Multi-modal computing power network system

Abstract

The invention relates to the technical field of multi-modal networks, and provides a multi-modal computing power network system which comprises a user side, a first computing power gateway, a transmission network, a second computing power gateway, a computing power server side and a control side. The transmission network is a network formed by a multi-mode network element, the multi-mode network element and the two computing power gateways are all configured with a computing network protocol implemented by software, the computing network protocol at least comprises four fields of a computing network mode message header, the first computing power gateway encapsulates the computing network mode message header according to user information and computing power requests, and the second computing power gateway encapsulates the computing network mode message header according to the user information and the computing power requests. And the control end selects a target computing power node according to the computing network modal message header and generates a flow table, and the two computing power gateways and the transmission network transmit the computing power request to the target computing power node according to the flow table. According to the invention, the computing power network is established based on the multi-modal network, a special network does not need to be established, the cost is saved, and the problem that the network utilization rate and the effective load proportion are relatively low due to the adoption of an IPv6 extension field computing network protocol is avoided through the SDN custom computing network protocol.

Description

Technical Field

[0001] The present invention relates to the field of multimodal network technology, and in particular to a multimodal computing power network system. Background Art

[0002] Most of the existing computing networks are based on the IPV6 network protocol, and the network element equipment and network protocols need to build a dedicated network, which is very expensive. In the existing computing networks, a large part of the traffic is information about network maintenance, and the network utilization rate is low. In addition, most of the information carried by the existing computing networks is based on the extended fields of IPv6, and the result messages returned to the user contain a lot of irrelevant information, that is, the effective load ratio is low. Summary of the invention

[0003] The present invention provides a multimodal computing power network system to solve the problems in the prior art of high computing power network construction cost, low network utilization rate and low effective load ratio.

[0004] The present invention provides a multimodal computing power network system, comprising: a user end, a first computing power gateway, a transmission network, a second computing power gateway, a computing power server end and a control end.

[0005] The transmission network is a network composed of multimodal network elements, the first computing power gateway is connected to the user end and at least one multimodal network element, the second computing power gateway is connected to the computing power service end and at least one multimodal network element, and the control end is connected to the multimodal network element, the first computing power gateway and the second computing power gateway.

[0006] The multimodal network element, the first computing power gateway and the second computing power gateway are all configured with a computing network protocol implemented by software, and the computing network protocol at least includes: a computing network modality message header.

[0007] The first computing power gateway is used to encapsulate the computing network modality message header according to the user information and computing power request of the current requesting user.

[0008] The control end is used to select a target computing power node from the computing power service end according to the computing network modal message header, and generate a flow table, and send the flow table to the multimodal network element, the first computing power gateway and the second computing power gateway. The flow table includes a routing path from the user end to the target computing power node.

[0009] The multimodal network element, the first computing power gateway, and the second computing power gateway are used to transmit the computing power request to the target computing power node according to the flow table.

[0010] According to a multimodal computing power network system provided by the present invention, the computing network modal message header includes the following fields: next extension header or protocol header field, intent option enable identification field, optional parameter enable identification field, ID field, intent identifier field and parameter option field, and the ID field includes: user ID and service ID of the computing power service.

[0011] The intent identifier field indicates an intent request for different computing network service quality levels.

[0012] The parameter option fields include: network bandwidth requirements, network delay requirements, network jitter requirements, network packet loss rate requirements, preference index of the service specified by the service ID in the computing power node, percentage of available service capacity of the computing power node and service operation delay prediction.

[0013] According to a multimodal computing power network system provided by the present invention, the computing power information includes a computing power resource identifier, and the user information includes a user IP and a user level.

[0014] The control end is also used to map the user IP to the user ID defined in the computing network protocol, map the computing resource identifier to the service ID defined in the computing network protocol, and establish a first mapping relationship table of user ID, user level and computing network service quality level. For each computing network service quality level, a second mapping relationship table is established between the service ID of the computing service provided by the computing power server and the field filling information of the computing network modal message header, and the first mapping relationship table and the second mapping relationship table are sent to the first computing power gateway and the second computing power gateway.

[0015] According to a multimodal computing power network system provided by the present invention, the first computing power gateway is used to determine the user source IP of the current requesting user and the target ID of the computing power service to be requested according to the computing power request when receiving the computing power request, and encapsulate the computing power request into a computing network modal message header according to the user source IP, the target ID, the first mapping relationship table and the second mapping relationship in the format of the computing network protocol to obtain a computing power service request message, and send the computing power service request message to the control end.

