Privacy computing node optimization method and device based on dynamic routing weight, storage medium and electronic equipment

Abstract

The application provides a privacy computing node optimization method and device based on dynamic routing weights, a storage medium and an electronic device, and relates to the field of privacy computing. After receiving a privacy computing task request, target computing nodes participating in the task can be determined from multiple computing nodes, and a dynamic routing weight matrix is obtained, which records the routing weights between each two computing nodes. The routing weights represent the communication difficulty between the corresponding computing nodes. Since the dynamic routing weight matrix can reflect the actual situation of the communication difficulty between nodes, when planning the communication path, a path with lower communication cost can be selected, so that the node selection can be adjusted according to the network condition, and the demand of the privacy computing task for low delay and high reliability can be met.

Description

Technical Field

[0001] This application relates to the field of privacy computing, and more specifically, to a method, apparatus, storage medium, and electronic device for selecting privacy computing nodes based on dynamic routing weights. Background Technology

[0002] Privacy-preserving computation is a solution for collaborative computation of multi-source data while protecting the data privacy of all parties. Its core idea is to enable multiple participants to jointly complete computational tasks without revealing their original data. In a typical privacy-preserving computation process, each participating node preprocesses its local private data, transmits the encrypted intermediate results to other nodes or an aggregation server through cryptographic protocols or secure aggregation mechanisms, and then performs joint computations based on these encrypted results to obtain the final computational output.

[0003] Therefore, the operation of privacy-preserving computing systems is highly dependent on data interaction between participating nodes. The node routing scheme determines the choice of communication path, directly affecting task execution efficiency. To address this, related technologies have proposed static configuration routing schemes and random or round-robin scheduling schemes. However, in practice, it has been found that these technologies cannot dynamically adapt to changes in actual network conditions, resulting in node selection accuracy and adaptability that are insufficient to meet the actual requirements of privacy-preserving computing tasks for low latency and high reliability. Summary of the Invention

[0004] To overcome at least one deficiency in the prior art, this application provides a method, apparatus, storage medium, and electronic device for selecting privacy computing nodes based on dynamic routing weights. This method can select paths with lower communication costs when planning communication paths, enabling node selection to be adjusted according to changes in network conditions, thereby meeting the requirements of privacy computing tasks for low latency and high reliability.

[0005] In a first aspect, this application provides a method for selecting privacy-preserving computing nodes based on dynamic routing weights, the method comprising: Receive privacy computing task requests and determine multiple target computing nodes to participate in the privacy computing task from multiple computing nodes; Obtain a dynamic routing weight matrix, wherein the dynamic routing weight matrix records the routing weight between every two computing nodes in the plurality of computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes; Based on the dynamic routing weight matrix, communication paths are planned for the multiple target computing nodes.

[0006] Secondly, this application provides a privacy computing node selection device based on dynamic routing weights, the device comprising: The service interaction module is used to receive privacy computing task requests and determine multiple target computing nodes to participate in the privacy computing task from multiple computing nodes. The weight matrix module is used to obtain a dynamic routing weight matrix, wherein the dynamic routing weight matrix records the routing weight between every two computing nodes in the plurality of computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes; The path planning module is used to plan communication paths for the multiple target computing nodes based on the dynamic routing weight matrix.

[0007] Thirdly, this application provides a storage medium storing a computer program that, when executed by a processor, implements the aforementioned privacy computing node optimization method based on dynamic routing weights.

[0008] Fourthly, this application provides an electronic device, which includes a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the aforementioned privacy computing node optimization method based on dynamic routing weights.

