Distributed storage method and system based on AI flow identification

Abstract

The invention discloses a distributed storage method and system based on AI flow identification, and the method comprises the steps: obtaining the state information of a current CDN network node, simulating the construction of a first GNN node in a GNN model, and constructing an adjacent node of the GNN model according to the communication relation between the current CDN network node and different CDN network nodes; acquiring communication flow information between a first GNN node and a corresponding adjacent node, identifying flow characteristic data in the communication flow information by using an AI model, and constructing a dynamic adjacent matrix between the first GNN node and the adjacent node; using a dynamic adjacency matrix to update the states of the first GNN node and the adjacent node, and executing a caching strategy or a flow processing strategy of the corresponding CDN network node according to the updated states of the first GNN node and the adjacent node; and routing the node of the current CDN network to the node of the optimal CDN network according to the caching strategy or the flow processing strategy.

Description

Technical Field

[0001] The present invention relates to the field of distributed storage technology, and in particular to a distributed storage method and system based on AI traffic identification. Background Art

[0002] The existing traditional distributed storage method achieves data redundancy and backup by storing data on servers in multiple physical locations. Even if a server fails, data copies stored on other servers are still available, thus ensuring data persistence and system stability. The above traditional distributed storage does not conduct a comprehensive analysis of traffic. For example, data traffic with network threats or data traffic with illegal content are stored according to the distributed data sharding method, resulting in low security and cleanliness of stored data, which is easy to cause malicious occupation of server storage resources when the server node is attacked. At the same time, in the process of storage resource scheduling, traditional distributed storage does not effectively analyze data traffic, resulting in low distributed storage efficiency. Summary of the invention

[0003] One of the inventive purposes of the present invention is to provide a distributed storage method and system based on AI traffic identification, wherein the method and system utilize AI models to identify traffic data of different storage nodes, wherein the present invention utilizes AI models to identify network attack data and illegal content data that may exist in the traffic data, and intercepts or schedules the traffic data based on the identified network attack data and illegal content data, thereby achieving targeted processing of the network attack data and illegal content data traffic, improving the data security of distributed storage, reducing the occupation of distributed storage resources by illegal content data, and improving the effect of distributed data storage.

[0004] Another inventive object of the present invention is to provide a distributed storage method and system based on AI traffic identification, wherein the method and system utilize CDN network (content distribution network) as a distributed storage system, utilize AI model to identify and analyze the traffic of different nodes of the CDN network, obtain network attack data and illegal content data of the traffic of different nodes of the CDN network, and perform corresponding CDN network node caching strategy based on the network attack data and illegal content data; the present invention utilizes the efficient local access of the CDN network node cache while also taking into account the security of access, and improves the compliance effect of accessing data content, and reduces the occupation of bandwidth resources or illegal content data cache resources by illegal access.

[0005] Another inventive object of the present invention is to provide a distributed storage method and system based on AI traffic identification, wherein the method and system utilize a GNN model (graph neural network model) to simulate node behavior of the CDN network nodes, including but not limited to traffic data, wherein the CDN network node-related data is used as the node features in the GNN model, and the results of the AI ​​analysis of traffic data between different CDN network nodes are used as the connection edge features (adjacency matrix) between different nodes in the GNN model, and the GNN model is used to construct a dynamic adjacency matrix based on the traffic features identified by the AI ​​model to perform data updates or cache strategy executions on the corresponding CDN network nodes; the present invention utilizes the description method of different node relationships between graph nodes of the GNN model to effectively solve the description method of traffic data between different CDN network nodes, thereby efficiently achieving the technical effect of secure and compliant distributed storage.

[0006] In order to achieve at least one of the above-mentioned invention objects, the present invention further provides a distributed storage method based on AI traffic identification, the method comprising:

[0007] Obtain current CDN network node status information, simulate and construct a first GNN node in the GNN model according to the current CDN network node status information, and construct adjacent nodes of the GNN model according to the communication relationship between the current CDN network node and different CDN network nodes;

[0008] Acquire communication flow information between the first GNN node and the corresponding adjacent node, identify flow characteristic data in the communication flow information using an AI model, and construct a dynamic adjacency matrix between the first GNN node and the adjacent node according to the flow characteristic data;

[0009] Using the dynamic adjacency matrix to update the status of the first GNN node and the adjacent nodes, and executing the caching strategy or traffic processing strategy of the corresponding CDN network node according to the updated status of the first GNN node and the adjacent nodes;

[0010] Obtain the current CDN network node routing link table, and route the current CDN network node to the optimal CDN network node according to the cache strategy or traffic processing strategy.