[0016] The control end is used to determine the target computing power node according to the computing network modality message header in the received computing power service request message, generate the flow table, and send the flow table to the multimodal network element, the first computing power gateway and the second computing power gateway.

[0017] The second computing power gateway is used to parse the computing power request from the computing power service request message and send the computing power request to the target computing power node.

[0018] According to a multimodal computing power network system provided by the present invention, a network status monitoring protocol is also configured in the multimodal network element, and the multimodal network element is also used to monitor the network status of the transmission network in real time according to the network status monitoring protocol, and upload the network status to the control end.

[0019] The second computing power gateway is also used to obtain the computing power status of each computing power node in the computing power server in real time, and upload the computing power status to the control end, where the computing power status includes the percentage of available service capacity.

[0020] The control end is also used to verify whether the routing path meets the computing network service quality level corresponding to the user source IP according to the network status and computing power status. If it does not meet the requirements, the target computing power node is reselected, and the corresponding routing path is planned, and the flow table corresponding to the re-planned routing path is sent to the multimodal network element, the first computing power gateway and the second computing power gateway.

[0021] According to a multimodal computing power network system provided by the present invention, the multimodal network element is also used to parse the computing network modality message header when forwarding the computing power service request message, and when the network status does not meet the computing network service quality level encapsulated in the computing network modality message header, send abnormal information to the control end, so that the control end re-plans the routing path.

[0022] According to a multimodal computing power network system provided by the present invention, the computing power resource identifier is composed of at least a computing power location identifier, a computing power identity identifier and a computing power attribute identifier.

[0023] According to a multimodal computing power network system provided by the present invention, the control end is also used to cluster and group computing power nodes, and computing power nodes of the same type form a computing power center.

[0024] According to a multimodal computing power network system provided by the present invention, the first computing power gateway is also used to obtain user information, upload the user information to the control end, and store it in the database of the control end.

[0025] According to a multimodal computing power network system provided by the present invention, the second computing power gateway is also used to detect computing power information in real time, upload the computing power information to the control end, and store it in the database of the control end.

[0026] The multimodal computing power network system provided by the present invention includes: a user end, a first computing power gateway, a transmission network, a second computing power gateway, a computing power service end and a control end; the transmission network is a network composed of multimodal network elements, the first computing power gateway connects the user end and at least one multimodal network element, the second computing power gateway connects the computing power service end and at least one multimodal network element, and the control end connects the multimodal network element, the first computing power gateway and the second computing power gateway; the multimodal network element, the first computing power gateway and the second computing power gateway are all configured with a computing network protocol implemented by software, and the computing network protocol At least includes: a computing network modal message header; the first computing power gateway is used to encapsulate the computing network modal message header according to the user information of the current requesting user and the computing power request; the control end is used to select the target computing power node from the computing power service end according to the computing network modal message header, and generate a flow table, and send the flow table to the multimodal network element, the first computing power gateway and the second computing power gateway, the flow table includes the routing path from the user end to the target computing power node; the multimodal network element, the first computing power gateway and the second computing power gateway are used to transmit the computing power request to the target computing power node according to the flow table. In the present invention, the computing power network is built based on a multimodal network. The computing network protocol can be customized in the multimodal network element through the software defined network (SDN) mechanism, thereby realizing a multimodal computing power network system. The multimodal computing power network system does not need to build a dedicated network, saving the network construction cost. Moreover, the computing network protocol is customized through the SDN mechanism, avoiding the problems of low network utilization and effective load ratio in the computing power network that uses the computing network protocol with an extended field of IPv6. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 It is a schematic diagram of the network structure of the multimodal computing power network system provided by the present invention.

[0029] Figure 2 It is a schematic diagram of the computing network protocol field of the multimodal computing power network system provided by the present invention.

[0030] Figure 3 It is a schematic diagram of the field structure of the computing network modality message header in the computing network protocol field of the multi-modal computing power network system provided by the present invention.

[0031] Figure 4It is a specific structural diagram of the parameters (Parameters) field in the computing network modality message header in the multi-modal computing power network system provided by the present invention.