[0009] Compared with the prior art, this application has the following beneficial effects: The privacy computing node selection method, apparatus, storage medium, and electronic device based on dynamic routing weights provided in this application, after receiving a privacy computing task request, can determine the target computing node to participate in the task from multiple computing nodes, and obtain a dynamic routing weight matrix that records the routing weight between every two computing nodes. This routing weight characterizes the communication difficulty between the corresponding computing nodes. Since the dynamic routing weight matrix can reflect the actual communication difficulty between nodes, a path with lower communication cost can be selected when planning the communication path, allowing node selection to be adjusted according to changes in network conditions, thus meeting the requirements of privacy computing tasks for low latency and high reliability. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A schematic diagram illustrating the principle of the static configuration routing scheme provided in the embodiments of this application; Figure 2 A schematic diagram illustrating the principle of the random or round-robin scheduling scheme provided in the embodiments of this application; Figure 3 A flowchart illustrating the privacy computing node selection method based on dynamic routing weights provided in this application embodiment; Figure 4 A schematic diagram of the structure of the privacy computing node selection device based on dynamic routing weight provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this application (hereinafter referred to as "the embodiments") clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0013] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0014] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0015] In the description of this application, it should be noted that the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0016] Based on the above statement, as introduced in the background section, related technologies have proposed static configuration routing schemes, random or round-robin scheduling schemes, but none of them can dynamically adapt to changes in network conditions.

[0017] Taking a statically configured routing scheme as an example, such as Figure 1As shown, the scheduling center maintains a fixed routing table through a configuration file. When a task is issued, a random algorithm or a round-robin algorithm is used to select a node. The communication path between nodes depends entirely on the default route of the underlying network.

[0018] In practice, it was found that fixed routing cannot detect changes in network quality between nodes. When a link becomes congested or fails, the system cannot automatically switch to an alternative path, causing tasks to time out or fail. Furthermore, the execution order is fixed, making it difficult to dynamically enable / disable or adjust steps based on different scenarios (such as test environments or different user levels). Adding new features requires modifying core business code, violating the open / closed principle.

[0019] like Figure 2 As described above, in random or round-robin scheduling schemes, the scheduling center maintains a list of nodes and a task queue. When tasks are distributed, a random or round-robin algorithm is used to select nodes, and the communication paths between nodes rely entirely on the default routes of the underlying network. In practice, it has been found that random or round-robin scheduling may assign computational tasks to geographically distant nodes, leading to a significant increase in data transmission latency. For iterative tasks such as federated learning, the cumulative latency is unacceptable.

[0020] Therefore, none of the aforementioned technologies can dynamically adapt to changes in network conditions.

[0021] It should be noted that the defects in the solutions in the prior art are the result of practice and careful research. Therefore, the discovery process of the above problems and the solutions proposed by the embodiments of this application in the following text should be regarded as contributions to this application in the process of invention and creation, and should not be understood as technical content known to those skilled in the art.

[0022] Based on the discovery of the above-mentioned technical problems, this embodiment provides a method for selecting privacy-preserving computing nodes based on dynamic routing weights. For example... Figure 3 As shown, the method includes: S1 receives privacy computing task requests and determines multiple target computing nodes to participate in the privacy computing task from multiple computing nodes.

[0023] S2, obtain the dynamic routing weight matrix.

[0024] The dynamic routing weight matrix records the routing weight between every two computing nodes in multiple computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes.

[0025] S3 plans communication paths for multiple target computing nodes based on the dynamic routing weight matrix.

[0026] This can be understood as follows: after receiving a privacy-preserving computation task request, the server can determine the target computation node to participate in the task from multiple computation nodes and obtain a dynamic routing weight matrix that records the routing weight between each pair of computation nodes. This routing weight represents the communication difficulty between the corresponding computation nodes. Since the dynamic routing weight matrix can reflect the actual communication difficulty between nodes, the server can choose the path with lower communication cost when planning the communication path, allowing the node selection to be adjusted according to changes in network conditions, thus meeting the requirements of privacy-preserving computation tasks for low latency and high reliability.

[0027] It should be noted that the server can be a single server or a group of servers. The server group can be centralized or distributed (e.g., the servers can be a distributed system). In some embodiments, the server can be local or remote relative to the user terminal. In some embodiments, the server can be implemented on a cloud platform; by way of example only, a cloud platform can include private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, inter-cloud, multi-cloud, etc., or any combination thereof. In some embodiments, the server can be implemented on an electronic device having one or more components.

[0028] To make the solution provided in this embodiment clearer, the following is combined with... Figure 3 Each step of the method is described in detail. However, it should be understood that the operations in the flowchart may not be implemented in sequence, and steps without logical contextual relationships may be reversed in order or implemented simultaneously. Furthermore, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0029] See also Figure 3 Next, we will explain step S1 in the diagram in detail: S1 receives privacy computing task requests and determines multiple target computing nodes to participate in the privacy computing task from multiple computing nodes.

