Data transfer method based on distributed asynchronous architecture

Abstract

The invention discloses a data transfer method based on a distributed asynchronous architecture, and the data transfer method based on the distributed asynchronous architecture comprises the following steps: S1, building a plurality of computer nodes; s2, when a source node transfers to a target node, monitoring data consistency by using a machine learning model, and detecting an abnormal mode and a potential data error; s3, fragmenting the data needing to be transferred, creating a copy in the source node, starting an asynchronous copying process, and copying the data from the source node to a target node; and S4, after the data transfer is completed, verifying the data in the target node, and ensuring the integrity and consistency of the data again. Through the asynchronous communication and monitoring system, the efficiency and reliability of data transfer are improved, the data consistency is monitored through the machine learning model, the exception detection and error correction capabilities in the data transfer process are enhanced, and the integrity and consistency of the data are ensured.

Description

Technical Field

[0001] The present invention relates to the field of data transfer, and in particular to a data transfer method based on a distributed asynchronous architecture. Background Art

[0002] In large-scale distributed systems, data transfer is a critical and complex task. The existing technology has the following problems in data transfer: Traditional data transfer methods usually rely on synchronous communication, which limits the speed and efficiency of data processing, especially in environments with high network latency or limited bandwidth. Synchronous communication may also lead to resource competition between source and target nodes, affecting the overall performance of the system, and lack an effective data consistency monitoring mechanism during data transfer. This may lead to inconsistencies in data during the transfer process, increase the risk of data errors and anomalies, and thus affect the integrity and reliability of the data. In addition, existing data sharding and handling tools are often inefficient, especially when processing large-scale data. These tools may not be able to adapt to the load capacity of different nodes, resulting in delays and bottlenecks in the data transfer process. Summary of the invention

[0003] The present invention aims to provide a data transfer method based on a distributed asynchronous architecture to solve the above problems.

[0004] The present invention is achieved through the following technical solutions: A data transfer method based on a distributed asynchronous architecture includes the following steps: S1: Establish multiple computer nodes, communicate asynchronously with each other through the network, clarify the configuration and interface between the source node and each node, and set up a monitoring system to track the progress and performance of data transfer; S2: Use machine learning models to monitor data consistency and detect abnormal patterns and potential data errors when transferring data from the source node to the target node; S3: Shard the data to be transferred and create a copy in the source node. Start the asynchronous replication process to copy the data from the source node to the target node. Configure the message queue to asynchronously process the transfer task. When the data is copied to a certain amount or completed, trigger an event to notify the target node to prepare to receive data. Partition the data according to the load capacity of the target node and use the data transfer tool to asynchronously transfer the data from the source node to the target node. S4: After the data transfer is completed, the data in the target node is verified to ensure the integrity and consistency of the data again. After the data verification passes, the source node sends a confirmation message to the target node to complete the transfer process.

[0005] The present invention improves the efficiency and reliability of data transfer through asynchronous communication and monitoring systems, monitors data consistency through machine learning models, enhances the anomaly detection and error correction capabilities during data transfer, and ensures data integrity and consistency.

[0006] Preferably, in S2, the following steps are specifically included: S2.1: Build an autoencoder model; The autoencoder includes an encoder and a decoder, wherein the encoder is used to compress input data into a low-dimensional representation, and the decoder is used to reconstruct the original data from the low-dimensional representation; S2.2: parameter setting; Input layer: Determine the dimension of the input layer based on the number of features of the data; Hidden layers: set to a hierarchy of 16, 8, 8, 16; Activation function: Use tanh function in the encoder and relu function in the decoder; Optimizer: Use the adam optimizer to adaptively adjust the learning rate; Loss function: Use mean square error function; S2.3: Model training; Use the adam optimizer, adaptively adjust the learning rate, and perform multiple iterations; S2.4: Anomaly detection. For new data, the trained model is used to reconstruct the data and the reconstruction error is calculated. If the reconstruction error exceeds the set threshold, the data is considered to be abnormal.

[0007] The present invention realizes advanced monitoring of data consistency and improves the anomaly detection capability during data transfer by establishing an automatic encoder model, wherein the optimization of parameter settings enables the model to better adapt to different data features, thereby improving the generalization capability and detection accuracy of the model.

