Calculation network sensing method and device and related equipment
Through the information interaction between the computing network perception module and the first computing power cerebellum module, the computing volume and network communication load of the computing network perception module are reduced, the efficiency and accuracy of computing network perception are improved, and the problem of inefficiency in the prior art is solved.
Patent Information
- Application Number
- CN202410005552.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-02
- Publication Date
- 2025-07-04
AI Technical Summary
The current computing network perception module needs to obtain a large amount of computing power nodes and link quality information, resulting in low perception efficiency of computing network.
The computing network perception module sends a second computing power perception request to the first computing power cerebellum module. The first computing power cerebellum module querys and provides resource information of the computing power node. The computing network perception module determines the computing network perception information based on this information, reducing the computing volume and network communication load.
The efficiency of the computing network perception module is improved, the computing volume and network communication load are reduced, and the real-time and accuracy of computing network perception is enhanced.
Smart Images
Figure CN120263844A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present invention relate to the technical field of computing power processing, and in particular, to a computing network perception method, device and related equipment. Background Art
[0002] Computing power perception is the comprehensive perception of the deployment location, real-time status, load information, and business requirements of computing power resources and computing power services by a computing power network, and is the basis and prerequisite for the reasonable allocation and scheduling of computing tasks in the computing power network. Currently, the computing network perception module of a computing power network usually needs to obtain the link quality information of each computing power node and between computing power nodes to complete computing network perception. However, due to the large number of computing power nodes, the time taken for the computing network perception module to complete computing network perception is too long. It can be seen that the efficiency of the current computing network perception module in completing computing network perception is low. Summary of the Invention
[0003] The embodiments of the present invention provide a computing network perception method, device and related equipment to solve the problem that the current computing network perception module has low efficiency in completing computing network perception.
[0004] To solve the above problems, the present invention is implemented as follows:
[0005] In a first aspect, the embodiments of the present invention provide a computing network perception method, which is applied to a computing network perception module. The computing network perception module is communicatively connected to a task scheduling module and a first computing power cerebellum module respectively. The method includes:
[0006] Receiving a first computing power perception request sent by the task scheduling module;
[0007] Sending a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request. The first computing power cerebellum module is used to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0008] Receiving the first computing power resource information sent by the first computing power cerebellum module;
[0009] Determining computing network perception information based on the first computing power resource information.
[0010] In a second aspect, the embodiments of the present invention provide a computing network perception method, which is applied to a first computing power cerebellum module. The first computing power cerebellum module is communicatively connected to a computing network perception module. The method includes:
[0011] Receiving a second computing power perception request sent by the computing network perception module;
[0012] Obtaining the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0013] Send the first computing power resource information to the computing network perception module, which is used to determine computing network perception information based on the first computing power resource information.
[0014] In a third aspect, an embodiment of the present invention provides a computing network perception device, which is applied to a computing network perception module. The computing network perception module is respectively communicatively connected to a task scheduling module and a first computing power cerebellum module. The computing network perception device includes:
[0015] A first receiving module, configured to receive a first computing power perception request sent by the task scheduling module;
[0016] A first sending module, configured to send a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request. The first computing power cerebellum module is used to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0017] A second receiving module, configured to receive the first computing power resource information sent by the first computing power cerebellum module;
[0018] A determining module, configured to determine computing network perception information based on the first computing power resource information.
[0019] In a fourth aspect, an embodiment of the present invention provides a computing network perception device, which is applied to a first computing power cerebellum module. The first computing power cerebellum module is communicatively connected to the computing network perception module. The computing network perception device includes:
[0020] A third receiving module, configured to receive a second computing power perception request sent by the computing network perception module;
[0021] An obtaining module, configured to obtain the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0022] A second sending module, configured to send the first computing power resource information to the computing network perception module, and the computing network perception module is used to determine computing network perception information based on the first computing power resource information.
[0023] In a fifth aspect, an embodiment of the present invention further provides an electronic device, including: a transceiver, a memory, a processor, and a program stored on the memory and executable on the processor; the processor is configured to read the program in the memory to implement the steps in the method described in the foregoing first aspect or second aspect.
[0024] Sixthly, an embodiment of the present invention further provides a readable storage medium for storing a program, which when executed by a processor implements the steps in the method described in the foregoing first aspect or second aspect.