[0011] According to one of the preferred embodiments of the present invention, the current CDN network node status information includes hardware resource status, network resource status and cache status, wherein the hardware resource status includes: CPU utilization, memory utilization, cache disk read and write performance data; the network resource status includes: the upstream and downstream bandwidth utilization of the current CDN network node, the packet loss rate of the current CDN network node, the network request rate, the number of concurrent connections and the number of URLs; the cache status includes: cache hit rate, cache elimination rate, cache preheating status and cache content type; the hardware resource status, network resource status and cache status are respectively preprocessed to obtain the feature matrix of the first GNN node, and the dynamic adjacency matrix is ​​constructed according to the traffic feature data between the first GNN node and the corresponding adjacent node identified by the AI ​​model.

[0012] According to another preferred embodiment of the present invention, the method for constructing the dynamic adjacency matrix includes: using an image recognition AI model, a text recognition AI model, a speech recognition AI model and a network attack recognition AI model to identify and obtain network traffic feature data between the first GNN node and the corresponding adjacent node; wherein the traffic feature data includes: illegal video frame data, illegal voice data, illegal text data, network attack type and quantity; pre-processing the different traffic feature data and converting them into the dynamic adjacency matrix, and updating the status of the first GNN node and the corresponding adjacent node according to the dynamic adjacency matrix.

[0013] According to another preferred embodiment of the present invention, the state update method of the first GNN node and the corresponding adjacent node includes: constructing a multi-head attention coefficient according to the traffic feature data obtained by identifying the traffic with the AI ​​model , where i represents the current first GNN node, j represents the adjacent node that has a communication relationship with the first GNN node, and h i represents the first GNN node feature, a and W represent different learnable parameters, where W is the weight matrix, a is the action parameter between nodes based on the transformation of traffic feature data, T represents the transposed matrix identifier, LeakyReLU represents the activation function, Softmax represents the linear classification function, and the attention coefficient α ij Represents the similarity feature weights of the first GNN node and the corresponding adjacent node.

[0014] According to another preferred embodiment of the present invention, the state updating method of the first GNN node and the corresponding adjacent node further includes: ij Parallel computation is performed to capture the relationship type between the first GNN node i and the adjacent node j, through the multi-head attention coefficient α ijThe content of calculating the updated first GNN node includes: using the multi-head attention coefficient α ij Construct a dynamic adjacency matrix A dynamic [i, j] k = , where k represents the corresponding attention head, and the dynamic adjacency matrix A dynamic [i, j] k Replace the original static adjacency matrix A static After replacement, the first GNN node is updated according to the following formula: ,in represents the first GNN node state predicted after the update, N represents the node relationship function, K represents the total number of attention heads, σ is the activation function, and the first GNN node state predicted after the update is obtained. Execute the node cache strategy or traffic processing strategy corresponding to the current CDN network.

[0015] According to another preferred embodiment of the present invention, the cache strategy or traffic processing strategy includes: according to the updated predicted first GNN node state Update the predicted cache state of the node in the current CDN network to allow the corresponding adjacent node data to be cached; according to the updated predicted first GNN node state The network attack data from the corresponding adjacent nodes predicted by the nodes of the current CDN network is limited or transferred to the nodes of a specific CDN network; according to the first GNN node state predicted after the update The node of the current CDN network predicts that the violation data from the corresponding adjacent node will be limited or prohibited from access and cache strategy.

[0016] According to another preferred embodiment of the present invention, the CDN network node includes a central CDN network node and an edge CDN network node. After obtaining the current CDN network node routing link table, the central CDN network node determines the edge CDN network node closest to the current CDN network node according to the current CDN network node routing link table, and transmits the data of the adjacent node to the closest edge CDN network node through routing conversion according to the caching strategy, thereby realizing efficient processing of cached data.