[0008] Preferably, in S2.3, each iteration performs the following steps: S2.3.1: Use DataLoader to load data in batches; S2.3.2: Forward propagate the input data through the autoencoder to obtain the reconstructed output; S2.3.3: Use the MSE loss function to calculate the difference between the reconstructed output and the original input; S2.3.4: Calculate the gradient according to the loss function and perform backpropagation; S2.3.5: Update the model weights based on the gradients.

[0009] The present invention ensures that the model can learn more accurate data representation through multiple iterations through detailed model training steps, thereby improving the accuracy of anomaly detection. The detailed steps of each iteration ensure the systematicness and effectiveness of model training, so that the model can better adapt to new data and improve the security of data transfer.

[0010] Preferably, the expression of the tanh function is: in is the input value, is the characteristic value of the preprocessed data; It is the base of natural logarithms, a mathematical constant; The output range of the tanh function is (-1, 1).

[0011] The present invention provides a specific expression of the tanh function, so that the model can maintain the nonlinear characteristics of data during data compression and reconstruction, thereby improving the accuracy of data reconstruction.

[0012] Preferably, the expression of the relu function is: in is the input value, is the characteristic value of the preprocessed data; The output value of the relu function is non-negative. If the input value x is negative, the output is 0; if x is positive, the output is x itself.

[0013] The present invention provides a specific expression of the relu function, which makes the model more efficient when processing non-negative data, while maintaining the nonlinear characteristics of the data, which helps to improve the expressiveness of the model.

[0014] Preferably, the expression of the MSE loss function is: in, Represents the forward propagation process of the model; Represents the weights and biases in the encoder and decoder; is the amount of transferred data; It is The original input value of the data; It is The data reconstruction value is the output of the model; It is The square of the difference between the original input value and the predicted value of the sample; By minimizing this MSE loss function, the autoencoder model learns how to adjust its parameters , so that the predicted value As close to the actual value as possible , thereby improving the accuracy of data reconstruction.

[0015] The present invention provides a specific expression of the MSE loss function, so that the goal can be clearly defined during the model training process, that is, minimizing the difference between the predicted value and the actual value, thereby improving the accuracy of data reconstruction.

[0016] Preferably, in the input layer, before the original data is input into the autoencoder, features need to be extracted or constructed, which specifically include the following: For data with statistical attributes, the mean, median, maximum, minimum, standard deviation, etc. are extracted as features and input into the autoencoder; For data with business meaning, timestamp, data version number, checksum, etc. are extracted as features and input into the autoencoder; The present invention enables the model to better understand and process different types of data by extracting or constructing features, thereby improving the anomaly detection and error correction capabilities during data transfer.

[0017] Preferably, in S3, it specifically includes the following steps: S3.1: Divide the data into multiple small blocks through a hash function to obtain shard data, and store them on multiple target node storage devices; S3.2: Save copies of the data shards on multiple target nodes; S3.3: Set up a master node to receive client requests and synchronize data to multiple standby nodes. When the master node goes down, the standby node takes over the master node to ensure system availability. The specific steps are as follows: S3.3.1: When a client sends a request, the request is first sent to the master node; S3.3.2: The primary node executes the request and asynchronously copies the result to the standby node; S3.3.3: The standby node receives the synchronization data and updates its own copy of the data.

[0018] The present invention improves the availability and reliability of data by storing data in fragments and replicas, and can ensure the continuity and consistency of data even when the master node is down.

[0019] Preferably, during asynchronous execution, the data transfer task is started and executed asynchronously in the background to reduce the impact on the performance of the source node. In addition, the progress and performance of the data transfer are monitored to ensure that the data is correctly and completely transferred from the source node to the target node.

[0020] The present invention reduces the impact on the performance of the source node by asynchronously executing the data transfer task, and simultaneously monitors the progress and performance of the data transfer, thereby ensuring the high efficiency and accuracy of the data transfer.