[0025] In an embodiment of the present invention, a first computing power awareness request sent by the task scheduling module is received; a second computing power awareness request is sent to the first computing power cerebellum module based on the first computing power awareness request, and the first computing power cerebellum module is configured to query first computing power resource information of computing power nodes controlled by the first computing power cerebellum module based on the second computing power awareness request; the first computing power resource information sent by the first computing power cerebellum module is received; and computing power-network awareness information is determined based on the first computing power resource information. In this way, the computing power-network awareness module only needs to send a second computing power awareness request to the first computing power cerebellum module, and the first computing power cerebellum module obtains the first computing power resource information of the computing power nodes it controls. Then the computing power-network awareness module determines the computing power-network awareness information according to the first computing power resource information, reducing the calculation amount of the computing power-network awareness module and improving the efficiency of the computing power-network awareness module to complete computing power-network awareness. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for description in the embodiments of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.
[0027] Figure 1 It is the overall architecture diagram of the computing power network applied in the embodiment of the present invention;
[0028] Figure 2 It is one of the structural schematic diagrams of the computing power network applied in the embodiment of the present invention;
[0029] Figure 3 It is one of the flow schematic diagrams of the computing power-network awareness method provided by the embodiment of the present invention;
[0030] Figure 4 It is the schematic diagram of the computing power resource utilization rate of the first computing power cerebellum module provided by the embodiment of the present invention;
[0031] Figure 5 It is the schematic diagram of the communication time consumption between the computing power nodes controlled by the first computing power cerebellum module and the corresponding first computing power cerebellum module provided by the embodiment of the present invention;
[0032] Figure 6 It is the schematic diagram of the number of computing power nodes controlled by the first computing power cerebellum module provided by the embodiment of the present invention;
[0033] Figure 7It is the second flow schematic diagram of the computing and network perception method provided by the embodiment of the present invention;
[0034] Figure 8 It is the third flow schematic diagram of the computing and network perception method provided by the embodiment of the present invention;
[0035] Figure 9 It is the second structure schematic diagram of the computing power network applied in the embodiment of the present invention;
[0036] Figure 10 It is the fourth flow schematic diagram of the computing and network perception method provided by the embodiment of the present invention;
[0037] Figure 11 It is the first structure schematic diagram of the computing and network perception device provided by the embodiment of the present invention;
[0038] Figure 12 It is the second structure schematic diagram of the computing and network perception device provided by the embodiment of the present invention;
[0039] Figure 13 It is the structure schematic diagram of the electronic device provided by the embodiment of the present invention. Specific embodiments
[0040] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.
[0041] Terms such as "first" and "second" in the embodiments of the present invention are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices. In addition, the use of "and / or" in this application means at least one of the connected objects. For example, A and / or B and / or C means including seven cases: A alone, B alone, C alone, A and B both present, B and C both present, A and C both present, and A, B, and C all present.
[0042] See Figure 1 , Figure 1 is the overall architecture diagram of the computing power network applied in the embodiment of the present invention, as Figure 1As shown in the figure, the computing power network includes a computing power service layer, a computing power routing layer, a computing network infrastructure layer, and a computing network orchestration and management layer. Among them, the computing power service layer and the computing power routing layer are connected through the I1 interface, the computing power routing layer and the computing network infrastructure layer are connected through the I2 interface, and the computing network orchestration and management layer and the computing power service layer are connected through the I3-1 interface, and the computing network orchestration and management layer and the computing power routing layer are connected through the I3-2 interface, and the computing network orchestration and management layer and the computing network infrastructure layer are connected through the I3-3 interface. It should be noted that the modules included in the computing power service layer, the computing power routing layer, the computing network infrastructure layer, and the computing network orchestration and management layer can be referred to Figure 1 as shown.
[0043] Refer to Figure 2 , Figure 2 which is a schematic structural diagram of the computing power network applied in the embodiment of the present invention. As Figure 2 shown, the computing power routing layer includes a computing network perception module and a task scheduling module, and the computing network infrastructure layer includes multiple computing power cerebella. The multiple computing power cerebella can include computing power cerebellum C1, computing power cerebellum C2, and computing power cerebellum CN, and the first computing power cerebellum module can be one of the multiple computing power cerebella. It should be noted that the computing network perception module can also be referred to as the computing power perception module.
[0044] Refer to Figure 3 , Figure 3 which is one of the flow schematic diagrams of the computing network perception method provided by the embodiment of the present invention and is applied to the computing network perception module. The computing network perception module is respectively communicatively connected to the task scheduling module and the first computing power cerebellum module. As Figure 3 shown, the computing network perception method can include the following steps:
[0045] Step 301, receive the first computing power perception request sent by the task scheduling module.
[0046] Among them, the computing network perception module can be referred to as the computing network brain or the computing network brain module.
[0047] Among them, the first computing power perception request can be a request sent by an electronic device (the electronic device can also be referred to as a terminal or a user device) to the computing power service layer and forwarded to the task scheduling module by the computing power service layer, and the first computing power perception request can be a request for determining the computing network perception information of the computing power network.