[0017] According to another preferred embodiment of the present invention, the method for calculating the distance L between the edge CDN network node closest to the current CDN network node includes: obtaining the round-trip time RTT between the current CDN network node and any edge CDN network node, and calculating the number of hops Hops to the edge CDN network node, and calculating the bandwidth utilization P of the current CDN network node. Then, the calculation formula for the distance L between the edge CDN network node closest to the current CDN network node is: , where β, λ and γ are weight coefficients respectively.

[0018] In order to achieve at least one of the above-mentioned invention purposes, the present invention further provides a distributed storage system based on AI traffic identification, and the system executes the above-mentioned distributed storage method based on AI traffic identification.

[0019] The present invention further provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the above-mentioned distributed storage method based on AI traffic identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Shown is a flow chart of a distributed storage method based on AI traffic identification of the present invention. DETAILED DESCRIPTION

[0021] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0022] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0023] Please combine Figure 1, the present invention discloses a distributed storage method and system based on AI traffic identification, the method includes: first, it is necessary to obtain CDN network node status information, and perform feature conversion simulation to construct GNN nodes (graph neural network nodes) according to the CDN network node status information, wherein the CDN network node is a content distribution network node, and the CDN network node includes a central CDN network node and an edge CDN network node, wherein the central CDN network node is equivalent to a main control node, which is used to monitor multiple edge CDN network nodes and cache content distribution for multiple edge CDN network nodes, wherein the central CDN network node can also perform traffic control between multiple edge network nodes, or distribute the corresponding edge CDN network node to a specific CDN network node. While the present invention uses the CDN network node for distributed storage, different cache strategies and traffic control strategies are implemented according to the traffic characteristics between different CDN network nodes, especially the cache strategy and traffic control strategy for traffic and illegal data including network attacks. The present invention uses the GNN model (graph neural network model) as the driving model of the distributed storage strategy, and uses the GNN model to simulate the state of CDN network nodes and the relationship between different CDN network nodes, wherein the adjacency relationship based on the GNN model (equivalent to the connection edge of different GNN model nodes) is constructed according to the traffic feature data between different CDN network nodes, and the state change of the GNN model node is judged according to the adjacency relationship, and the corresponding CDN network node cache strategy and traffic control strategy are executed based on the updated predicted GNN model node state. Therefore, the present invention can effectively improve the security performance in the distributed cache, effectively reduce the malicious occupation of distributed storage resources caused by abnormal traffic, and improve the resource utilization of distributed storage.

[0024] Specifically, the method for obtaining the CDN network node status information to construct the GNN model node specifically includes: obtaining the CDN network node status information including: hardware resource status, network resource status and cache status, wherein the hardware resource status includes but is not limited to the CPU usage rate, GPU usage rate, memory usage rate, cache disk read and write performance data of the current CDN network node, and the hardware resource status is used to represent the overall performance of the edge device where the current CDN network node is located. Generally speaking, the CDN network node of the edge device is usually used as a distributed cache hardware resource, and the performance of the hardware devices deployed by different edge devices is different. The network resource status includes but is not limited to the upstream and downstream bandwidth usage rate of the current CDN network node, the packet loss rate of the current CDN network node, the network request rate, the number of concurrent connections and the number of URLs. The network resource status of the CDN network node is automatically identified according to the changes in the network environment of the current CDN network node. For example, there may be insufficient network bandwidth and a decrease in network request speed during the high concurrent interaction of multiple CDN network nodes. The identification of the above network resource status can provide reference data for the CDN network node cache strategy and traffic restriction strategy to a certain extent, thereby improving the distributed storage efficiency of the CDN network node as a whole. The cache status includes but is not limited to cache hit rate, cache elimination rate, cache preheating status and cache content type, etc.; the cache hit rate is used to describe the performance of the cache system, especially in specific business scenarios. If the cache hit performance is high, it means that the cache system of the CDN network node has a better distribution effect on the current business content request, and can achieve high adaptability of the edge device CDN network node and the corresponding business content. Therefore, using the cache hit rate as the state parameter of the GNN graph node can better describe the adaptability of different CDN network nodes to related business traffic. Similarly, the state parameters of the GNN graph node converted based on the cache elimination rate, cache preheating status and cache content type can describe the adaptation effect of the corresponding edge device CDN network node and the corresponding business content, which will not be described in detail in the present invention.