[0021] Preferably, the hash function specifically includes the following: S3.1.1: Divide the input data into blocks of 512 bits in size. If the data length is not a multiple of 512, zero padding is added at the end. S3.1.2: Use a set of predefined initialization vectors and expand it into 8 independent 128-bit sub-blocks. These initial hash values ​​are taken from the fractional part of the square root of the first 8 prime numbers in natural numbers (2, 3, 5, 7, 11, 13, 17, 19), and take the first 32 bits; S3.1.3: For each 512-bit block in data, do the following: S3.1.3.1: Use a series of logic functions and operations to process each block. These operations include logical AND, OR, NOT, XOR, circular left shift, etc. S3.1.3.2: After each block is processed, add the result to the initial hash value to update the hash value; S3.1.4: After all blocks have been processed, the final hash value is the concatenation of eight 32-bit hash values ​​to form a 256-bit hash value.

[0022] Through specific hash function steps, the security and consistency of data sharding are improved, ensuring the integrity and reliability of data in distributed systems.

[0023] Beneficial effects: The present invention improves the efficiency and reliability of data transfer through asynchronous communication and monitoring system, and uses machine learning model to monitor data consistency, enhances the ability of anomaly detection and error correction, and ensures the integrity and consistency of data. Secondly, by establishing an automatic encoder model and optimizing parameter settings, the ability of anomaly detection in the process of data transfer is improved, so that the model can better adapt to different data features, and the generalization ability and detection accuracy of the model are improved. In addition, the detailed model training steps ensure that the model can learn more accurate data representation through multiple iterations, and improve the accuracy of anomaly detection. The specific function expressions provided, such as tanh and relu, enable the model to maintain the nonlinear characteristics of data during data compression and reconstruction, and improve the accuracy of data reconstruction. The introduction of MSE loss function enables the goal to be clear during the model training process, that is, to minimize the difference between the predicted value and the actual value, and further improves the accuracy of data reconstruction. The step of feature extraction or construction enables the model to better understand and process different types of data, and improves the ability of anomaly detection and error correction in the process of data transfer. The strategy of data sharding and copy preservation improves the availability and reliability of data, and ensures the continuity and consistency of data even when the master node is down. Finally, by asynchronously executing data transfer tasks and specific hash function steps, the impact on the performance of the source node is reduced. At the same time, the progress and performance of data transfer are monitored, ensuring the efficiency and accuracy of data transfer and improving the security and consistency of data sharding. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of the steps of the present invention. DETAILED DESCRIPTION

[0025] The technical solution of the present invention is further described in detail below, but the protection scope of the present invention is not limited to the following.

[0026] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments.

[0027] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention. It should be noted that relational terms such as the terms "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0028] Moreover, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that includes a list of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus. In the absence of more limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0029] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0030] Example according to Figure 1 As shown, a data transfer method based on a distributed asynchronous architecture includes the following steps: S1: Establish multiple computer nodes, communicate asynchronously with each other through the network, clarify the configuration and interface between the source node and each node, and set up a monitoring system to track the progress and performance of data transfer; In this embodiment, the monitoring system uses existing monitoring tools, such as Prometheus, Grafana, etc., to track the progress and performance of data transfer in real time to ensure the transparency and traceability of the transfer process.

[0031] S2: Use machine learning models to monitor data consistency and detect abnormal patterns and potential data errors when transferring data from the source node to the target node; In this embodiment, the model can be implemented based on the existing TensorFlow or PyTorch.

[0032] S3: Shard the data to be transferred and create a copy in the source node. Start the asynchronous replication process to copy the data from the source node to the target node. Configure the message queue to asynchronously process the transfer task. When the data is copied to a certain amount or completed, trigger an event to notify the target node to prepare to receive data. Partition the data according to the load capacity of the target node and use the data transfer tool to asynchronously transfer the data from the source node to the target node. S4: After the data transfer is completed, the data in the target node is verified to ensure the integrity and consistency of the data again. After the data verification passes, the source node sends a confirmation message to the target node to complete the transfer process.

[0033] The present invention improves the efficiency and reliability of data transfer through asynchronous communication and monitoring systems, monitors data consistency through machine learning models, enhances the anomaly detection and error correction capabilities during data transfer, and ensures data integrity and consistency.