[0048] In addition, when the task scheduling module receives the first computing power perception request forwarded by the computing power service layer, it can parse the type of computing power resources requested by the first computing power perception request and forward the first computing power perception request to the computing network perception module in the computing power routing layer.
[0049] Step 302: Send a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request. The first computing power cerebellum module is used to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request.
[0050] Among them, the computing power and network perception module can directly send the second computing power perception request to the first computing power cerebellum module, or, see Figure 2 , the computing power and network perception module can also forward the second computing power perception request to the first computing power cerebellum module through the corresponding forwarding module of the computing power routing layer.
[0051] Optionally, the first computing power perception request and the second computing power perception request can be the same computing power perception request, that is, after receiving the first computing power perception request, the computing power and network perception module can forward the first computing power perception request as the second computing power perception request to the first computing power cerebellum module.
[0052] Alternatively, the first computing power perception request and the second computing power perception request can be different computing power perception requests. The second computing power perception request can be a computing power perception request obtained after encapsulation processing based on the first computing power perception request, that is, the second computing power perception request can include both the first computing power perception request and the target information at the same time, and the target information can be the identification information of the first computing power cerebellum module, the identification information of the task scheduling module, etc. The specific content of the target information is not limited here.
[0053] Among them, the first computing power cerebellum module can control computing power nodes, see Figure 2 , when the first computing power cerebellum module is the computing power cerebellum C1, it can control computing power nodes 1-C1, computing power nodes 2-C1, and computing power nodes 3-C1, etc.; when the first computing power cerebellum module is the computing power cerebellum C2, it can control computing power nodes 1-C2, computing power nodes 2-C2, and computing power nodes 3-C2, etc.; when the first computing power cerebellum module is the computing power cerebellum CN, it can control computing power nodes 1-CN, computing power nodes 2-CN, and computing power nodes 3-CN, etc.
[0054] Among them, the specific content of the first computing power resource information is not limited here. Optionally, the first computing power resource information can include at least one of the following: the computing power resource utilization rate of the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, and the communication time consumption between the computing power nodes controlled by the first computing power cerebellum module and the corresponding first computing power cerebellum module.
[0055] The computing power resource utilization rate of the first computing power cerebellum module can be seen in Figure 4 as shown, the communication time consumption between the computing power nodes controlled by the first computing power cerebellum module and the corresponding first computing power cerebellum module can be seen in Figure 5As shown, the number of computing power nodes controlled by the first computing power cerebellum module can be referred to Figure 6 as shown.
[0056] Step 303: Receive the first computing power resource information sent by the first computing power cerebellum module.
[0057] Among them, the computing power and network perception module can directly receive the first computing power resource information sent by the first computing power cerebellum module, or, referring to Figure 2 , it can also forward the first computing power resource information sent by the first computing power cerebellum module to the computing power and network perception module through the corresponding forwarding module of the computing power routing layer.
[0058] Step 304: Determine the computing power and network perception information based on the first computing power resource information.
[0059] Among them, the computing power and network perception information can be the computing power and network topology information, or the computing power and network perception information can also be the computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, which is not specifically limited here.
[0060] In the embodiment of the present invention, through steps 301 to 304, receive the first computing power perception request sent by the task scheduling module; send a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request, where the first computing power cerebellum module is used to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; receive the first computing power resource information sent by the first computing power cerebellum module; determine the computing power and network perception information based on the first computing power resource information. In this way, the computing power and network perception module only needs to send a second computing power perception request to the first computing power cerebellum module, and the first computing power cerebellum module obtains the first computing power resource information of the computing power nodes it controls, and then the computing power and network perception module determines the computing power and network perception information according to the first computing power resource information, reducing the calculation amount of the computing power and network perception module and improving the efficiency of the computing power and network perception module to complete the computing power and network perception.
[0061] In addition, since the computing power and network perception module only needs to complete the computing power and network perception through the first computing power cerebellum module, without performing a global computing power and network perception on the computing power network, it reduces the frequency of the computing power and network perception module to perform the computing power and network perception, and reduces the network communication load of the computing power and network perception and the control load of the computing power and network perception module.
[0062] As an optional implementation manner, the computing power and network perception module is communicatively connected to a plurality of second computing power cerebellum modules, and the first computing power cerebellum module is one of the plurality of second computing power cerebellum modules. The sending the second computing power perception request to the first computing power cerebellum module based on the first computing power perception request includes:
[0063] Obtain the perception activation score of the second computing power cerebellar module, where the perception activation score is used to represent the computing power of the second computing power cerebellar module;
[0064] Determine the target perception activation score with the largest value among multiple said perception activation scores;
[0065] Determine the second computing power cerebellar module corresponding to the target perception activation score as the first computing power cerebellar module;
[0066] Send the second computing power perception request to the first computing power cerebellar module based on the first computing power perception request.