[0025] Furthermore, after obtaining the relevant status data of the hardware resource status, network resource status and cache status of the current CDN network node, the present invention adopts data preprocessing including but not limited to normalization, standardization and feature conversion, so that the relevant status data of the current CDN network node can be more standardized, wherein the present invention classifies the standardized relevant status data using label encoding (labelEncoder) to construct different features. The above-mentioned label encoding (labelEncoder) method is an existing method for constructing classification features, and the present invention will not go into details. The relevant status data of the hardware resource status, network resource status and cache status of the current CDN network node are constructed using the label encoding (labelEncoder) method to construct the GNN graph node feature matrix, and the feature matrix is ​​defined as the feature matrix of the first GNN node, which is used to describe the relevant status data of the current CDN network node.

[0026] It is worth mentioning that one of the core technical points of the present invention is to construct an adjacency matrix of the first GNN node, wherein the adjacency matrix of the first GNN node is a matrix that describes the relationship between the first GNN node and the adjacent nodes, wherein the adjacent nodes and the first GNN node are defined in the present invention as CDN network nodes with a communication relationship, especially CDN network nodes with a content distribution scheduling relationship. The adjacency matrix contains the traffic characteristic data between the first GNN node and the adjacent nodes, and the adjacency matrix constructed by the traffic characteristic data is used to update the predicted state changes of the first GNN node and the adjacent nodes themselves. It should be noted that the traffic data can be from the first GNN node to the adjacent node, or from the adjacent node to the first GNN node. The above-mentioned defined names can be converted according to the actual traffic transmission situation, and the present invention is no longer specifically limited to this.

[0027] The present invention needs to identify the traffic feature data present in the traffic data. Since the traffic data may contain traffic data including but not limited to network attack data, illegal videos, audios and texts, the present invention needs to utilize the existing image recognition AI model, text recognition AI model, speech recognition AI model and network attack recognition AI model to identify and obtain the network traffic feature data between the first GNN node and the corresponding adjacent node, and use the traffic features identified by the existing different AI models to output the traffic feature label (label), and further construct the traffic feature matrix based on the traffic feature label (label) using the same label encoding (labelEncoder) method mentioned above. The present invention uses the traffic feature matrix as the adjacency matrix between the first GNN node and the adjacent nodes to drive the status update of the first GNN node and the adjacent nodes. The traffic feature labels include but are not limited to illegal video frame data, illegal voice data, illegal text data, network attack types and quantities, etc.

[0028] It should be noted that the adjacency matrix of the GNN network node in the traditional GNN model is generally a static matrix used for message passing to update the state of the GNN network node. For example, the adjacency matrix message passing mechanism formula of the traditional GNN model is: Of which A Static is a regular static adjacency matrix, is the GNN node state before updating, It represents the predicted GNN node state after update, which exists in the form of node state feature matrix; N represents the node relationship function; i and j represent different nodes respectively; W is a learnable parameter, which is generally a weight matrix corresponding to the node state feature.

[0029] In order to adapt to the dynamic traffic characteristics and dynamically update the GNN network node status, the present invention replaces the conventional static adjacency matrix A with a dynamic adjacency matrix A based on traffic characteristic data. dynamic [i, j] k , where the dynamic adjacency matrix A dynamic [i, j] k The implementation method includes: constructing a multi-head attention coefficient according to the traffic feature data obtained by identifying the traffic with the AI ​​model , where i represents the current first GNN node, j represents the adjacent node that has a communication relationship with the first GNN node, and h irepresents the first GNN node feature, a and W represent different learnable parameters, where W is the weight matrix, a is the action parameter between nodes based on the transformation of traffic feature data, T represents the transposed matrix identifier, LeakyReLU represents the activation function, Softmax represents the linear classification function, and the attention coefficient α ij represents the similar feature weights of the first GNN node and the corresponding adjacent node, and further converts the multi-head attention coefficient α constructed based on the traffic feature data ij Construct the dynamic adjacency matrix A dynamic [i, j] k =α ij , where due to the multi-head attention coefficient α ij It shows the influence of different types of traffic characteristics on the update of the first GNN node and the corresponding adjacent nodes, so the dynamic adjacency matrix A can be constructed directly by parameter conversion dynamic [i, j] k .

