provider edge devices PE does not need to be modified, that is, the existing network is not impacted, and the applicability of the computing power scheduling method and the stability of the computing power network are improved.

[0085] In an embodiment of the application, the edge device comprises a provider edge device PE and / or a user edge device CE (Customer Edge).

[0086] Next, the case that the edge device is a provider edge device PE and the case that the edge device is a user edge device CE are described respectively in combination with specific embodiments.

[0087] Figure 3 is a flowchart of a second embodiment of a computing power sensing method of a computing power network provided by the embodiment of the application, in combination with Figure 3 The method comprises the following steps.

[0088] Step S301: receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE.

[0089] The computing power node in the embodiment of the application specifically refers to computing power resources connected by each edge router in the computing power network. In an embodiment of the application, the computing power node refers to an edge cloud.

[0090] The node computing power state information is specifically used for representing the computing power resource condition of each computing power node.

[0091] In an embodiment of the application, the node computing power state information includes, but is not limited to, the CPU (Central Processing Unit) usage rate of the computing power node, the IOPS (Input / Output Operations Per Second) value, and the like.

[0092] Taking the edge cloud as an example, in an embodiment, the edge cloud includes a computing power perception module, and the node computing power state information of the edge cloud is collected by the computing power perception module. In this case, the node computing power state information notified by the computing power node is received by the distributed database, and specifically, the node computing power state information notified by the computing power perception module of the edge cloud.

[0093] The link communication state information is specifically used for representing the communication state between the PEs.

[0094] In an embodiment of the application, the link communication state information includes, but is not limited to, the delay, the jitter, the bandwidth, the packet loss rate, and the like of the path.

[0095] It should be understood that the node computing power state information and the link communication state information involved in the embodiments of the application may be different in content at different times, and the embodiments of the application do not limit the number and frequency of receiving the node computing power state information and the link communication state information. In an embodiment of the application, the distributed database receives the node computing power state information and the link communication state information periodically based on a preset time period.

[0096] Step S302: publishing the node computing power state information and the link communication state information to the PEs, so that the PEs determine the target computing power node and the packet forwarding path based on the node computing power state information and the link communication state information in response to the computing power request of the user equipment.

[0097] The computing power request of the user equipment is specifically a request for scheduling computing power sent by the user equipment to the computing power network. The user equipment is a user terminal,

[0098] The target computing power node is specifically a computing power node finally determined to process the computing power request of the user equipment.

[0099] The message forwarding path specifically refers to path information of one or more provider edge devices PE in the forwarding path when the message is forwarded from the head node router to the provider edge device PE connected with the first computing power node. The head node router specifically refers to the provider edge device PE connected with the user edge device CE used when the user equipment accesses the computing power network. As an example, assuming that the head node router is A, the provider edge device PE connected with the target computing power node is B, and the provider edge devices C, D and E are sequentially included in the forwarding path during the forwarding of the message from the head node router A to B, the message forwarding path can be understood as A-C-D-E-B.

[0100] In an embodiment of the present application, the message forwarding path supports a specified message forwarding mode.

[0101] As an example, the message forwarding path is an SRv6 (Segment Routing IPv6, a segment routing based on an IPv6 forwarding plane) path.

[0102] In the publishing of the computing power state information and the link communication state information by the distributed database execution node, the embodiments of the present application do not limit the specific publishing to several provider edge devices PE or to which provider edge devices PE.

[0103] In an embodiment of the present application, the distributed database publishes the computing power state information and the link communication state information to the head node router.

[0104] The embodiments of the present application also do not limit the specific algorithm used when the target computing power node and the message forwarding path are determined based on the node computing power state information and the link communication state information.

[0105] In an embodiment of the present application, the node computing power state information and the link communication state information are used for the provider edge device PE to maintain a computing power routing table, and the target computing power node and the message forwarding path are determined according to the computing power routing table.

[0106] As an example, the computing power routing table includes a comprehensive computing power value of the next hop, and the comprehensive computing power value is specifically calculated by combining the node computing power state information and the link communication state information, and the embodiments of the present application do not limit the specific algorithm of the comprehensive computing power value. The next hop specifically refers to the next provider edge device PE in the message forwarding path relative to the current provider edge device PE and / or the computing power node connected with the next hop provider edge device PE.