[0030] In practical applications, when a server receives a privacy computing task request, it can parse the task requirements to obtain a list of participating nodes, denoted as { , ,..., The computing nodes in this list are the multiple target computing nodes identified in this privacy computing task.

[0031] For example, the privacy-preserving computation task can be a multi-party secure computation task or a federated learning task. Taking a federated learning task as an example, after receiving a task request for federated learning, the server can determine multiple participating nodes that need to collaborate to complete model training from all registered computing nodes, and use them as the target computing nodes for this task.

[0032] Based on the above descriptions of privacy-preserving computation tasks and target computing nodes in the embodiments, the following will continue to discuss... Figure 3 Step S2 will be explained in detail: S2, obtain the dynamic routing weight matrix.

[0033] The dynamic routing weight matrix records the routing weight between every two computing nodes in multiple computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes.

[0034] It should be noted that in the field of privacy computing, when multiple target computing nodes collaborate to complete a computing task, the quality of data transmission between nodes directly affects the overall task execution efficiency. However, related technologies do not perceive or quantify the actual communication status between each computing node. Factors such as the physical distance between nodes, network latency and packet loss, and whether they belong to the same network operator are not comprehensively included in the routing decision-making process. Therefore, it is impossible to predict in advance which communication links between nodes are more efficient and which have higher transmission overhead, which can easily lead to increased communication latency or even task failure due to improper path selection.

[0035] To address the above problems, this embodiment provides the following optional implementation methods for step S2: S2-1, obtain the physical distance between each pair of computing nodes, network quality indicators, and operator matching coefficients among multiple computing nodes.

[0036] The operator matching coefficient indicates whether the network operators of the corresponding computing nodes are the same.

[0037] In practical applications, each computing node can call the Map Application Programming Interface (API) upon startup to obtain the latitude and longitude coordinates of its location in the data center or cloud availability zone, and report this geographic location information to the pre-deployed Eureka service registry. When the server needs to obtain the physical distance, it can read the latitude and longitude information of each node from the registry and use the geodesy geocomputing library to calculate the spherical distance between any two nodes, thereby constructing an N×N static distance matrix D, where the matrix elements can be represented as: , This element represents the node. With nodes The physical distance between them.

[0038] Meanwhile, the server can periodically perform network quality probes through an independent thread pool, sending Transmission Control Protocol (TCP) probe packets or Internet Control Message Protocol (ICMP) Echo requests to each compute node to obtain network quality metrics, which may include network latency. Packet loss rate And network jitter Among them, network latency Indicates round-trip time and packet loss rate. Jitter represents the ratio of packets sent to packets received. The standard deviation represents the delayed variation.

[0039] In addition, the server can also obtain the carrier matching coefficient. The carrier matching coefficient here This serves as a quantitative indicator to determine whether two compute nodes share the same network operator. For example, when two compute nodes belong to the same network operator, the operator matching coefficient... The value can be 0, which is the operator matching coefficient when the two compute nodes belong to different network operators. It can take the value 1.

[0040] Based on the above embodiments regarding physical distance, network quality indicators, and operator matching coefficients, step S2 further includes: S2-2, For every two computing nodes, the routing weight between the two computing nodes is obtained based on the physical distance between the two computing nodes, network quality indicators, and operator matching coefficient.

[0041] As an optional implementation, the server can normalize the physical distance and network quality indicators to obtain normalized physical distance and normalized network quality indicators; and then weight and fuse the normalized physical distance, normalized network quality indicators and operator matching coefficients to obtain the routing weight between the two computing nodes.

[0042] Optionally, network quality metrics include at least one of network latency, packet loss rate, and network jitter.

[0043] In practical applications, after obtaining the physical distance, network quality indicators, and carrier matching coefficients between each pair of computing nodes, the server can further process this data to obtain routing weights. However, the dimensions and numerical ranges of indicators such as physical distance, network latency, and packet loss rate are different, and there is no unified benchmark for direct numerical comparison or calculation. Therefore, the server can normalize the physical distance and network quality indicators separately, mapping them to the interval between zero and one, thereby obtaining normalized physical distance and normalized network quality indicators.