[0030] Furthermore, the present invention requires the multi-head attention coefficient α ij Parallel computation is performed to capture the relationship type between the first GNN node i and the adjacent node j, through the multi-head attention coefficient α ij The method for calculating the updated first GNN node includes: using the multi-head attention coefficient α ij Construct a dynamic adjacency matrix A dynamic [i, j] k = , where k represents the corresponding attention head, and the dynamic adjacency matrix A dynamic [i, j] k Replace the original static adjacency matrix A static After replacement, the first GNN node is updated according to the following formula: ,in represents the first GNN node state predicted after the update, N represents the node relationship function, K represents the total number of attention heads, σ is the activation function, and the first GNN node state predicted after the update is obtained. Execute the node cache strategy or traffic processing strategy corresponding to the current CDN network.

[0031] The cache strategy or traffic processing strategy includes: according to the updated predicted first GNN node state Update the predicted cache state of the node in the current CDN network to allow the corresponding adjacent node data to be cached; according to the updated predicted first GNN node state The network attack data from the corresponding adjacent nodes predicted by the nodes of the current CDN network is limited or transferred to the nodes of a specific CDN network; according to the first GNN node state predicted after the update The node prediction of the current CDN network from the corresponding adjacent node violation data includes flow limiting or access prohibition and cache strategy. The above cache strategy and flow control strategy are examples, and the present invention does not make any specific restrictions on this.

[0032] In one of the preferred embodiments of the present invention, in order to find the CDN network node closest to the current CDN network node, after obtaining the current CDN network node routing link table, the central CDN network node determines the edge CDN network node closest to the current CDN network node according to the current CDN network node routing link table, and transmits the data of the adjacent node to the closest edge CDN network node through routing conversion according to the cache strategy, so as to realize efficient processing of cached data. The method for calculating the distance L between the edge CDN network node closest to the current CDN network node includes: obtaining the round-trip time RTT from the current CDN network node to any edge CDN network node, and calculating the number of hops Hops to the edge CDN network node, and calculating the bandwidth utilization P of the current CDN network node, then the calculation formula of the distance L between the edge CDN network node closest to the current CDN network node is: , where β, λ and γ are weight coefficients respectively.

[0033] The embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. The embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present application are executed. It should be noted that the computer-readable medium mentioned above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wire segments, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, electrical wire, optical cable, RF, etc., or any suitable combination of the foregoing.

[0034] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0035] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the principles, the implementation methods of the present invention may be deformed or modified in any way.

Claims

1. A distributed storage method based on AI traffic identification, characterized in that: The method comprises: Obtain current CDN network node status information, simulate and construct a first GNN node in the GNN model according to the current CDN network node status information, and construct adjacent nodes of the GNN model according to the communication relationship between the current CDN network node and different CDN network nodes; Acquire communication flow information between the first GNN node and the corresponding adjacent node, identify flow characteristic data in the communication flow information using an AI model, and construct a dynamic adjacency matrix between the first GNN node and the adjacent node according to the flow characteristic data; Using the dynamic adjacency matrix to update the status of the first GNN node and the adjacent nodes, and executing the caching strategy or traffic processing strategy of the corresponding CDN network node according to the updated status of the first GNN node and the adjacent nodes; Obtain the current CDN network node routing link table, and route the current CDN network node to the optimal CDN network node according to the cache strategy or traffic processing strategy.

2. According to the distributed storage method based on AI traffic identification described in claim 1, it is characterized in that: The current CDN network node status information includes hardware resource status, network resource status and cache status, wherein the hardware resource status includes: CPU usage, memory usage, cache disk read and write performance data; The network resource status includes: the upstream and downstream bandwidth utilization rate of the current CDN network node, the packet loss rate of the current CDN network node, the network request rate, the number of concurrent connections and the number of URLs; the cache status includes: cache hit rate, cache elimination rate, cache preheating status and cache content type; the hardware resource status, network resource status and cache status are respectively preprocessed to obtain the feature matrix of the first GNN node, and the dynamic adjacency matrix is ​​constructed according to the traffic feature data between the first GNN node and the corresponding adjacent node identified by the AI ​​model.