[0107] Figure 4 is an example diagram of the computing power routing table provided by the embodiments of the present application, Figure 4Specifically, the computing power routing table maintained by the provider edge device PE marked as PE1, and the provider edge device PE1 is connected with the edge cloud marked as MEC1, and the like, Figure 4 The provider edge device PE2 is connected with the edge cloud MEC2, the provider edge device PE3 is connected with the edge cloud MEC3, and the provider edge device PE4 is connected with the edge cloud MEC4. The computing power routing table of the provider edge device PE1 includes the integrated computing power values of the edge clouds MEC2, MEC3 and MEC4 as the next hop, that is, the provider edge devices PE2, PE3 and PE4 as the next hop router, and the integrated computing power values are 60, 20 and 10 respectively. Therefore, according to the computing power resource table, the edge cloud MEC2 is determined as the next hop, that is, the message is forwarded from PE1 to PE2.

[0108] In order to facilitate understanding of the content of step S302, the following is described with a specific example. In an embodiment of the present application, the user initiates an access to an application, and the application is used to process a rendering task submitted by the user, but the task exceeds the computing capacity of the end side, so the user equipment sends a computing power request to the computing power network for completing the task. The distributed database publishes the node computing power state information and the link communication state information to the nodes of the provider edge device PE, so that the provider edge device PE responds to the computing power request and determines the target computing power node and the message forwarding path. Subsequently, the message is forwarded to the target computing power node along the message forwarding path, and after the target computing power node processes the rendering task submitted by the user, the processing result is returned to the user equipment.

[0109] The computing power scheduling method of the computing power network provided by the embodiment of the present application introduces a distributed database on the basis of the original computing power network architecture, receives the node computing power state information and the link communication state information in the computing power network by the distributed database, and publishes them to the edge device, so that the edge device determines the target computing power node and the message forwarding path based on the node computing power state information and the link communication state information, and realizes the computing power scheduling. Therefore, the node computing power state information and the link communication state information do not need to be transmitted between the provider edge devices PE, and the underlying routing protocol between the provider edge devices PE does not need to be modified, that is, the present network will not be impacted, and the applicability of the computing power scheduling method and the stability of the computing power network are improved. In an embodiment of the present application, the distributed database pre-establishes a communication mode based on publication and subscription with at least one edge device in the computing power network.

[0110] In this embodiment, after step S201 in the foregoing, the computing power scheduling method of the computing power network provided by the embodiment of the present application further includes:

[0111] The node computing power state information is integrated to obtain network computing power state information of the computing power network, and the link communication state information is integrated to obtain network communication state information of the computing power network.

[0112] The step S202 of publishing the node computing power state information and the link communication state information to the edge device specifically includes the following steps.

[0113] When it is detected that the network computing power state information and / or the network communication state information is changed, the changed network computing power state information and / or the network communication state information is published to the corresponding information channel, so that the edge device pulls the pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0114] In the second embodiment shown in the figure, the distributed database is pre-established with at least a provider edge device in the computing power network in a communication mode based on publishing and subscribing. Figure 3 On the basis of the second embodiment shown in the figure, that is, the distributed database is pre-established with at least a provider edge device in the computing power network in a communication mode based on publishing and subscribing. The communication mode based on publishing and subscribing is described below in combination with specific embodiments.

[0115] In an embodiment of the present application, after receiving the node computing power state information announced by each computing power node in the computing power network and the link communication state information announced by each provider edge device PE, before publishing the computing power node state information and the link communication state information to the provider edge device PE, that is, before step S301, after step S302, the method further includes the following steps.

[0116] The node computing power state information is integrated to obtain network computing power state information of the computing power network, and the link communication state information is integrated to obtain network communication state information of the computing power network.

[0117] It should be understood that when the first target computing power node and the first message forwarding path are determined according to the node computing power state information and the link communication state information, the first target computing power node and the first message forwarding path need to be determined from each computing power node and edge router in the computing power network, and the distributed database naturally needs to integrate the node computing power state information and the link communication state information in the entire computing power network.

[0118] In this communication mode based on publishing and subscribing, the step S302 of publishing the node computing power state information and the link communication state information to the provider edge device PE specifically can include the following steps.