[0067] In the embodiment of the present invention, the second computing power cerebellar module corresponding to the target perception activation score with the largest value among the perception activation scores is determined as the first computing power cerebellar module, and the computing power of the second computing power cerebellar module corresponding to the target perception activation score is the best. In this way, only the second computing power cerebellar module with the best computing power needs to be screened for computing network perception, and the computing network perception module only needs to send the second computing power perception request to the second computing power cerebellar module with the best computing power, thereby greatly reducing the number of computing power nodes for computing network perception and further improving the efficiency of the computing network perception module to complete computing network perception.
[0068] It should be noted that the determination method of the perception activation score of the second computing power cerebellar module is not limited here. As an optional implementation manner, the perception activation score corresponding to each second computing power cerebellar module can be determined according to the computing power usage of each second computing power cerebellar module; alternatively, the perception activation score of each second computing power cerebellar module can be a pre-set parameter.
[0069] As another optional implementation manner, the obtaining of the perception activation score of the second computing power cerebellar module includes:
[0070] Obtain the second computing power resource information of the second computing power cerebellar module, where the second computing power resource information includes at least one of the following: the computing power resource utilization rate of the second computing power cerebellar module, the number of computing power nodes controlled by the second computing power cerebellar module, and the communication time consumption between the second computing power cerebellar module and the controlled computing power nodes;
[0071] Calculate the perception activation score of the second computing power cerebellar module according to the second computing power resource information.
[0072] In the embodiment of the present invention, the perception activation score of the second computing power cerebellar module is calculated according to the second computing power resource information, and the second computing power resource information includes at least one of the following: the computing power resource utilization rate of the second computing power cerebellar module, the number of computing power nodes controlled by the second computing power cerebellar module, and the communication latency between the second computing power cerebellar module and the controlled computing power nodes. That is, the diversity of the content of the second computing power resource information is increased. When the content of the second computing power resource information is relatively large, the accuracy of the result of the calculated perception activation score of the second computing power cerebellar module can be improved.
[0073] Optionally, the above embodiment can also be described as follows:
[0074] When the second computing power resource information includes at least one of the following: the computing power resource utilization rate of the second computing power cerebellar module, the number of computing power nodes controlled by the second computing power cerebellar module, and the communication latency between the second computing power cerebellar module and the controlled computing power nodes, the perception activation score of the second computing power cerebellar module is calculated according to the second computing power resource information. That is, it can be understood that multiple pieces of information included in the second computing power resource information are input into Formula 1, and the perception activation score of the second computing power cerebellar module is calculated. Formula 1 can be seen as follows:
[0075]
[0076] Where:
[0077] i represents the index identifier of all computing power resource types of the current second computing power cerebellar module;
[0078] j represents the index identifier of the second computing power cerebellar module;
[0079] usage i represents the computing power resource usage of the i-th computing power resource type information of the j-th second computing power cerebellar module;
[0080] num j represents the number of computing power nodes controlled by the j-th second computing power cerebellar module;
[0081] time_consumed j represents the communication latency from the computing network perception layer (i.e., the computing network perception module) to the j-th second computing power cerebellar module;
[0082] is the resource type compliance score of the j-th second computing power cerebellar module, indicating whether the resource type held by this second computing power cerebellar module conforms to the computing power resource type of the current request (i.e., the second computing power perception request). The calculation formula is as shown in Formula 2:
[0083]
[0084] In addition, the second computing power cerebellum module corresponding to the target perception activation score can be determined as the first computing power cerebellum module according to Formula 3:
[0085] controller optimum = Max(score j ) (Formula 3);
[0086] where controller optimum represents the second computing power cerebellum module corresponding to the target perception activation score, and Max(score j ) represents the perception activation score with the largest value among multiple perception activation scores.
[0087] As an alternative implementation, the computing network perception information is computing network topology information, and the first computing power resource information includes at least one of the following: the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, and the communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
[0088] Among them, the communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module can be used to represent the smoothness of the communication link between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module. When the communication latency is greater, it means the communication link is more congested; when the communication latency is smaller, it means the communication link is smoother.
[0089] In the embodiment of the present invention, when the content of the first computing power resource information is relatively large, that is, including: the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, the communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module, etc., the accuracy of the result of the calculated computing network topology information can be improved.
[0090] It should be noted that when the first computing power resource information includes the remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module and the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the calculation method of the utilization rate of the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module can refer to Formula 4, and Formula 4 can be referred to as the following expression:
[0091]
[0092] Among them, j and n represent the index identifiers of the first computing power cerebellar module;
[0093] usage represents the utilization rate of the computing power resource type information of the computing power nodes controlled by the first computing power cerebellar module;
[0094] remain j represents the remaining amount of the computing power resources of the computing power resource type information of the j-th first computing power cerebellar module;
[0095] total j represents the total amount information of the computing power resources of the computing power nodes controlled by the j-th first computing power cerebellar module;
[0096] num represents the total number of the first computing power cerebellar modules.