[0032] Figure 5 It is a schematic diagram of the information structure of computing power information in the multimodal computing power network system provided by the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] The multi-modal computing power network system of the embodiment of the present invention is as follows: Figure 1 As shown, it includes: a user end 100, a first computing power gateway 200, a transmission network 300, a second computing power gateway 400, a computing power service end 500 and a control end 600.

[0035] The transmission network 300 is a network composed of multimodal network elements 301. The first computing power gateway 200 is connected to the user end 100 and at least one multimodal network element 301. The second computing power gateway 400 is connected to the computing power service end 500 and at least one multimodal network element 301. The control end 500 is connected to the multimodal network element 301 (connected to each multimodal network element 301), the first computing power gateway 200 and the second computing power gateway 400.

[0036] Specifically, the user end 100 includes two types of users: home broadband users and campus private network (universities and enterprises). The first computing power gateway 200 and the second computing power gateway 400 can be deployed on a separate server, and the computing network protocol can be implemented by software, or they can overlap with the edge multimodal network element 301 in the transmission network 300, that is, the edge multimodal network element 301 located on the side where the user end 100 is located is used as the first computing power gateway 200, and the edge multimodal network element 301 located on the side where the computing power service end 400 is located is used as the second computing power gateway 400. The transmission network 300, that is, the data plane, is composed of multiple multimodal network elements 301 to form a transmission path. The computing power service end 500 can be a Kubernetes (K8s) cluster deployed on the server, providing computing power services in the form of containers. The control end 600, that is, the control plane, includes an orchestrator, a network management system, and a database. The orchestrator has a control function, and the network management system is the overall resource management of the modality and equipment in the engineering implementation plan. The southbound interface of the orchestrator is responsible for routing orchestration, path planning and flow table delivery, and is connected to the network management system through the northbound interface. The database is used to store user information of the user terminal 100 and computing power information of the computing power server 500. The computing power server 500 includes multiple computing power centers, each of which is composed of multiple computing power nodes, and each computing power node corresponds to its own computing power information.

[0037] The multimodal network element 301, the first computing power gateway 200 and the second computing power gateway 400 are all configured with a computing network protocol implemented by software, and the computing network protocol at least includes: a computing network modality message header field. Specifically, Figure 2 As shown, the computing network protocol includes four fields: destination physical address (DMAC), source physical address (SMAC), computing network mode type (Type) and computing network mode message header.

[0038] The source physical address SMAC is the MAC address of the user terminal 100; the destination physical address DMAC is the MAC address of the first computing power gateway 200 corresponding to the user terminal 100; the computing network modality type Type is the type of the custom computing network modality message header. In this embodiment, the computing network modality message header type is TLV (Type-Length-Value), which is a custom field information that conforms to the computing power network. Since the computing network protocol is a network protocol in a custom format, and the protocol format is different from the format defined by the extended field of IPv6, the multi-modal computing power network system of this embodiment avoids the problem of low network utilization and load effective ratio in the computing power network using the computing network protocol with the extended field of IPv6.

[0039] The first computing power gateway 200 is used to encapsulate the computing network modality message header according to the user information and computing power request of the current requesting user.

[0040] The control end 600 is used to select a target computing power node from the computing power service end 500 according to the computing network modal message header, and generate a flow table, and send the flow table to the multimodal network element 301, the first computing power gateway 200 and the second computing power gateway 400. The flow table includes a routing path from the user end 100 to the target computing power node.

[0041] The multimodal network element 301, the first computing power gateway 200 and the second computing power gateway 400 are used to transmit the computing power request to the target computing power node according to the flow table.

[0042] In the multimodal computing power network system of the present embodiment, since the computing power network is built based on the multimodal network, the computing network protocol can be customized in the multimodal network element through the Software Defined Network (SDN) mechanism, thereby realizing a multimodal computing power network system. The multimodal computing power network system does not need to build a dedicated network, saving the network construction cost, and customizes the computing network protocol through the SDN mechanism, avoiding the problem of low network utilization and effective load ratio in the computing power network that uses the computing network protocol with an extended field of IPv6.

[0043] like Figure 3 As shown, in some embodiments, the computing network modality message header includes the following fields: the next extension header or protocol header field (Next Header), the intent option enable identification field (flags), the optional parameter enable identification field (Para-Type), the ID field, the intent identifier field (Intent, which is an optional item) and the parameter option field (Parameters, which is an optional item), and the ID field includes: the user ID and the service ID of the computing power service. Among them, a computing power node can provide multiple types of computing power services, and a type of computing power service can also be provided by different computing power nodes.