[0119] When it is detected that the network computing power state information and / or the network communication state information is changed, the changed network computing power state information and / or the network communication state information is published to the corresponding information channel, so that the provider edge device PE pulls the pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0120] Specifically, the provider edge device PE can subscribe to part of the node computing power state information and link communication state information based on actual needs. After the subscription is completed, once the subscribed part of the node computing power state information and link communication state information is changed, the information will be published to the corresponding information channel, and the provider edge device PE can pull the pre-subscribed node computing power state information and / or link communication state information from the corresponding information channel.

[0121] In actual application scenarios, the provider edge device PE can not need to call the computing power resources of any computing power node in the computing power network, and the distributed database can naturally not need to publish the network computing power state information and network communication state information to it, but can publish information based on the communication mode of subscription publishing.

[0122] For example, when a certain computing power node and the corresponding provider edge device PE are too far away from the physical location of the head node router, calling this computing power node does not meet the computing efficiency requirement, and therefore the head node router does not need to subscribe to the corresponding node computing power state information and / or link communication state information.

[0123] The embodiment of the application determines which node computing power state information and / or link communication state information the distributed database publishes to the provider edge device PE through the communication mode of subscription publishing, reduces the data transmission amount of information published by the distributed database, and thus can improve the efficiency of computing power scheduling.

[0124] In an embodiment of the application, the storage node in the distributed database stores the node computing power state information and / or link communication state information in the form of a data table; the data table includes key-value pairs; wherein the key is a field representing the device category, device identifier, and / or device computing power state attribute, and inter-device communication state attribute of the device in the computing power network; and the value is a value corresponding to the key.

[0125] Here, the device specifically refers to an edge device and / or a computing power node.

[0126] As an example, the key of the key-value pair related to the node computing power state information is designed as a combination of the device category, device identifier, and device computing power state attribute, and the key of the key-value pair related to the link communication state information is designed as a combination of the device category, device identifier, and inter-device communication state attribute. As an example, Figure 5 is an example diagram of the data table provided by the embodiment of the application, referring to Figure 5 , Figure 5Two key-value pairs are shown. Among them, the Key in the first key-value pair is MEC / Edge Cloud 2 / Total Remaining vCPU (virtual Central Processing Unit) number, and the Value is 100, which represents that Edge Cloud 2 still has 100 vCPUs that can provide computing power services to the outside. The Key of the second key-value pair is PE / PE2 / node IP (Internet Protocol), and the Value is 2.2.2.2, which represents that the IP of PE2 is 2.2.2.2.

[0127] For example, in combination with the data table in Figure 5 and the foregoing communication mode based on subscription and publication, the node computing power state information and / or link communication state information that the provider edge device PE pre-subscribes can be understood as the key that the provider edge device PE is interested in being determined in advance. When the distributed database detects that the value of the corresponding key-value pair changes, the distributed database will send the changed key-value pair to the corresponding information channel based on the subscription of the provider edge device PE.

[0128] The embodiment of the application can simplify the data stored in the distributed database by storing the node computing power state information and / or link communication state information in the data table including the key-value pair.

[0129] Figure 6 is an example diagram of the computing power awareness method of the computing power network provided by the embodiment of the application. In order to facilitate the understanding of the second embodiment in the foregoing, the following will be combined with Figure 6 The computing power awareness method of the computing power network provided by the embodiment of the application is described as follows, which comprises the following steps:

[0130] The provider edge device PE subscribes to the data of the distributed database on demand;

[0131] The computing power awareness module of the edge cloud collects the node computing power state information of the edge cloud;

[0132] The computing power awareness module notifies the node computing power state information to the distributed database;

[0133] The provider edge device PE notifies the link communication state information to the distributed database;

[0134] The distributed database publishes the received node computing power state information and link communication state information to the head node router;

[0135] The user initiates access to the application and requests computing power;

[0136] The provider edge device PE performs computing power routing according to a computing power routing table maintained by the provider edge device PE, determines a target computing power node and a packet forwarding path, and transmits the packet out.

[0137] Figure 7 is a flowchart of a third embodiment of the computing power aware method of the computing power network provided by the embodiment of the application, in combination with Figure 7 The method comprises the following steps:

[0138] Step S701: receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE.