[0097] It should be noted that, optionally, when the target perception activation score cannot be determined based on the multiple perception activation scores, the computing network perception module may send a second computing power perception request to all the first computing power cerebellar modules in the computing network; all the first computing power cerebellar modules receive the second computing power perception request, send the second computing power perception request to the computing power nodes they control, and calculate the first computing power resource information of the computing power nodes controlled by the first computing power cerebellar module (i.e., the utilization rate of various types of resources in the first computing power cerebellar module (see formula 4), the number of nodes, the communication time-consuming, etc.), and report the first computing power resource information to the computing network perception module; the computing network perception module stores the first computing power resource information of the computing power nodes controlled by the first computing power cerebellar module.
[0098] As an optional implementation manner, after determining the computing network perception information based on the first computing power resource information, the method further includes:
[0099] Sending the computing network topology information to the task scheduling module, and the task scheduling module is used to determine the target computing power node based on the computing network topology information and use the target computing power node to process the target task.
[0100] Among them, the specific content of the target task is not limited here. Optionally, the target task may be an Extended Reality (XR) task, and the above XR task may also be referred to as a cloud rendering task.
[0101] Among them, the target computing power node may be the node with the optimal computing power among the corresponding computing power nodes in the computing network topology information.
[0102] In an embodiment of the present invention, a target computing power node is determined through computing network topology information, and the target task is processed using the target computing power node. The target computing power node can be the node with the optimal computing power among the corresponding computing power nodes in the computing network topology information. In this way, the computing power of the target computing power node can be fully utilized, thereby improving the completion efficiency of the target task.
[0103] It should be noted that, for the convenience of describing the embodiments of the present invention, each embodiment of the present invention can refer to Figure 7 the steps included in the embodiments shown, which will not be specifically described here.
[0104] Refer to Figure 8 , Figure 8 FIG. 3 is a flowchart of a computing network perception method provided by an embodiment of the present invention, which is applied to a first computing power cerebellum module. The first computing power cerebellum module is communicatively connected to a computing network perception module. As Figure 8 shown, the method includes the following steps:
[0105] Step 801: Receive a second computing power perception request sent by the computing network perception module.
[0106] Step 802: Obtain first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request.
[0107] Step 803: Send the first computing power resource information to the computing network perception module, and the computing network perception module is used to determine computing network perception information based on the first computing power resource information.
[0108] In an embodiment of the present invention, through steps 801 to 803, the computing network perception module only needs to send a second computing power perception request to the first computing power cerebellum module. The first computing power cerebellum module obtains the first computing power resource information of the computing power nodes it controls, and then the computing network perception module determines the computing network perception information according to the first computing power resource information, reducing the calculation amount of the computing network perception module and improving the efficiency of the computing network perception module to complete computing network perception.
[0109] As an optional implementation manner, the obtaining of the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request includes:
[0110] Sending a third computing power perception request to the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0111] Receiving the first computing power resource information replied by the computing power nodes controlled by the first computing power cerebellum module for the third computing power perception request;
[0112] Wherein, the first computing power resource information includes at least one of the following:
[0113] The computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module;
[0114] The remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module;
[0115] The total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module;
[0116] The number of computing power nodes controlled by the first computing power cerebellum module;
[0117] The communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
[0118] Optionally, the second computing power perception request and the third computing power perception request may be the same perception request, that is, the first computing power cerebellum module may use the second computing power perception request as the third computing power perception request and forward the third computing power perception request to the computing power nodes controlled by the first computing power cerebellum module.
[0119] Alternatively, the second computing power perception request and the third computing power perception request may be different perception requests, that is, the first computing power cerebellum module may perform target processing on the second computing power perception request to obtain the third computing power perception request, and the third computing power perception request may include the second computing power perception request and target information, and the target information may be the identification information of the first computing power cerebellum module, etc. The specific content of the target information is not limited herein.
[0120] In the embodiment of the present invention, when the content of the first computing power resource information is relatively large, that is, including: the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, the communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module, etc., the accuracy of the result of the computing network perception information calculated based on the first computing power resource information can be improved.
[0121] It should be noted that this embodiment is the embodiment of the first computing power cerebellum module corresponding to the above method embodiment. Therefore, the relevant descriptions in the above method embodiment can be referred to, and the same beneficial effects can be achieved. To avoid repeated description, it will not be elaborated herein.