[0044] Among them, the intent identifier field represents the intent request for different computing network service quality levels, and the parameter option field includes: network bandwidth requirements, network delay requirements, network jitter requirements, network packet loss rate requirements, the preference index of the service specified by the service ID in the computing power node, the available service capacity percentage of the computing power node and the service operation delay prediction.

[0045] Specifically, each field in the computing network modality message header is explained as follows.

[0046] - Next header: Indicates the type of the next extension header or protocol header.

[0047] - Flags: Defines whether to enable the intent option (Intent).

[0048] - Para-Type: 16 bits, specifies optional parameters in the order of the bitmap. Setting it to 1 means enabling the parameter.

[0049] bit0: User computing power network requests have bandwidth requirements.

[0050] bit1: There is a delay requirement for user computing power network requests.

[0051] bit2: There is a jitter requirement for user computing power network requests.

[0052] bit3: There is a packet loss rate requirement for user network calculation requests.

[0053] bit4: Services are divided into many types, such as high-speed computing, high bandwidth or large storage. This parameter indicates the preference index of the service specified by the service ID (Service ID) in the computing power node (Site-Id).

[0054] bit5: There are many running processes and services in the computing power node (Site-ID), so there needs to be an indicator to indicate the percentage of available service capacity that is free or can be provided by the site. The percentage of available service capacity means: the ratio of the available service capacity of the computing power node to the total service capacity. The service capacity refers to the amount of computing power resources that the computing power node can provide within a specific time. When it is 0, it means that this is an empty node and no service is running. When it reaches 100%, it means that the site is overloaded. If it exceeds 100%, it is considered abnormal.

[0055] bit6: For users, the delay between a request and a service response should include network delay and internal service operation delay. These two items together constitute the user experience. This item specifies whether to enable site service prediction delay, which is a predicted value for the service running time.

[0056] It should be noted that the predicted value of the service operation time is predicted by the orchestrator of the control end 600. After the network is built or a new computing node is added to the network, the orchestrator predicts the predicted value of the service operation time of the computing service provided by each computing node. During the operation of the entire computing network system, when the percentage of available service capacity of the computing node of the computing service end 500 changes, the second computing gateway 400 can determine whether to report the abnormality to the orchestrator based on whether the percentage of available service capacity after the change meets the service operation delay prediction.

[0057] - ID: 32-bit identifier, including service ID (Sevice-ID) and user ID (User-ID), which indicates the unique identifier of the application session.

[0058] - Intent: Intent identifier, indicating the intent request for different network service quality levels. When the above Flags is set to 1, it means that the requirement for the network service quality level corresponding to the intent identifier is enabled.

[0059] - Parameters: Specific parameter information of the computing power network mode. The existence of the parameter is indicated by Para-Type. If the corresponding bit of Para-Type is 1, it means that the parameter corresponding to the bit is enabled. Figure 4 As shown, each parameter occupies four bytes.

[0060] For the four parameters of network bandwidth requirement (bandwidth), network delay requirement (Delay), network jitter requirement (DelayVariation) and network packet loss rate requirement (Packet Loss Rate), the first byte is a reserved bit and the following three bytes are specific parameter values.

[0061] For the three parameters of the service specified by the service ID, namely, the preference index value of the computing node, the available service capacity percentage of the computing node, and the service delay prediction, the first byte is the computing node ID (site-id), and the last three bytes are the specific parameter values.

[0062] In some embodiments, the computing power information includes a computing power resource identifier, and the user information includes a user IP and a user level.

[0063] The control end 600 is also used to map the user IP to the user ID defined in the computing network protocol, map the computing resource identifier to the service ID defined in the computing network protocol, and establish a first mapping relationship table of user ID, user level and computing network service quality level. For each computing network service quality level, a second mapping relationship table is established between the service ID of the computing service provided by the computing power server 500 and the field filling information of the computing network modal message header, and the first mapping relationship table and the second mapping relationship table are sent to the first computing power gateway 200 and the second computing power gateway 400.