[0139] The specific content of this part can refer to the description in step S301, which will not be repeated here.

[0140] Step S702: publishing the node computing power state information and the link communication state information to the user edge device CE, so that the user edge device CE determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user device.

[0141] The embodiment of the application does not limit the specific algorithm for determining the target computing power node and the packet forwarding path based on the node computing power state information and the link communication state information. In an embodiment of the application, the user edge device CE is built-in with a pre-determined routing algorithm, and the embodiment of the application does not limit the specific content of the routing algorithm.

[0142] It should be understood that in the computing power scheduling method of the computing power network provided by steps S301 and S302 of the foregoing second embodiment, the database publishes the node computing power state information and the link communication state information to the provider edge device PE, so that the provider edge device PE acts as a role of computing power routing for users. In step S702, the database publishes the node computing power state information and the link communication state information to the user edge device CE, so that the user edge device CE acts as a role of computing power routing for users.

[0143] In actual applications, different applications may correspond to different computing power requirements, and only the application can better understand its own computing power requirements, so the user edge device CE can be pre-built-in with a routing algorithm corresponding to such computing power requirements. Therefore, the user edge device CE performs computing power routing based on the pre-determined built-in routing algorithm, which can improve the scheduling effect of the computing power scheduling method.

[0144] It can be seen that when the distributed database publishes the collected node computing power state information and link communication state information to the user edge device CE, the perception of the computing power network is extended to the user edge device CE side, the user edge device CE can directly make strategy decision according to the node computing power state information and the link communication state information, the performance requirement of the provider edge device PE is reduced, and the scheduling effect of the computing power scheduling method can be improved.

[0145] Similar to the foregoing, based on the computing power scheduling method of the computing power network provided in steps S701 and S702, in an embodiment of the present application, the distributed database pre-establishes a communication mode based on publication and subscription with at least one user edge device CE accessing the computing power network.

[0146] In an embodiment of the present application, after receiving the node computing power state information announced by each computing power node in the computing power network and the link communication state information announced by each provider edge device PE, before publishing the node computing power state information and the link communication state to the user edge device CE, that is, before step S701 and after step S702, the method further comprises:

[0147] Integrating the node computing power state information to obtain network computing power state information of the computing power network; and integrating the link communication state information to obtain network communication state information of the computing power network.

[0148] In this communication mode based on subscription and publication, the step S702 of publishing the node computing power state information and the link communication state information to the user edge device CE specifically comprises:

[0149] When it is detected that the network computing power state information and / or the network communication state information is changed, the changed network computing power state information and / or the network communication state information is published to the corresponding information channel, so that the user edge device CE pulls the pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0150] The specific content and beneficial effects of this part can be referred to the foregoing description, which will not be repeated here.

[0151] Figure 8 is another example diagram of the computing power perception method of the computing power network provided by the embodiments of the present application, in order to facilitate understanding of the third embodiment in the foregoing, the following will be combined with Figure 8 The computing power perception method of the computing power network provided by the embodiments of the present application is described, which comprises:

[0152] The user edge device CE subscribes data to the distributed database on demand;

[0153] The computing power perception module of the edge cloud collects node computing power state information of the edge cloud;

[0154] The computing power awareness module notifies the node computing power state information to the distributed database.

[0155] The provider edge device PE notifies the link communication state information to the distributed database.

[0156] The distributed database publishes the received node computing power state information and link communication state information to the user edge device CE.

[0157] The user edge device CE is built-in with a routing algorithm, and performs routing calculation according to the received node computing power state information and link communication state information, determines the target computing power node and the message forwarding path, and forwards the message to the target computing power node through the message forwarding path.

[0158] In an embodiment of the present application, in order to improve reliability, the distributed database applying the computing power scheduling method of the computing power network provided by the embodiment of the present application can have the following mechanisms:

[0159] The distributed database includes a non-primary storage node and a primary storage node determined through an election mechanism; wherein the primary storage node is used to receive the node computing power state information announced by each computing power node in the computing power network and the link communication state information announced by the provider edge device PE, and synchronizes the received data to the non-primary storage node.

[0160] When a node in the distributed database fails, the primary storage node is determined again through the election mechanism, and the determined primary storage node contains full data; when the failed node re-joins the distributed database, the primary storage node synchronizes the full data stored in itself to the re-joined node.