[0034] In this embodiment, in S2, the following steps are specifically included: S2.1: Build an autoencoder model; The autoencoder includes an encoder and a decoder, wherein the encoder is used to compress input data into a low-dimensional representation, and the decoder is used to reconstruct the original data from the low-dimensional representation; S2.2: parameter setting; Input layer: Determine the dimension of the input layer based on the number of features of the data; Hidden layers: set to a hierarchy of 16, 8, 8, 16; Activation function: Use tanh function in the encoder and relu function in the decoder; Optimizer: Use the adam optimizer to adaptively adjust the learning rate; Loss function: Use mean square error function; S2.3: Model training; Use the adam optimizer, adaptively adjust the learning rate, and perform multiple iterations; S2.4: Anomaly detection. For new data, the trained model is used to reconstruct the data and the reconstruction error is calculated. If the reconstruction error exceeds the set threshold, the data is considered to be abnormal.

[0035] In this embodiment, input data preprocessing includes normalization, standardization, missing value processing, etc. to meet the input requirements of the model.

[0036] The present invention realizes advanced monitoring of data consistency and improves the anomaly detection capability during data transfer by establishing an automatic encoder model, wherein the optimization of parameter settings enables the model to better adapt to different data features, thereby improving the generalization capability and detection accuracy of the model.

[0037] In this embodiment, in S2.3, the following steps are performed in each iteration: S2.3.1: Use DataLoader to load data in batches; S2.3.2: Forward propagate the input data through the autoencoder to obtain the reconstructed output; S2.3.3: Use the MSE loss function to calculate the difference between the reconstructed output and the original input; S2.3.4: Calculate the gradient according to the loss function and perform backpropagation; S2.3.5: Update the model weights based on the gradients.

[0038] The present invention ensures that the model can learn more accurate data representation through multiple iterations through detailed model training steps, thereby improving the accuracy of anomaly detection. The detailed steps of each iteration ensure the systematicness and effectiveness of model training, so that the model can better adapt to new data and improve the security of data transfer.

[0039] In this embodiment, the expression of the tanh function is: in is the input value, is the characteristic value of the preprocessed data; It is the base of natural logarithms, a mathematical constant; The output range of the tanh function is (-1, 1).

[0040] The present invention provides a specific expression of the tanh function, so that the model can maintain the nonlinear characteristics of data during data compression and reconstruction, thereby improving the accuracy of data reconstruction.

[0041] In this embodiment, the expression of the relu function is: in is the input value, is the characteristic value of the preprocessed data; The output value of the relu function is non-negative. If the input value x is negative, the output is 0; if x is positive, the output is x itself.

[0042] The present invention provides a specific expression of the relu function, which makes the model more efficient when processing non-negative data, while maintaining the nonlinear characteristics of the data, which helps to improve the expressiveness of the model.

[0043] In this embodiment, the expression of the MSE loss function is: in, Represents the forward propagation process of the model; Represents the weights and biases in the encoder and decoder; is the amount of transferred data; It is The original input value of the data; It is The data reconstruction value is the output of the model; It is The square of the difference between the original input value and the predicted value of the sample; By minimizing this MSE loss function, the autoencoder model learns how to adjust its parameters , so that the predicted value As close to the actual value as possible , thereby improving the accuracy of data reconstruction.

[0044] The present invention provides a specific expression of the MSE loss function, so that the goal can be clearly defined during the model training process, that is, minimizing the difference between the predicted value and the actual value, thereby improving the accuracy of data reconstruction.

[0045] In this embodiment, in the input layer, before the original data is input into the autoencoder, features need to be extracted or constructed, which specifically include the following: For data with statistical attributes, the mean, median, maximum, minimum, standard deviation, etc. are extracted as features and input into the autoencoder; For data with business meaning, timestamp, data version number, checksum, etc. are extracted as features and input into the autoencoder; The present invention enables the model to better understand and process different types of data by extracting or constructing features, thereby improving the anomaly detection and error correction capabilities during data transfer.

[0046] In this embodiment, an abnormal pattern is defined as a data behavior that is significantly different from a normal data pattern, and a potential data error is defined as an error that may cause data inconsistency or loss.