[0064] Specifically, the control end 600 is also used to manage the user information of the user end 100 and the computing power information of the computing power service end 500. When the multimodal computing power network system is built, or when a user subsequently joins the network, the orchestrator in the control end 600 maps the user IP to the user ID defined in the computing network protocol, so that the user IP and the user ID defined in the computing network protocol are bound. When the multimodal computing power network system is built, or when a new computing power node is subsequently deployed, the computing power resource identifier corresponding to the computing power node is mapped to the service ID defined in the computing network protocol, so that the computing power resource identifier is bound to the service ID defined in the computing network protocol. A first mapping relationship table of user ID, user level and computing network service quality level is established by defining a YANG model. For each computing network service quality level, a second mapping relationship table of the service ID of the computing power service provided by the computing power service end 500 and the field information filled in the computing network modality message header is established.

[0065] Among them, the user level of the user can be distinguished according to the preset rules. The higher the user level, the higher the computing network service quality level. For example, the user level is divided into three levels: high, medium and low. Correspondingly, the computing network service quality level is also divided into three levels: high, medium and low. The same level corresponds to each other, so that a first mapping relationship table from user IP to computing network service quality level can be constructed. For a computing network service quality level, the filling field information (especially: the values ​​of Flags, Para-Type, Intent and Parameters fields) of the computing network modal message header corresponding to different computing power services is also different. Therefore, for a computing network service quality level, a second mapping relationship table is established between the service ID of the computing power service provided by the computing power service end 500 and the filling field information of the computing network modal message header. This facilitates the subsequent first computing power gateway 200 to determine the computing network service quality level corresponding to the user source IP according to the computing power request of the current requesting user and the two mapping relationship tables, and also determines the filling field information of the computing network modal message header, and constructs the computing network modal message header.

[0066] In some embodiments, the first computing power gateway 200 is used to determine the user source IP of the current requesting user and the target ID of the computing power service to be requested according to the computing power request when receiving the computing power request, and encapsulate the computing power request in the computing network modal message header according to the user source IP, the target ID, the first mapping relationship table and the second mapping relationship in the format of the computing network protocol to obtain a computing power service request message, and send the computing power service request message to the control end 600. Specifically, the first computing power gateway 200 can determine the type of server or host corresponding to the target address IP according to the target address IP in the computing power request, thereby determining the target ID of the computing power service to be requested, and can also determine the target ID of the computing power service to be requested according to the characteristics of the fields in the computing power request. For example, if there is a characteristic field describing matrix calculation in the computing power request, it can be determined that the computing power service to be requested is the target ID of the matrix calculation service.

[0067] The control end 600 is used to determine the target computing power node according to the computing network modality message header in the received computing power service request message, generate the flow table, and send the flow table to the multimodal network element 301, the first computing power gateway 200 and the second computing power gateway 400. Specifically, the orchestrator in the control end 600 determines the target computing power node according to the information in the computing network modality message header, and the target computing power node can be determined by machine learning, expert library or prior knowledge.

[0068] The second computing power gateway 400 is used to parse the computing power request from the computing power service request message and send the computing power request to the target computing power node.

[0069] In some embodiments, the multimodal network element 301 is further configured with a network status monitoring protocol, and the multimodal network element 301 is further used to monitor the network status of the transmission network in real time according to the network status monitoring protocol, and upload the network status to the control terminal 600. Specifically, a network status monitoring protocol such as Two-way Active Measurement Protocol (TWAMP) and Packet Conservation Algorithm for Internet (IPCA) is configured in an in-band or out-of-band manner, and the network status information is sent to the orchestrator in the control terminal 600 through the southbound interface.

[0070] The second computing power gateway 400 is also used to obtain the computing power status of each computing power node in the computing power server 500 in real time, and upload the computing power status to the control terminal 600, wherein the computing power status includes the percentage of available service capacity.

[0071] The control end 600 is also used to verify whether the routing path meets the computing network service quality level corresponding to the user source IP according to the network status and computing power status, that is, to verify whether the network status and computing power status can reach the parameter value in the fill field information of the computing network modal message header corresponding to the computing network service quality level, for example: the bandwidth of the network status does not meet the conditions, etc. If it does not meet the conditions, reselect the target computing power node, plan the corresponding routing path, and send the flow table corresponding to the re-planned routing path to the multimodal network element 301, the first computing power gateway 200 and the second computing power gateway 400. Specifically, the orchestrator in the control end 600 plans the route and generates the flow table, so that the subsequent computing power request message data flow of the same computing power service of the user source IP can meet the user's requirements for the computing network service quality level.