[0161] Therefore, when most nodes in the network can work normally, the entire cluster can still provide normal data access services to the outside, and can tolerate less than half of the node failures.

[0162] As can be seen, through the above mechanism, the reliability of the distributed database can be improved, thereby improving the data reliability of the computing power scheduling.

[0163] As an example, the distributed database applying the computing power scheduling method of the computing power network provided by the embodiment of the present application can be etcd.

[0164] Specifically, etcd is a distributed database in the form of key-value pairs, and is built-in with a Raft consistency algorithm. The Raft is a method for maintaining data consistency between nodes, so that even if part of the nodes in the system fail, the system can still provide services to the outside.

[0165] Figure 9 is an example diagram of etcd provided by the embodiment of the present application, and the following will be described in combination withFigure 9 The method for maintaining data consistency of etcd and election of database nodes is described.

[0166] Figure 9 Three database nodes are shown in the middle, and one leader is selected by the election mechanism to provide database services externally. The election mechanism of the leader is that each node votes for the candidate, and the node with more than half of the votes obtains the leader qualification.

[0167] In combination Figure 9 When data needs to be stored, the leader node stores the data locally and sends it to other nodes. As long as more than half of the nodes return success, it is considered that the storage is successful. When reading data, the leader node reads and returns the data. Therefore, each node in the cluster has full data. When a node in the cluster fails, the cluster starts a new round of leader election and re-elects a leader. The new leader can provide database services again because it has full data. The node that fails after restarts and rejoins the cluster, and the leader synchronizes the latest database to the new node.

[0168] As can be seen, when most nodes in the network can work normally, the entire cluster can still provide data access services externally and can tolerate less than half of the node failures.

[0169] Through the above mechanism, the reliability of the distributed database is improved, thereby improving the data reliability of the computing power scheduling.

[0170] etcd is a distributed database with strong consistency and high reliability. Therefore, when the computing power scheduling method of the computing power network provided in the embodiments of the present application is applied to etcd, the acquisition and publication of information in the computing power network have high reliability, and thus the computing power scheduling method also has high reliability and high stability.

[0171] In an embodiment of the present application, each database node of the distributed database can be deployed on a different server. For example, on the basis of the existing network equipment of the computing power network, a plurality of external servers are additionally selected for deploying each database node.

[0172] Alternatively, each database node of the distributed database can be deployed on a different provider edge device PE. Specifically, since the provider edge device PE is distributed by itself, it can meet the deployment requirements of the distributed database nodes, that is, the underlying provider edge device PE can be directly used to deploy the database nodes, so that it is not necessary to additionally build external servers on the basis of the existing network equipment of the computing power network, thereby reducing resource consumption to a certain extent.

[0173] Exemplarily, Figure 10is an example diagram of database node deployment provided by an embodiment of the present application, Figure 10 Three database nodes are shown, which are respectively deployed in provider edge devices PE1, PE2 and PE3.

[0174] Referring to Figure 11 , Figure 11 is a structural schematic diagram of a computing power scheduling device of a computing power network provided by an embodiment of the present application, as Figure 11 shown, comprising the following modules:

[0175] The receiving module 1101 is configured to receive node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE;

[0176] The publishing module 1102 is configured to publish the node computing power state information and the link communication state information to the edge device, so that the edge device determines a target computing power node and a packet forwarding path based on the node computing power state information and the link communication state information in response to a computing power request of a user device.

[0177] In an embodiment of the present application, the edge device includes a provider edge device PE and / or a user edge device CE.

[0178] In an embodiment of the present application, the distributed database pre-establishes a communication mode based on publishing and subscribing with at least one provider edge device PE in the computing power network;

[0179] The device further comprises:

[0180] The integration module is configured to integrate the node computing power state information to obtain network computing power state information of the computing power network, and integrate the link communication state information to obtain network communication state information of the computing power network;

[0181] The publishing module is specifically configured to:

[0182] When it is detected that the network computing power state information and / or the network communication state information changes, the changed network computing power state information and / or network communication state information is published to a corresponding information channel, so that the provider edge device PE pulls the pre-subscribed node computing power state information and / or link communication state information from the information channel.