[0044] For example, the normalized physical distance can be denoted as... Network latency in the normalized network quality index can be denoted as... Packet loss rate can be denoted as... .

[0045] After normalization, the server can weight and fuse the normalized physical distance, normalized network quality index, and carrier matching coefficient to obtain the routing weight between the two computing nodes. Specifically, the server can calculate the node using the following formula. With nodes Routing weights between :

[0046] In the formula, For normalized physical distance, To normalize network latency, To normalize the packet loss rate, For operator matching coefficient, , , , The weights are configurable and satisfy the following conditions: In practical applications, the aforementioned weighting coefficients can be dynamically adjusted through the configuration center, thereby supporting flexible switching of scheduling strategies according to different business scenarios. For example, in scenarios with high real-time requirements, the weighting coefficient corresponding to network latency can be increased, while in scenarios where cost control is the priority, the weighting coefficient corresponding to physical distance can be increased.

[0047] Based on the above description of routing weights in the embodiments, step S2 further includes: S2-3: Construct a dynamic routing weight matrix based on the routing weights between every two computing nodes.

[0048] In practical applications, after the server calculates the routing weights between every two computing nodes, it can arrange and combine these routing weights according to the node correspondence to construct a dynamic routing weight matrix.

[0049] For example, assume that there are a total of computing nodes currently participating in the privacy computing task. If there are one, then the server can build one The matrix, where the first... Line number The elements of a column are nodes. With nodes Routing weights between This matrix can comprehensively reflect the communication costs between computing nodes across multiple dimensions, including physical distance, network transmission quality, and carrier compatibility.

[0050] It should be noted that since network quality metrics change over time, such as network latency increasing due to link congestion and packet loss rate increasing due to network fluctuations, the server can periodically re-execute network quality probing and update routing weights so that the dynamic routing weight matrix can reflect the current communication status between computing nodes in real time.

[0051] In practical applications, the server can update the calculated routing weights to the Redis cache in real time. This way, when planning communication paths for privacy computing tasks, the latest dynamic routing weight matrix can be read directly from Redis without recalculation, thereby improving the efficiency of path planning.

[0052] Based on the above description of the dynamic routing weight matrix in the embodiments, step S3 will be further explained below: S3 plans communication paths for multiple target computing nodes based on the dynamic routing weight matrix.

[0053] Regarding step S3, it should be noted that in the field of privacy computing, for point-to-point transmission tasks, related technologies often fail to comprehensively consider the actual communication costs between nodes, making it difficult to find the path with the minimum cumulative transmission overhead. For multi-point collaborative tasks, related technologies only focus on the connectivity between some nodes, failing to guarantee that an effective communication connection can be established between any two target computing nodes, and also failing to control the overall communication cost to a low level.

[0054] To address the above problems, this embodiment provides the following optional implementation methods for step S3: S3-1, If ​​the privacy computing task is a point-to-point transmission task, then according to the dynamic routing weight matrix, under the first constraint, a communication path is planned for the source node and the destination node among multiple target computing nodes.

[0055] The first constraint is that the cumulative weight generated by the communication path is minimized.

[0056] The point-to-point transmission task here can be interpreted as a transmission mode in which data is transmitted from a source node to a destination node. For example, in a multi-party secure computation scenario, when it is necessary to transfer secret sharing fragments from one participant node to another participant node, it is a typical point-to-point transmission task.

[0057] During this process, the server can read the routing weights between each computing node from the dynamic routing weight matrix and use these routing weights as the cost basis for path search to plan a communication path between the source node and the destination node. This communication path can be a sequence of nodes formed by the source node passing through one or more intermediate nodes to finally reach the destination node. The cumulative weight generated by this communication path can be interpreted as the sum of the routing weights between each pair of adjacent nodes on the path. The first constraint followed by the server when planning the communication path is to minimize the cumulative weight generated by that communication path.

[0058] For example, the server can use Dijkstra's shortest path algorithm, which uses the routing weights between each pair of nodes in the dynamic routing weight matrix as the weights of the edges, and searches step by step from the source node to the destination node, thereby finding the communication path with the minimum cumulative weight.

