A secure batch configuration method and system for coffee machine clusters using edge gateway proxy

By constructing a topology diagram and analyzing time-series data of the coffee machine cluster through an edge gateway, and combining authentication and cross-attention mechanisms, the secure and stable operation and maintenance of the coffee machine cluster was achieved. This solved the problem of the separation between device authentication and control, and improved the security and reliability of operation and maintenance.

CN122845457BActive Publication Date: 2026-11-10JIANGSU HAOHAN INFORMATION TECH +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611350172.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-10
Estimated Expiration
2046-09-02

AI Technical Summary

Technical Problem

The existing operation and maintenance methods for coffee machine clusters have significant security risks, as the equipment communication links are vulnerable to forged command attacks. They cannot push updates based on the real-time operating conditions of the equipment, the update process lacks fault tolerance, and local equipment failures can easily cause the entire batch of equipment to shut down, resulting in low overall operation and maintenance control accuracy.

Method used

An explicit topology diagram of the coffee machine cluster is constructed through an edge gateway. Time-series data is collected to generate an environmental health vector. Random number challenge codes are used to verify device identity, a secure session channel is established, and a grayscale sequence is generated by combining a cross-attention mechanism. Grayscale test groups and follower groups are divided, and differentiated configuration packages are configured. In case of an anomaly, a global circuit breaker and configuration rollback are triggered.

Benefits of technology

It has achieved safe and stable operation and maintenance of coffee machine clusters, improved the security, consistency and reliability of configuration operations, enhanced the self-healing capability of faults, and ensured the linkage verification and hierarchical distribution of equipment safety status.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122845457B_ABST
    Figure CN122845457B_ABST
Patent Text Reader

Abstract

This invention discloses a secure batch configuration method and system for coffee machine clusters using an edge gateway proxy, relating to the field of edge data processing technology. The method includes: the edge gateway constructing an explicit topology diagram and environmental health vector of the coffee machine cluster; establishing a secure session channel after authentication; generating a grayscale sequence through a cross-attention mechanism and grouping it into a grayscale test group and a follower group; distributing configuration packages for pre-check execution to the grayscale test group; and triggering a global circuit breaker and configuration rollback in case of anomalies. This invention solves the technical problems of fragmented device authentication and control in batch configuration operations of coffee machine clusters, which easily introduce risks of unauthorized access and configuration errors, and lack of differentiated management and anomaly recovery mechanisms for the dynamic state of the cluster. It achieves linked verification of configuration operations and device security status, and performs hierarchical distribution and anomaly circuit breaker rollback based on the real-time status of the coffee machine cluster, thereby improving the security, consistency, and reliability of the configuration process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of edge data processing technology, and in particular to a secure batch configuration method and system for coffee machine clusters using an edge gateway proxy. Background Technology

[0002] Unmanned self-service coffee machines are mostly deployed in various offline stores, with multiple machines forming clusters for operation. Routine parameter adjustments and remote firmware updates are routine maintenance tasks. Communication security and the overall reliability of cluster maintenance directly affect the normal operation of stores. Currently, the mainstream cluster maintenance model in the market uses terminal devices to directly connect to the cloud server to complete command interaction, and device identity verification relies on fixed keys for protection, resulting in a single overall protection method. At the same time, a large amount of operational data generated by the devices is only used for basic fault alarms and does not participate in the scheduling decision-making of task issuance. Configuration updates generally adopt a full cluster synchronous push method. The existing maintenance method has many shortcomings: the device communication link is vulnerable to forged command attacks, resulting in significant security risks; it cannot push update content differently based on the real-time operating conditions of each device, the update process lacks fault tolerance, and local device anomalies can easily cause the entire batch of devices to shut down; the overall maintenance control precision is low, and it cannot adapt to the stable maintenance requirements of large-scale commercial coffee machine clusters. Summary of the Invention

[0003] This application provides a secure batch configuration method and system for coffee machine clusters using an edge gateway proxy, which solves the technical problems of the separation of device authentication and control in the batch configuration operation of coffee machine clusters, which easily introduces the risks of unauthorized access and configuration errors, and lacks differentiated management and abnormal recovery mechanisms for the dynamic state of the cluster.

[0004] The first aspect of this application provides a secure batch configuration method for a coffee machine cluster proxied by an edge gateway. The method includes: the edge gateway constructing an explicit topology diagram of multiple coffee machines in the coffee machine cluster and collecting time-series data of each coffee machine during operation to generate an environmental health vector; the edge gateway verifying the response codes returned by the multiple coffee machines according to a random number challenge code, and establishing a secure session channel when the verification is successful and the corresponding communication response delay meets a preset deviation; the edge gateway using a cross-attention mechanism to match the batch configuration task, the explicit topology diagram, and the environmental health vector to generate a grayscale sequence, and dividing the grayscale sequence into a grayscale test group and a follower group; the edge gateway sending differentiated configuration packages to the grayscale test group, and issuing an execution instruction after receiving a pre-check ready signal, triggering a global circuit breaker and performing configuration rollback using local old parameter baselines when the execution state is abnormal.