[0122] In the embodiments of the present invention, the various optional embodiments introduced can be combined with each other or implemented separately without conflict. The embodiments of the present invention do not make any limitations in this regard.
[0123] For easy understanding, the examples are described as follows:
[0124] See Figure 9 , Figure 9 , which is a schematic diagram of the computing power network structure when the embodiment of the present invention is applied to the XR cloud rendering scenario. Figure 9 The terminal side in Figure 9 can be understood as the user side, while Figure 9 the computing network perception in Figure 9 can be understood as the computing network perception module in the embodiment of the present invention.
[0125] And the specific flowchart of the XR cloud rendering scenario can be seen in Figure 10 as shown.
[0126] Such as Figure 10 shown, specifically including the following steps:
[0127] 1. The cloud rendering service layer (which can be understood as the computing power service layer) receives the cloud rendering request (i.e., the first computing power perception request) from the terminal side (which can also be referred to as the terminal or user device), and issues a rendering task to the task scheduling layer (i.e., the task scheduling module).
[0128] 2. The task scheduling layer receives the issued task and parses the request resource type of the task, and at the same time transmits the computing network perception request (i.e., the first computing power perception request) to the computing power perception layer (i.e., the computing network perception module).
[0129] 3. The computing network perception layer receives the computing power perception request from the task scheduling layer, calculates the computing network perception activation score (i.e., the perception activation score) of each computing power cerebellum (i.e., the second computing power cerebellum module) according to the computing power resource utilization rate of each computing power cerebellum, the number of computing power nodes controlled by the computing power cerebellum, and their corresponding communication time consumption (i.e., the second computing power resource information), determines the optimal computing power cerebellum X (i.e., the first computing power cerebellum module) for computing power perception, and sets the computing network perception activation state (activation / inhibition) of each computing power cerebellum.
[0130] 4. The computing power cerebellum X receives the computing power perception request (i.e., the second computing power perception request) from the computing network perception layer, issues a computing power resource query request (i.e., the third computing power perception request) to the computing power nodes it controls, queries information such as the computing power resource type, the remaining amount of each type of computing power resource, and the total amount of computing power resources of each node (i.e., the first computing power resource information), and reports it to the computing network perception layer.
[0131] 5. The computing and network perception layer receives the computing power resource information reported by the optimal computing power cerebellum X (i.e., the first computing power resource information), uses this computing power resource information to calculate and update the utilization rate, the number of computing power nodes, and the communication time consumption of each type of computing power resource of the computing power cerebellum X, and at the same time generates the computing and network topology (i.e., the computing and network perception information) corresponding to the computing power nodes controlled by the computing power cerebellum network.
[0132] 6. The task scheduling layer determines the scheduling node (i.e., the target computing power node) through the computing and network topology discovery, and performs the scheduling and distribution of tasks.
[0133] 7. The scheduling node executes the task (i.e., the target task is the rendering task), and returns the rendering result to the terminal side.
[0134] When applying the embodiment of the present invention to this cloud rendering application scenario, the overall performance of the cloud rendering application can be significantly improved, which is mainly reflected in:
[0135] 1. In the computing power network corresponding to cloud rendering, the scale of the number of nodes overall controlled by the computing and network perception module in the computing power network can be reduced by one order of magnitude, which can greatly reduce the network control overhead and communication link load for the computing and network perception module to perform computing power perception.
[0136] 2. Due to a large reduction in the number of nodes that need to be controlled by the computing and network perception module for computing power perception, the scale of the computing and network topology information generated by the computing and network is smaller, and the time consumption for node selection during task scheduling is less.
[0137] 3. During the computing power perception process, only the control of some computing power nodes (i.e., the computing power nodes controlled by the first computing power cerebellum module) needs to be implemented, which also greatly improves the real-time performance of the computing and network computing power perception.
[0138] 4. Due to the computing power cerebellum activation scoring algorithm, the computing power resources controlled by the computing power cerebellum (i.e., the first computing power cerebellum module) are also the most abundant in the current computing power network. When a cloud rendering request is scheduled, a computing power node with sufficient computing power resources can be obtained for scheduling and processing, and the processing speed is faster.
[0139] See Figure 11 , Figure 11 which is one of the structural diagrams of the computing and network perception device provided by the embodiment of the present invention. The computing and network perception device is applied to the computing and network perception module, and the computing and network perception module is respectively communicatively connected to the task scheduling module and the first computing power cerebellum module. As Figure 11 shown, the computing and network perception device 1100 includes:
[0140] The first receiving module 1101 is used to receive the first computing power perception request sent by the task scheduling module.
[0141] The first sending module 1102 is configured to send a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request. The first computing power cerebellum module is configured to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request;
[0142] The second receiving module 1103 is configured to receive the first computing power resource information sent by the first computing power cerebellum module;
[0143] The determining module 1104 is configured to determine the computing network perception information based on the first computing power resource information.