[0047] In this embodiment, in S3, it specifically includes the following steps: S3.1: Divide the data into multiple small blocks through a hash function to obtain shard data, and store them on multiple target node storage devices; S3.2: Save copies of the data shards on multiple target nodes; S3.3: Set up a master node to receive client requests and synchronize data to multiple standby nodes. When the master node goes down, the standby node takes over the master node to ensure system availability. The specific steps are as follows: S3.3.1: When a client sends a request, the request is first sent to the master node; S3.3.2: The primary node executes the request and asynchronously copies the result to the standby node; S3.3.3: The standby node receives the synchronization data and updates its own copy of the data.

[0048] The present invention improves the availability and reliability of data by storing data in fragments and replicas, and can ensure the continuity and consistency of data even when the master node is down.

[0049] In this embodiment, data sharding can be performed based on the size and characteristics of the data using a distributed file system such as Hadoop HDFS. The transfer tool can be a custom-developed script or an existing data synchronization tool such as Rsync, SCP, etc. In this embodiment, during asynchronous execution, the data transfer task is started and executed asynchronously in the background to reduce the impact on the performance of the source node. In addition, the progress and performance of the data transfer are monitored to ensure that the data is correctly and completely transferred from the source node to the target node.

[0050] In this embodiment, the message queue can use popular message middleware such as RabbitMQ and Kafka, and manage the transfer task by configuring queues, switches and bindings. The present invention reduces the impact on the performance of the source node by asynchronously executing the data transfer task, and simultaneously monitors the progress and performance of the data transfer, thereby ensuring the high efficiency and accuracy of the data transfer.

[0051] In this embodiment, the hash function specifically includes the following: S3.1.1: Divide the input data into blocks of 512 bits in size. If the data length is not a multiple of 512, zero padding is added at the end. S3.1.2: Use a set of predefined initialization vectors and expand it into 8 independent 128-bit sub-blocks. These initial hash values ​​are taken from the fractional part of the square root of the first 8 prime numbers in natural numbers (2, 3, 5, 7, 11, 13, 17, 19), and take the first 32 bits; S3.1.3: For each 512-bit block in data, do the following: S3.1.3.1: Use a series of logic functions and operations to process each block. These operations include logical AND, OR, NOT, XOR, circular left shift, etc. S3.1.3.2: After each block is processed, add the result to the initial hash value to update the hash value; S3.1.4: After all blocks have been processed, the final hash value is the concatenation of eight 32-bit hash values ​​to form a 256-bit hash value.

[0052] Through specific hash function steps, the security and consistency of data sharding are improved, ensuring the integrity and reliability of data in distributed systems.

[0053] In this embodiment, a heartbeat detection, fault detection and automatic switching mechanism is adopted to ensure that when the master node fails, the backup node can quickly take over the task, thereby ensuring the high availability of the system.

[0054] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A data transfer method based on a distributed asynchronous architecture, characterized in that: The steps include: S1: Establish multiple computer nodes, communicate asynchronously with each other through the network, clarify the configuration and interface between the source node and each node, and set up a monitoring system to track the progress and performance of data transfer; S2: Use machine learning models to monitor data consistency and detect abnormal patterns and potential data errors when transferring data from the source node to the target node; S3: Shard the data to be transferred and create a copy in the source node. Start the asynchronous replication process to copy the data from the source node to the target node. Configure the message queue to asynchronously process the transfer task. When the data is copied to a certain amount or completed, trigger an event to notify the target node to prepare to receive the data. Partition the data according to the load capacity of the target node, and use data transfer tools to asynchronously transfer data from the source node to the target node; S4: After the data transfer is completed, the data in the target node is verified to ensure the integrity and consistency of the data again. After the data verification passes, the source node sends a confirmation message to the target node to complete the transfer process.

2. The data transfer method based on a distributed asynchronous architecture according to claim 1, characterized in that: In S2, the following steps are specifically included: S2.1: Build an autoencoder model; The autoencoder includes an encoder and a decoder, wherein the encoder is used to compress input data into a low-dimensional representation, and the decoder is used to reconstruct the original data from the low-dimensional representation; S2.2: parameter setting; Input layer: Determine the dimension of the input layer based on the number of features of the data; Hidden layers: set to a hierarchy of 16, 8, 8, 16; Activation function: Use tanh function in the encoder and relu function in the decoder; Optimizer: Use the adam optimizer to adaptively adjust the learning rate; Loss function: Use mean square error function; S2.3: Model training; Use the adam optimizer, adaptively adjust the learning rate, and perform multiple iterations; S2.4: Anomaly detection. For new data, the trained model is used to reconstruct the data and the reconstruction error is calculated. If the reconstruction error exceeds the set threshold, the data is considered to be abnormal.