[0183] In an embodiment of the present application, each database node of the distributed database is respectively deployed in a different server;

[0184] Or,

[0185] Each database node of the distributed database is respectively deployed in a different provider edge device.

[0186] In one embodiment of the present invention, the storage nodes in the distributed database store the node computing power status information and / or the link communication status information in the form of data tables;

[0187] The data table includes key-value pairs; where the key is a field that represents the device category, device identifier and / or device computing power status attribute, and inter-device communication status attribute in the computing power network; and the value is the value corresponding to the key.

[0188] In one embodiment of the present invention, the distributed database includes non-master storage nodes and master storage nodes determined through an election mechanism;

[0189] The primary storage node is used to receive node computing power status information announced by each computing power node in the computing power network and link communication status information announced by each provider edge device (PE), and to synchronize the received data to the non-primary storage node.

[0190] The device further includes:

[0191] The selection module is used to re-elect the master storage node through an election mechanism when a node in the distributed database fails. The determined master storage node contains all the data.

[0192] The synchronization module is used so that when a failed node rejoins the distributed database, the master storage node synchronizes all the data it stores to the rejoined node.

[0193] The computing power scheduling device for computing networks provided in this invention introduces a distributed database into the existing computing power network architecture. The distributed database receives node computing power status information and link communication status information from the computing power network and publishes this information to edge devices. The edge devices then determine the target computing power node and packet forwarding path based on this information, thus achieving computing power scheduling. Consequently, the provider's edge devices (PEs) do not need to transmit node computing power status information and link communication status information, and therefore do not need to modify the underlying routing protocol between PEs, thus avoiding any impact on the existing network and improving the applicability of the computing power scheduling method and the stability of the computing power network.

[0194] Based on the same inventive concept, and according to the above-described embodiments of the computing power scheduling method for computing power networks, this invention also provides an electronic device, such as... Figure 12 As shown, it includes a processor 1201, a communication interface 1202, a memory 1203 and a communication bus 1204, wherein the processor 1201, the communication interface 1202 and the memory 1203 communicate with each other through the communication bus 1204.

[0195] a memory 1203 for storing a computer program;

[0196] a processor 1201 for executing the program stored in the memory, and implementing the following steps.

[0197] receiving node computing power state information announced by each computing power node in the computing power network and link communication state information announced by each provider edge device PE;

[0198] publishing the node computing power state information and the link communication state information to the edge device, so that the edge device determines target computing power nodes and packet forwarding paths based on the node computing power state information and the link communication state information in response to a computing power request of a user device.

[0199] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. 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 only one type of bus.

[0200] The communication interface is used for communication between the above electronic device and other devices.

[0201] The memory can include a Random Access Memory (RAM) and can also include a Non-Volatile Memory (NVM), such as at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0202] The processor mentioned above can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0203] In a further embodiment provided by the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. The computer program is executed by a processor to implement the steps of the computing power scheduling method of any of the computing power networks.

[0204] In a further embodiment provided by the present application, a computer program product including instructions, which when executed on a computer, cause the computer to perform the computing power scheduling method of any of the computing power networks in the above embodiments.

[0205] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) manner. The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as DVD), or semiconductor media (such as solid state disk (SSD)) and the like.

[0206] It should be noted that, in this document, the relationship 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 any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the processes, methods, articles or devices including a series of elements not only include those elements, but also include other elements not explicitly listed or inherent to such processes, methods, articles or devices. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0207] Various embodiments in the specification are described in a related manner, and the same or similar parts between various embodiments can be referred to each other, and each embodiment focuses on the difference from other embodiments. In particular, for the computing power scheduling device, electronic device, and computer storage medium embodiments of the computing power network, since they are basically similar to the computing power scheduling method embodiment of the computing power network, the description is relatively simple, and the relevant parts can refer to the part of the computing power scheduling method embodiment of the computing power network.