[0144] As an optional implementation manner, the computing network perception module is communicatively connected to multiple second computing power cerebellum modules. The first computing power cerebellum module is one of the multiple second computing power cerebellum modules. The first sending module 1102 includes:
[0145] An obtaining sub-module, configured to obtain the perception activation score of the second computing power cerebellum module, where the perception activation score is used to represent the computing power of the second computing power cerebellum module;
[0146] A first determining sub-module, configured to determine a target perception activation score with the largest value among the multiple perception activation scores;
[0147] A second determining sub-module, configured to determine the second computing power cerebellum module corresponding to the target perception activation score as the first computing power cerebellum module;
[0148] A first sending sub-module, configured to send the second computing power perception request to the first computing power cerebellum module based on the first computing power perception request.
[0149] As an optional implementation manner, the obtaining sub-module includes:
[0150] An obtaining unit, configured to obtain the second computing power resource information of the second computing power cerebellum module, where the second computing power resource information includes at least one of the following: the computing power resource utilization rate of the second computing power cerebellum module, the number of computing power nodes controlled by the second computing power cerebellum module, and the communication latency between the second computing power cerebellum module and the controlled computing power nodes;
[0151] A calculating unit, configured to calculate the perception activation score of the second computing power cerebellum module according to the second computing power resource information.
[0152] As an alternative implementation, the computing and network awareness information is computing and network topology information, and the first computing power resource information includes at least one of the following: the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the remaining quantity information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, and the communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
[0153] As an alternative implementation, the computing and network awareness device 1100 further includes:
[0154] A third sending module, configured to send the computing and network topology information to the task scheduling module, and the task scheduling module is configured to determine a target computing power node based on the computing and network topology information and use the target computing power node to process a target task.
[0155] The computing and network awareness device 1100 can implement each process of the method embodiment in the present invention Figure 3 and achieve the same beneficial effects. To avoid repetition, details are not described herein again.
[0156] See Figure 12 , Figure 12 which is the second structural diagram of the computing and network awareness device provided by the embodiment of the present invention. The computing and network awareness device is applied to the first computing power cerebellum module, and the first computing power cerebellum module is communicatively connected to the computing and network awareness module. As shown in Figure 12 , the computing and network awareness device 1200 includes:
[0157] A third receiving module 1201, configured to receive a second computing power awareness request sent by the computing and network awareness module;
[0158] An obtaining module 1202, configured to obtain first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power awareness request;
[0159] A second sending module 1203, configured to send the first computing power resource information to the computing and network awareness module, and the computing and network awareness module is configured to determine computing and network awareness information based on the first computing power resource information.
[0160] As an alternative implementation, the obtaining module 1202 includes:
[0161] A second sending sub-module, configured to send a third computing power awareness request to the computing power nodes controlled by the first computing power cerebellum module based on the second computing power awareness request;
[0162] A receiving sub-module, configured to receive the first computing power resource information replied by the computing power nodes controlled by the first computing power cerebellum module for the third computing power perception request;
[0163] Wherein, the first computing power resource information includes at least one of the following:
[0164] The computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module;
[0165] The remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module;
[0166] The total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module;
[0167] The number of computing power nodes controlled by the first computing power cerebellum module;
[0168] The communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
[0169] The computing power and network perception device 1200 can implement each process of the method embodiments in the present invention Figure 8 and achieve the same beneficial effects. To avoid repetition, it will not be elaborated here.
[0170] The embodiments of the present invention further provide an electronic device. Please refer to Figure 13 , the electronic device may include a processor 1301, a memory 1302, and a program 13021 stored on the memory 1302 and executable on the processor 1301.
[0171] When the electronic device includes a computing power and network perception module, when the program 13021 is executed by the processor 1301, it can implement Figure 3 any step in the corresponding method embodiment and achieve the same beneficial effects, which will not be elaborated here.
[0172] When the electronic device includes a first computing power cerebellum module, when the program 13021 is executed by the processor 1301, it can implement Figure 8 any step in the corresponding method embodiment and achieve the same beneficial effects, which will not be elaborated here.
[0173] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the program can be stored in a readable medium. The embodiments of the present invention further provide a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it can implement the above Figure 3 or Figure 8Any step in the corresponding method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.
[0174] The storage medium described above, such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.
[0175] The above is the preferred implementation manner of the embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.
Claims
1. A computing and network perception method, which is applied to a computing and network perception module. The computing and network perception module is respectively communicatively connected to a task scheduling module and a first computing power cerebellum module, and is characterized in that The method includes: Receiving a first computing power perception request sent by the task scheduling module; Sending a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request, where the first computing power cerebellum module is used to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; Receiving the first computing power resource information sent by the first computing power cerebellum module; Determining computing power-network perception information based on the first computing power resource information.