[0059] Based on the above embodiments regarding point-to-point transmission tasks, step S3 further includes: S3-2, If the privacy computing task is a multi-point collaborative task, then according to the dynamic routing weight matrix, communication paths are planned for multiple target computing nodes under the second constraint.

[0060] The second constraint is that the communication path can establish a communication connection between any two target computing nodes and the resulting cumulative weight is minimized.

[0061] The multi-point collaborative task here can be interpreted as a collaborative computing mode that requires the joint participation of multiple computing nodes and data interaction between them. For example, in a multi-party secure computing scenario, when multiple participating nodes need to collaborate to complete a secure intersection operation, each participating node needs to establish a data transmission channel, which is a multi-point collaborative task.

[0062] In this process, the second constraint followed by the server comprises two levels. First, the planned communication path must be able to establish a communication connection between any two target computing nodes. This means that regardless of which two target computing nodes are selected, there should be a reachable data transmission path between them, and there should be no situation where some node pairs cannot communicate with each other. Second, under the premise of satisfying the above connectivity requirements, the sum of the cumulative weights generated by all communication paths should be as small as possible. Here, the cumulative weight can be understood as the sum of the routing weights between each adjacent node pair along the communication path.

[0063] For example, to satisfy the second constraint mentioned above, the server can use the minimum spanning tree algorithm to construct the connection topology. Specifically, the server can treat multiple target computing nodes as vertices in a graph, and the routing weights between each pair of nodes in the dynamic routing weight matrix as the weights of the edges connecting the corresponding vertices. Then, it can use the Prim algorithm or the Kruskal algorithm to select several edges between all vertices, such that the selected edges can connect all vertices and minimize the sum of the edge weights.

[0064] Thus, the connection topology constructed by the minimum spanning tree algorithm not only ensures that there is a communication connection between any two target computing nodes, but also keeps the overall communication cost at a minimum.

[0065] Based on the communication path obtained from the above embodiments, the server can also distribute the communication path to each target computing node.

[0066] In practice, when the privacy-preserving computation task is a point-to-point transmission task, each target computing node can transmit data from the source node to the destination node sequentially through the intermediate nodes along the communication path issued by the server. When the privacy-preserving computation task is a multi-point collaborative task, each target computing node can establish a data transmission channel with each other according to the connection topology issued by the server, thereby completing collaborative computation among multiple parties.

[0067] It should be noted that, in addition to the path planning method based on the dynamic routing weight matrix mentioned above, other alternative solutions can be used to achieve node selection and path optimization in privacy computing scenarios.

[0068] For example, a reinforcement learning scheduling scheme can be used. In this scheme, the server can use a reinforcement learning model to train the scheduling strategy, using historical communication data between nodes as the model's state input and path selection as the model's action output, and continuously optimizing the scheduling decision through a reward function. In this way, the model can gradually learn a better path selection strategy over a long period of operation.

[0069] In addition, a hierarchical regional aggregation scheme can be adopted. In this scheme, the server can divide geographically proximate computing nodes into different regions. Nodes within a region communicate first, while data is aggregated between regions through a pre-designated regional central node before cross-regional transmission. This reduces the frequency of cross-regional communication, thereby lowering overall communication costs.

[0070] Alternatively, the Anycast broadcast routing scheme can be adopted. In this scheme, multiple computing nodes can share the same Internet Protocol (IP) address. After the sender sends data to this IP address, the underlying network infrastructure can automatically select the nearest node to receive it. In this way, routing decisions are automatically completed at the network device level, and the application layer server does not need to participate in the specific path calculation process.

[0071] In addition, a centralized routing scheme based on Software Defined Networking (SDN) can be adopted. In this scheme, the SDN controller can centrally manage the routing tables of the entire network and dynamically distribute the optimal flow tables to each switch based on network conditions, thereby achieving path optimization at the network layer. Thus, the centralized scheduling by the SDN controller improves the transmission efficiency at the network layer.