3. The data transfer method based on a distributed asynchronous architecture according to claim 2, characterized in that: In S2.3, each iteration performs the following steps: S2.3.1: Use DataLoader to load data in batches; S2.3.2: Forward propagate the input data through the autoencoder to obtain the reconstructed output; S2.3.3: Use the MSE loss function to calculate the difference between the reconstructed output and the original input; S2.3.4: Calculate the gradient according to the loss function and perform backpropagation; S2.3.5: Update the model weights based on the gradients.

4. The data transfer method based on a distributed asynchronous architecture according to claim 3 is characterized in that: The expression of the tanh function is: in is the input value, is the characteristic value of the preprocessed data; It is the base of natural logarithms, a mathematical constant; The output range of the tanh function is (-1, 1).

5. The data transfer method based on a distributed asynchronous architecture according to claim 3 is characterized in that: The expression of the relu function is: in is the input value, is the characteristic value of the preprocessed data; The output value of the relu function is non-negative. If the input value x is negative, the output is 0; if x is positive, the output is x itself.

6. The data transfer method based on a distributed asynchronous architecture according to claim 3 is characterized in that: The expression of the MSE loss function is: in, Represents the forward propagation process of the model; Represents the weights and biases in the encoder and decoder; is the amount of transferred data; It is The original input value of the data; It is The data reconstruction value is the output of the model; It is The square of the difference between the original input value and the predicted value of the sample; By minimizing this MSE loss function, the autoencoder model learns how to adjust its parameters , so that the predicted value As close to the actual value as possible , thereby improving the accuracy of data reconstruction.

7. The data transfer method based on a distributed asynchronous architecture according to claim 3 is characterized in that: In the input layer, before the raw data is input into the autoencoder, features need to be extracted or constructed, which include the following: For data with statistical attributes, the mean, median, maximum, minimum, and standard deviation are extracted as features and input into the autoencoder; For data with business meaning, timestamp, data version number, and checksum are extracted as features and input into the autoencoder.

8. The data transfer method based on a distributed asynchronous architecture according to claim 1, characterized in that: In S3, it specifically includes the following steps: S3.1: Divide the data into multiple small blocks through a hash function to obtain shard data, and store them on multiple target node storage devices; S3.2: Save copies of the data shards on multiple target nodes; S3.3: Set up a master node to receive client requests and synchronize data to multiple standby nodes. When the master node goes down, the standby node takes over the master node to ensure system availability. The specific steps are as follows: S3.3.1: When a client sends a request, the request is first sent to the master node; S3.3.2: The primary node executes the request and asynchronously copies the result to the standby node; S3.3.3: The standby node receives the synchronization data and updates its own copy of the data.

9. The data transfer method based on a distributed asynchronous architecture according to claim 8, characterized in that: In asynchronous execution, the data transfer task is started and executed asynchronously in the background to reduce the impact on the performance of the source node. In addition, the progress and performance of the data transfer are monitored to ensure that the data is correctly and completely transferred from the source node to the target node.

10. The data transfer method based on a distributed asynchronous architecture according to claim 9, characterized in that: The hash functions are as follows: S3.1.1: Divide the input data into blocks of 512 bits in size. If the data length is not a multiple of 512, zero padding is added at the end. S3.1.2: Use a set of predefined initialization vectors and expand it into 8 independent 128-bit sub-blocks. These initial hash values ​​are taken from the fractional part of the square root of the first 8 prime numbers in natural numbers, and the first 32 bits are taken; S3.1.3: For each 512-bit block in data, do the following: S3.1.3.1: Process each block using a series of logic functions and operations, including logical AND, OR, NOT, XOR, and rotate left; S3.1.3.2: After each block is processed, add the result to the initial hash value to update the hash value; S3.1.4: After all blocks have been processed, the final hash value is the concatenation of eight 32-bit hash values ​​to form a 256-bit hash value.

Images