2. The method according to claim 1, wherein The computing power-network perception module is communicatively connected to multiple second computing power cerebellum modules, and the first computing power cerebellum module is one of the multiple second computing power cerebellum modules. The sending the second computing power perception request to the first computing power cerebellum module based on the first computing power perception request includes: Obtaining the perception activation score of the second computing power cerebellum module, where the perception activation score is used to represent the computing power of the second computing power cerebellum module; Determining a target perception activation score with the largest value among the multiple perception activation scores; Determining the second computing power cerebellum module corresponding to the target perception activation score as the first computing power cerebellum module; Sending the second computing power perception request to the first computing power cerebellum module based on the first computing power perception request.
3. The method according to claim 2, wherein The obtaining the perception activation score of the second computing power cerebellum module includes: Obtaining the second computing power resource information of the second computing power cerebellum module, where the second computing power resource information includes at least one of the following: the computing power resource utilization rate of the second computing power cerebellum module, the number of computing power nodes controlled by the second computing power cerebellum module, and the communication time consumption between the second computing power cerebellum module and the controlled computing power nodes; Calculating the perception activation score of the second computing power cerebellum module according to the second computing power resource information.
4. The method according to any one of claims 1 to 3, characterized in that The computing power-network perception information is computing power-network topology information, and the first computing power resource information includes at least one of the following: the computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module, the total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module, the number of computing power nodes controlled by the first computing power cerebellum module, and the communication time consumption between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
5. The method according to claim 4, characterized in that After determining the computing power-network perception information based on the first computing power resource information, the method further includes: Sending the computing power-network topology information to the task scheduling module, where the task scheduling module is used to determine a target computing power node based on the computing power-network topology information and use the target computing power node to process a target task.
6. A computing-network perception method is applied to a first computing power cerebellum module, and the first computing power cerebellum module is communicatively connected to a computing-network perception module, characterized in that The method includes: Receiving a second computing power perception request sent by the computing power-network perception module; Obtaining the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; Sending the first computing power resource information to the computing power-network perception module, where the computing power-network perception module is used to determine computing power-network perception information based on the first computing power resource information.
7. The method according to claim 6, wherein Obtaining the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request includes: Sending a third computing power perception request to the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; Receiving the first computing power resource information replied by the computing power nodes controlled by the first computing power cerebellum module for the third computing power perception request; Wherein, the first computing power resource information includes at least one of the following: The computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module; The remaining amount information of each computing power resource type information of the computing power nodes controlled by the first computing power cerebellum module; The total computing power resource information of the computing power nodes controlled by the first computing power cerebellum module; The number of computing power nodes controlled by the first computing power cerebellum module; The communication latency between the first computing power cerebellum module and the computing power nodes controlled by the first computing power cerebellum module.
8. A computing-network perception device is applied to a computing-network perception module. The computing-network perception module is respectively communicatively connected to a task scheduling module and a first computing power cerebellum module, and is characterized in that The computing power and network perception device includes: A first receiving module, configured to receive the first computing power perception request sent by the task scheduling module; A first sending module, configured to send a second computing power perception request to the first computing power cerebellum module based on the first computing power perception request, where the first computing power cerebellum module is configured to query the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; A second receiving module, configured to receive the first computing power resource information sent by the first computing power cerebellum module; A determining module, configured to determine the computing power and network perception information based on the first computing power resource information.
9. A computing network perception device is applied to the first computing power cerebellum module, and the first computing power cerebellum module is communicatively connected to the computing network perception module, and is characterized in that, The computing power and network perception device includes: A third receiving module, configured to receive the second computing power perception request sent by the computing power and network perception module; An obtaining module, configured to obtain the first computing power resource information of the computing power nodes controlled by the first computing power cerebellum module based on the second computing power perception request; A second sending module, configured to send the first computing power resource information to the computing power and network perception module, where the computing power and network perception module is configured to determine the computing power and network perception information based on the first computing power resource information.
10. An electronic device, comprising: A transceiver, a memory, a processor, and a program stored on the memory and executable on the processor; characterized in that the processor is configured to read the program in the memory to implement the steps in the computing power and network perception method according to any one of claims 1 to 5, or the processor is configured to read the program in the memory to implement the steps in the computing power and network perception method according to any one of claims 6 to 7.
11. A readable storage medium for storing a program, characterized in that, When the program is executed by the processor, it implements the steps in the computing power and network perception method according to any one of claims 1 to 5, or when the program is executed by the processor, it implements the steps in the computing power and network perception method according to any one of claims 6 to 7.