[0072] Based on the same inventive concept as the privacy computing node selection method based on dynamic routing weights provided in this embodiment, this embodiment also provides a privacy computing node selection device based on dynamic routing weights. This device includes at least one software functional module that can be stored in memory or embedded in an electronic device. The processor in the electronic device executes the executable module stored in memory. For example, the software functional modules and computer programs included in this device. Please refer to... Figure 4 Functionally, the device may include: Service interaction module 11 is used to receive privacy computing task requests and determine multiple target computing nodes to participate in the privacy computing task from multiple computing nodes; The weight matrix module 12 is used to obtain the dynamic routing weight matrix, which records the routing weight between every two computing nodes in multiple computing nodes. The routing weight represents the communication difficulty between the corresponding computing nodes. The path planning module 13 is used to plan communication paths for multiple target computing nodes based on the dynamic routing weight matrix.

[0073] Optionally, the weight matrix module 12 obtains the dynamic routing weight matrix in the following ways: Obtain the physical distance, network quality index, and carrier matching coefficient between every two computing nodes in multiple computing nodes. The carrier matching coefficient indicates whether the network carriers of the corresponding computing nodes are the same. For every two computing nodes, the routing weight between the two computing nodes is obtained based on the physical distance between the two computing nodes, network quality indicators, and operator matching coefficient. A dynamic routing weight matrix is ​​constructed based on the routing weights between every two computing nodes.

[0074] Optionally, the weight matrix module 12 obtains the routing weights between two computing nodes based on the physical distance between them, network quality metrics, and operator matching coefficients, including: The physical distance and network quality indicators are normalized to obtain normalized physical distance and normalized network quality indicators. The routing weight between the two computing nodes is obtained by weighting and fusing the normalized physical distance, normalized network quality index, and operator matching coefficient.

[0075] Optionally, the path planning module 13 plans communication paths for multiple target computing nodes based on a dynamic routing weight matrix, including: If the privacy computing task is a peer-to-peer transmission task, then based on the dynamic routing weight matrix, communication paths are planned for the source and destination nodes among multiple target computing nodes under the first constraint condition, where the cumulative weight generated by the communication path is minimized.

[0076] Optionally, the path planning module 13 plans communication paths for multiple target computing nodes based on a dynamic routing weight matrix, including: If the privacy computing task is a multi-point collaborative task, then based on the dynamic routing weight matrix, communication paths are planned for multiple target computing nodes under the second constraint. The second constraint is that the communication path can establish a communication connection between any two target computing nodes and generates the minimum cumulative weight.

[0077] Optionally, the path planning module 13 is also used to distribute the communication path to each target computing node.

[0078] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0079] It should also be understood that if the above embodiments are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0080] Therefore, this embodiment also provides a storage medium, which is a computer-readable storage medium. This storage medium stores a computer program, which, when executed by a processor, implements the privacy computing node selection method based on dynamic routing weights provided in this embodiment. The storage medium can be any medium capable of storing program code, such as a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0081] Please refer to Figure 5 The electronic device provided in this embodiment may include a processor 22 and a memory 21. The memory 21 stores a computer program, and the processor implements the privacy computing node selection method based on dynamic routing weights provided in this embodiment by reading and executing the computer program corresponding to the above embodiments in the memory 21.

[0082] See also Figure 5 The electronic device also includes a communication unit 23. The memory 21, processor 22 and communication unit 23 are electrically connected to each other directly or indirectly through system bus 24 to realize data transmission or interaction.

[0083] The memory 21 can be an information recording device based on any electronic, magnetic, optical, or other physical principles, used to record execution instructions, data, etc. In some embodiments, the memory 21 can be, but is not limited to, volatile memory, non-volatile memory, memory drive, etc.

[0084] In some embodiments, the volatile memory may be random access memory (RAM); in some embodiments, the non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, etc.; in some embodiments, the storage drive may be a disk drive, solid-state drive, any type of storage disk (such as optical disc, DVD, etc.), or similar storage media, or a combination thereof.

[0085] The communication unit 23 is used to send and receive data over a network. In some embodiments, the network may include a wired network, a wireless network, a fiber optic network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, or a near field communication (NFC) network, or any combination thereof. In some embodiments, the network may include one or more network access points. For example, the network may include wired or wireless network access points, such as base stations and / or network switching nodes, through which one or more components of the service request processing system can connect to the network to exchange data and / or information.