[0072] In some embodiments, the multimodal network element 301 is also used to parse the computing network modality message header when forwarding the computing power service request message, and when the network status does not meet the computing network service quality level encapsulated in the computing network modality message header, send exception information to the control end 600, so that the control end 600 re-plans the routing path.

[0073] Specifically, during the operation of the multimodal computing power network system, the entire network status is also changing in real time. When the computing power service request message is transmitted to a multimodal network element 301, the multimodal network element 301 monitors the network status of the transmission network in real time. When it finds that the current network status no longer meets the computing network service quality level encapsulated in the computing network modality message header, it sends an exception message through the Packetout of the southbound interface, and the orchestrator in the control end 600 re-plans and configures the guarantee path so that the subsequent computing power request message data flow of the same computing power service of the user source IP can meet the user's requirements for the computing network service quality level.

[0074] In some embodiments, the computing power resource identifier is composed of at least a computing power location identifier, a computing power identity identifier, and a computing power attribute identifier. Figure 5 As shown in the figure, the computing power resources are hierarchically abstracted into a tree model with a mechanism similar to DNS resolution mapping, and then encoded, registered and managed according to the domain name organization rules. The computing power resource abstraction hierarchy tree is divided into national domain, computing power domain, computing power provider, product code, computing power form, computing power model, computing power capability level, and computing power instance name from the root node to the bottom. The computing power location identifier includes the national domain and the computing power domain, which refers to the location attribute; the computing power identity identifier includes the computing power provider and product code; and the computing power attribute identifier refers to the computing power service attribute. By sorting the computing power resources by Figure 5 The structural management can achieve effective and efficient management of trivial computing resources.

[0075] In some embodiments, the control terminal 600 is also used to cluster computing nodes. The computing nodes of the same type form a computing center. The orchestrator performs clustering and grouping management on the computing nodes, thereby realizing distributed allocation and scheduling of computing resources, making full use of computing resources, and improving the utilization rate and performance of the overall computing network. Figure 5 The last layer is the computing power attribute identification for classification.

[0076] In some embodiments, the first computing power gateway 200 is also used to obtain user information, upload the user information to the control terminal 600, and store it in the database of the control terminal 600. Specifically, when the multimodal computing power network system is built, the first computing power gateway 200 obtains all user information of the user terminal 100, uploads the user information to the control terminal 600, and stores it in the database of the control terminal 600. During the operation of the multimodal computing power network, for new users of the user terminal 100, that is, users who newly join the multimodal computing power network system, when connecting to the first computing power gateway 200, the first computing power gateway 200 obtains the new user information of the user, uploads the new user information to the control terminal 600, and stores it in the database of the control terminal 600, so as to update the user information stored in the database in real time.

[0077] In some embodiments, the second computing power gateway 400 is also used to detect computing power information in real time, upload the computing power information to the control end 600, and store it in the database of the control end 600. Specifically, when the multimodal computing power network system is built, the second computing power gateway 400 obtains all computing power information in the computing power service end 500, uploads the computing power information to the control end 600, and stores it in the database of the control end 600. During the operation of the multimodal computing power network, the change of computing power resources is dynamic. The activation or termination of services and the dynamic changes of service loads will affect the state of computing power resources. Therefore, the second computing power gateway 400 needs to detect and obtain the computing power information of computing power resources in real time, and upload the latest computing power information to the control end 600, and store it in the database of the control end 600, so as to update the computing power information stored in the database in real time. In actual use, the second computing power gateway 400 can send the computing power information to the orchestrator of the control end 600 through the southbound interface by configuring to read the information provided by the cloud service provider or sending probes to detect cloud computing resources.