3. According to the distributed storage method based on AI traffic identification according to claim 2, it is characterized in that: The method for constructing the dynamic adjacency matrix includes: using an image recognition AI model, a text recognition AI model, a speech recognition AI model and a network attack recognition AI model to identify and obtain network traffic feature data between the first GNN node and the corresponding adjacent node; wherein the traffic feature data includes: illegal video frame data, illegal voice data, illegal text data, network attack type and quantity; pre-processing the different traffic feature data and converting them into the dynamic adjacency matrix, and updating the status of the first GNN node and the corresponding adjacent node according to the dynamic adjacency matrix.

4. According to the distributed storage method based on AI traffic identification according to claim 1, it is characterized in that: The state update method of the first GNN node and the corresponding adjacent node includes: constructing a multi-head attention coefficient according to the traffic feature data obtained by identifying the traffic with the AI ​​model , where i represents the current first GNN node, j represents the adjacent node that has a communication relationship with the first GNN node, and h i represents the first GNN node feature, a and W represent different learnable parameters, where W is the weight matrix, a is the action parameter between nodes based on the transformation of traffic feature data, T represents the transposed matrix identifier, LeakyReLU represents the activation function, Softmax represents the linear classification function, and the attention coefficient α ij Represents the similarity feature weights of the first GNN node and the corresponding adjacent node.

5. According to the distributed storage method based on AI traffic identification according to claim 4, it is characterized in that: The state updating method of the first GNN node and the corresponding adjacent node also includes: ij Parallel computation is performed to capture the relationship type between the first GNN node i and the adjacent node j, through the multi-head attention coefficient α ij The content of calculating the updated first GNN node includes: using the multi-head attention coefficient α ij Construct a dynamic adjacency matrix A dynamic [i, j] k = , where k represents the corresponding attention head, and the dynamic adjacency matrix A dynamic [i, j] k Replace the original static adjacency matrix A static After replacement, the first GNN node is updated according to the following formula: ,in represents the first GNN node state predicted after the update, N represents the node relationship function, K represents the total number of attention heads, σ is the activation function, and the first GNN node state predicted after the update is obtained. Execute the node cache strategy or traffic processing strategy corresponding to the current CDN network.

6. A distributed storage method based on AI traffic identification according to claim 5, characterized in that: The cache strategy or traffic processing strategy includes: according to the updated predicted first GNN node state Update the predicted cache state of the node in the current CDN network to allow the corresponding adjacent node data to be cached; according to the updated predicted first GNN node state The network attack data from the corresponding adjacent nodes predicted by the nodes of the current CDN network is limited or transferred to the nodes of a specific CDN network; according to the first GNN node state predicted after the update The node prediction of the current CDN network is processed with the violation data from the corresponding adjacent node, including flow limiting or access prohibition and caching strategy.

7. According to the distributed storage method based on AI traffic identification according to claim 1, it is characterized in that: CDN network nodes include central CDN network nodes and edge CDN network nodes. After obtaining the current CDN network node routing link table, the central CDN network node determines the edge CDN network node closest to the current CDN network node based on the current CDN network node routing link table, and transmits the data of the adjacent node to the closest edge CDN network node through routing conversion according to the cache strategy, thereby realizing efficient processing of cached data.

8. A distributed storage method based on AI traffic identification according to claim 7, characterized in that: The method for calculating the distance L between the edge CDN network node closest to the current CDN network node includes: obtaining the round-trip time RTT between the current CDN network node and any edge CDN network node, calculating the number of hops Hops to the edge CDN network node, and calculating the bandwidth utilization P of the current CDN network node. The calculation formula for the distance L between the edge CDN network node closest to the current CDN network node is: , where β, λ and γ are weight coefficients respectively.

9. A distributed storage system based on AI traffic identification, characterized in that: The system executes a distributed storage method based on AI traffic identification as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement a distributed storage method based on AI traffic identification as described in any one of claims 1-8.

Images