[0086] The processor 22 may be an integrated circuit chip with signal processing capabilities, and may include one or more processing cores (e.g., a single-core processor or a multi-core processor). By way of example only, the processor described above may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application Specific Instruction-set Processor (ASIP), a Graphics Processing Unit (GPU), a Physics Processing Unit (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a controller, a microcontroller unit, a Reduced Instruction Set Computing (RISC) computer, or a microprocessor, or any combination thereof.

[0087] Understandable. Figure 5The structure shown is for illustrative purposes only. Electronic devices may also have more advanced features. Figure 5 Showing more or fewer components, or having with Figure 5 The different configurations shown. Figure 5 The components shown can be implemented using hardware, software, or a combination thereof.

[0088] It should be understood that the apparatus and methods disclosed in the above embodiments can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0089] The above descriptions are merely various embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for selecting privacy-preserving computation nodes based on dynamic routing weights, characterized in that, The method includes: Receive privacy computing task requests and determine multiple target computing nodes to participate in the privacy computing task from multiple computing nodes; Obtain a dynamic routing weight matrix, wherein the dynamic routing weight matrix records the routing weight between every two computing nodes in the plurality of computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes; Based on the dynamic routing weight matrix, communication paths are planned for the multiple target computing nodes.

2. The privacy computing node selection method based on dynamic routing weights according to claim 1, characterized in that, Obtain the dynamic routing weight matrix, including: Obtain the physical distance, network quality index, and operator matching coefficient between every two computing nodes in the plurality of computing nodes, wherein the operator matching coefficient indicates whether the network operators of the corresponding computing nodes are the same; For every two computing nodes, the routing weight between the two computing nodes is obtained based on the physical distance between the two computing nodes, network quality indicators, and operator matching coefficient. The dynamic routing weight matrix is ​​constructed based on the routing weights between every two computing nodes.

3. The privacy computing node selection method based on dynamic routing weights according to claim 1, characterized in that, Based on the physical distance between the two computing nodes, network quality indicators, and carrier matching coefficients, the routing weight between the two computing nodes is obtained, including: The physical distance and the network quality index are normalized respectively to obtain normalized physical distance and normalized network quality index; The normalized physical distance, the normalized network quality index, and the operator matching coefficient are weighted and fused to obtain the routing weight between the two computing nodes.

4. The privacy computing node selection method based on dynamic routing weights according to claim 2 or 3, characterized in that, The network quality metrics include at least one of network latency, packet loss rate, and network jitter.

5. The privacy computing node selection method based on dynamic routing weights according to claim 1, characterized in that, Based on the dynamic routing weight matrix, communication paths are planned for the multiple target computing nodes, including: If the privacy computing task is a peer-to-peer transmission task, then according to the dynamic routing weight matrix, a communication path is planned for the source node and the destination node among the multiple target computing nodes under the first constraint condition, wherein the first constraint condition is that the cumulative weight generated by the communication path is minimized.

6. The privacy computing node selection method based on dynamic routing weights according to claim 1, characterized in that, Based on the dynamic routing weight matrix, communication paths are planned for the multiple target computing nodes, including: If the privacy computing task is a multi-point collaborative task, then according to the dynamic routing weight matrix, a communication path is planned for the multiple target computing nodes under the second constraint condition, wherein the second constraint condition is that the communication path can establish a communication connection between any two of the target computing nodes and the generated cumulative weight is minimized.

7. The privacy computing node selection method based on dynamic routing weights according to claim 1, characterized in that, The method further includes: The communication path is sent to each of the target computing nodes.

8. A privacy computing node selection device based on dynamic routing weights, characterized in that, The device includes: The service interaction module is used to receive privacy computing task requests and determine multiple target computing nodes to participate in the privacy computing task from multiple computing nodes. The weight matrix module is used to obtain a dynamic routing weight matrix, wherein the dynamic routing weight matrix records the routing weight between every two computing nodes in the plurality of computing nodes, and the routing weight represents the communication difficulty between the corresponding computing nodes; The path planning module is used to plan communication paths for the multiple target computing nodes based on the dynamic routing weight matrix.

9. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the privacy computing node selection method based on dynamic routing weights as described in any one of claims 1-7.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the privacy computing node selection method based on dynamic routing weights as described in any one of claims 1-7.

Images