[0208] The above only describes the preferred embodiments of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for scheduling computing power in a computing power network, characterized in that, Applied to a distributed database, wherein the distributed database pre-establishes a publish-subscribe communication pattern with at least one edge device in the computing power network; the method includes: Receive node computing power status information announced by each computing power node in the computing power network and link communication status information announced by each provider edge device (PE); the link communication status information represents the communication status between each provider edge device (PE); The node computing power status information is integrated to obtain the network computing power status information of the computing power network; the link communication status information is integrated to obtain the network communication status information of the computing power network. When a change is detected in the network computing power status information and / or the network communication status information, the changed network computing power status information and / or network communication status information is published to the corresponding information channel, so that the edge device can pull the pre-subscribed node computing power status information and / or link communication status information from the information channel. In response to the computing power request of the user equipment, the edge device determines the target computing power node and the packet forwarding path based on the node computing power status information and the link communication status information. The edge device includes provider edge device (PE) and / or user edge device (CE). The packet forwarding path is the path information of one or more provider edge devices (PE) in the forwarding path when the packet is forwarded by the head node router to the provider edge device (PE) connected to the computing power node. The head node router is the provider edge device (PE) connected to the user edge device (CE) used by the user equipment to access the computing power network.

2. The method according to claim 1, characterized in that, The database nodes of the distributed database are deployed on different servers. or, Each database node of the distributed database is deployed on a different edge device (PE) of the provider.

3. The method according to claim 1, characterized in that, The storage nodes in the distributed database store the node computing power status information and / or the link communication status information in the form of data tables; The data table includes key-value pairs; where the key is a field that represents the device category, device identifier and / or device computing power status attribute, and inter-device communication status attribute in the computing power network; and the value is the value corresponding to the key.

4. The method according to any one of claims 1-3, characterized in that, The distributed database includes non-master storage nodes and master storage nodes determined through an election mechanism; The primary storage node is used to receive node computing power status information announced by each computing power node in the computing power network and link communication status information announced by each provider edge device (PE), and to synchronize the received data to the non-primary storage node. When a node in the distributed database fails, a new primary storage node is determined through an election mechanism, and the determined primary storage node contains all the data. When a failed node rejoins the distributed database, the primary storage node synchronizes all the data it stores to the rejoined node.

5. A computing power scheduling device for a computing power network, characterized in that, Applied to a distributed database, wherein the distributed database has pre-established a publish-subscribe communication pattern with at least one edge device in the computing power network; the device includes: The receiving module is used to receive node computing power status information announced by each computing power node in the computing power network and link communication status information announced by each provider edge device (PE); the link communication status information represents the communication status between each provider edge device (PE). An integration module is used to integrate the node computing power status information to obtain the network computing power status information of the computing power network; and to integrate the link communication status information to obtain the network communication status information of the computing power network. The publishing module is used to publish the changed network computing power status information and / or network communication status information to the corresponding information channel when a change is detected in the network computing power status information and / or network communication status information. This allows the edge device to retrieve pre-subscribed node computing power status information and / or link communication status information from the information channel. In response to a user device's computing power request, the edge device determines the target computing power node and packet forwarding path based on the node computing power status information and the link communication status information. The edge device includes a provider edge device (PE) and / or a user edge device (CE). The packet forwarding path is the path information of one or more provider edge devices (PEs) in the forwarding path when a packet is forwarded by a head node router to a provider edge device (PE) connected to the computing power node. The head node router is a provider edge device (PE) connected to the user edge device (CE) used by the user device to access the computing power network.

6. The apparatus according to claim 5, characterized in that, The database nodes of the distributed database are deployed on different servers. or, Each database node of the distributed database is deployed on a different edge device (PE) of the provider.

7. The apparatus according to claim 5, characterized in that, The storage nodes in the distributed database store the node computing power status information and / or the link communication status information in the form of data tables; The data table includes key-value pairs; where the key is a field that represents the device category, device identifier and / or device computing power status attribute, and inter-device communication status attribute in the computing power network; and the value is the value corresponding to the key.

8. The apparatus according to any one of claims 5-7, characterized in that, The distributed database includes non-master storage nodes and master storage nodes determined through an election mechanism; The primary storage node is used to receive node computing power status information announced by each computing power node in the computing power network and link communication status information announced by each provider edge device (PE), and to synchronize the received data to the non-primary storage node. The device further includes: The selection module is used to re-elect the master storage node through an election mechanism when a node in the distributed database fails. The determined master storage node contains all the data. The synchronization module is used so that when a failed node rejoins the distributed database, the master storage node synchronizes all the data it stores to the rejoined node.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-4.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-4.

Images