[0078] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0079] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multimodal computing power network system, characterized in that: include: A user terminal, a first computing power gateway, a transmission network, a second computing power gateway, a computing power server, and a control terminal; The transmission network is a network composed of multimodal network elements, the first computing power gateway is connected to the user end and at least one multimodal network element, the second computing power gateway is connected to the computing power service end and at least one multimodal network element, and the control end is connected to the multimodal network element, the first computing power gateway and the second computing power gateway; The multimodal network element, the first computing power gateway and the second computing power gateway are all configured with a computing network protocol implemented by software, and the computing network protocol at least includes: a computing network modality message header; The first computing power gateway is used to encapsulate the computing network modality message header according to the user information and computing power request of the current requesting user; The control end is used to select a target computing power node from the computing power service end according to the computing network modality message header, and generate a flow table, and send the flow table to the multimodal network element, the first computing power gateway and the second computing power gateway, wherein the flow table includes a routing path from the user end to the target computing power node; The multimodal network element, the first computing power gateway, and the second computing power gateway are used to transmit the computing power request to the target computing power node according to the flow table.

2. The multimodal computing power network system according to claim 1, characterized in that: The computing network modality message header includes the following fields: a next extension header or protocol header field, an intent option enable identification field, an optional parameter enable identification field, an ID field, an intent identifier field, and a parameter option field, wherein the ID field includes: a user ID and a service ID of a computing power service; The intent identifier field indicates an intent request for different computing network service quality levels; The parameter option fields include: network bandwidth requirements, network delay requirements, network jitter requirements, network packet loss rate requirements, preference index of the service specified by the service ID in the computing power node, percentage of available service capacity of the computing power node and service operation delay prediction.

3. The multimodal computing power network system according to claim 2, characterized in that: The computing power information includes the computing power resource identifier, and the user information includes the user IP and user level; The control end is also used to map the user IP to the user ID defined in the computing network protocol, map the computing resource identifier to the service ID defined in the computing network protocol, and establish a first mapping relationship table of user ID, user level and computing network service quality level. For each computing network service quality level, a second mapping relationship table is established between the service ID of the computing service provided by the computing power server and the field filling information of the computing network modal message header, and the first mapping relationship table and the second mapping relationship table are sent to the first computing power gateway and the second computing power gateway.

4. The multimodal computing power network system according to claim 3, characterized in that: The first computing power gateway is used to determine the user source IP of the current requesting user and the target ID of the computing power service to be requested according to the computing power request when receiving the computing power request, and encapsulate the computing power request into a computing network modal message header according to the user source IP, the target ID, the first mapping relationship table and the second mapping relationship in the format of the computing network protocol to obtain a computing power service request message, and send the computing power service request message to the control end; The control end is used to determine the target computing power node according to the computing network modality message header in the received computing power service request message, generate the flow table, and send the flow table to the multimodal network element, the first computing power gateway, and the second computing power gateway; The second computing power gateway is used to parse the computing power request from the computing power service request message and send the computing power request to the target computing power node.

5. The multimodal computing power network system according to claim 4, characterized in that: The multimodal network element is also configured with a network status monitoring protocol, and the multimodal network element is also used to monitor the network status of the transmission network in real time according to the network status monitoring protocol, and upload the network status to the control end; The second computing power gateway is further used to obtain the computing power status of each computing power node in the computing power server in real time, and upload the computing power status to the control terminal, wherein the computing power status includes the percentage of available service capacity; The control end is also used to verify whether the routing path meets the computing network service quality level corresponding to the user source IP according to the network status and computing power status. If it does not meet the requirements, the target computing power node is reselected, and the corresponding routing path is planned, and the flow table corresponding to the re-planned routing path is sent to the multimodal network element, the first computing power gateway and the second computing power gateway.

6. The multimodal computing power network system according to claim 5, characterized in that: The multimodal network element is also used to parse the computing network modality message header when forwarding the computing power service request message, and when the network status does not meet the computing network service quality level encapsulated in the computing network modality message header, send abnormal information to the control end so that the control end can re-plan the routing path.

7. The multimodal computing power network system according to claim 3, characterized in that: The computing power resource identifier is composed of at least a computing power location identifier, a computing power identity identifier, and a computing power attribute identifier.

8. The multimodal computing power network system according to claim 3, characterized in that: The control end is also used to cluster and group computing nodes, and computing nodes of the same type form a computing center.

9. The multimodal computing power network system according to any one of claims 1 to 8, characterized in that: The first computing power gateway is also used to obtain user information, upload the user information to the control end, and store it in the database of the control end.

10. The multimodal computing power network system according to any one of claims 1 to 8, characterized in that: The second computing power gateway is also used to detect computing power information in real time, upload the computing power information to the control end, and store it in the database of the control end.

Images