[0005] A second aspect of this application provides a secure batch configuration system for a coffee machine cluster proxies by an edge gateway. The system includes: an environment health vector generation module, used by the edge gateway to construct an explicit topology diagram of multiple coffee machines in the cluster and collect time-series data of each coffee machine during operation to generate an environment health vector; a secure session channel construction module, used by the edge gateway to verify the response codes returned by multiple coffee machines according to random number challenge codes, and to establish a secure session channel when verification is successful and the corresponding communication response delay meets a preset deviation; a grayscale sequence generation module, used by the edge gateway to use a cross-attention mechanism to match the batch configuration task, the explicit topology diagram, and the environment health vector to generate a grayscale sequence, and to divide the grayscale sequence into a grayscale test group and a follower group; and a signal receiving and execution module, used by the edge gateway to send differentiated configuration packages to the grayscale test group, and to issue execution instructions after receiving a pre-check ready signal, triggering a global circuit breaker and performing configuration rollback using local old parameter baselines when the execution state is abnormal.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application establishes a coffee machine cluster topology through an edge gateway and collects device runtime sequence data. After frequency domain feature extraction and feature mapping, device operating status data is obtained. A secure communication channel is established by combining dynamic identity verification. Based on multi-dimensional feature matching, the system automatically divides and updates machine groups in batches, and performs front-end device self-checks. In case of anomalies, the system immediately blocks the distribution and restores historical configurations, thus securely completing the batch distribution of differentiated parameters for multiple coffee machines. This makes the remote operation and maintenance process of the distributed coffee machine cluster safer and more stable, with stronger fault self-healing capabilities. It achieves linkage verification between configuration operations and device security status, and performs hierarchical distribution and anomaly circuit breaker rollback based on the real-time status of the coffee machine cluster, improving the security, consistency, and reliability of the configuration process. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating a secure batch configuration method for a coffee machine cluster using an edge gateway proxy, as provided in an embodiment of this application.

[0009] Figure 2 This is a schematic diagram of the frequency domain amplitude distribution of four operating parameters of a secure batch configuration method for a coffee machine cluster using an edge gateway proxy, provided in an embodiment of this application.

[0010] Figure 3 This is a schematic diagram of the structure of a secure batch configuration system for a coffee machine cluster using an edge gateway proxy, as provided in an embodiment of this application.

[0011] Figure labeling: Module 1 for generating environmental health vector, Module 2 for constructing secure session channel, Module 3 for generating grayscale sequence, and Module 4 for receiving and executing signal. Detailed Implementation

[0012] This application provides a secure batch configuration method and system for coffee machine clusters using an edge gateway proxy, which solves the technical problems of the separation of device authentication and control in the batch configuration operation of coffee machine clusters, which easily introduces the risks of unauthorized access and configuration errors, and lacks differentiated management and abnormal recovery mechanisms for the dynamic state of the cluster.

[0013] 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 a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., used in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, a secure batch configuration method for a coffee machine cluster proxies via an edge gateway includes: The edge gateway constructs an explicit topology diagram of multiple coffee machines in the coffee machine cluster and collects time-series data of each coffee machine during operation to generate an environmental health vector.

[0016] In this embodiment, the edge gateway is an embedded network hardware device deployed locally in the store, connecting to the cloud and multiple self-service coffee machines, responsible for local data processing, device management, and secure communication forwarding. The coffee machine cluster is a unified system of multiple commercial self-service coffee machines within the same store or operating area, all subject to scheduling and maintenance by the edge gateway. The explicit topology diagram is a digital device network relationship data model stored by the edge gateway, clearly recording the attributes of each coffee machine node, the gateway root node, and the communication links between devices.

[0017] Specifically, the edge gateway first collects the device number, physical deployment location, store identifier, and communication address information of each coffee machine in the cluster through the device registration interface, while also including itself as the root node of the topology. For example, the edge gateway sends ICMP link probe messages to each coffee machine at a fixed 5-second interval, collecting the online status of each coffee machine, the communication latency and packet loss rate data between the gateway and the device, and recording the point-to-point connectivity between the coffee machines. The edge gateway uses an adjacency matrix as the computer storage form of the topology structure. Each row and column of the matrix corresponds one-to-one with all topology nodes, including the edge gateway and each coffee machine. The elements within the matrix store the link attribute data between the corresponding row and column nodes. Each node in the adjacency matrix has a store identifier attribute field. Nodes are organized using a block storage method, with coffee machine nodes within the same store corresponding to independent sub-matrix blocks in the adjacency matrix, achieving logical grouping of store partitions.

[0018] The edge gateway uses both the edge gateway and a single coffee machine as topology nodes, and the control link from the gateway to the device as directed topology edges, with the direction of each edge uniformly pointing from the gateway to the terminal device. The edge gateway maps basic device attributes and online status to node attributes, and communication latency and packet loss rate data to edge attributes, constructing a star-shaped directed explicit topology graph with attributes. After the topology graph is constructed, the edge gateway synchronously updates the node and edge attributes with each link probe result, ensuring that the topology structure is synchronized with the actual operating status of the cluster in real time.

[0019] Next, the time-series data of the coffee machine operation is subjected to discrete Fourier transform to extract high-frequency harmonic component features, and then a lightweight multilayer perceptron is used to complete feature mapping to obtain an environmental health vector that can characterize the operating conditions of the equipment. The specific implementation process of this step will be described in detail later.

[0020] The edge gateway verifies the response codes returned by multiple coffee machines based on random number challenge codes, and establishes a secure session channel when the verification is successful and the corresponding communication response delay meets the preset deviation.

[0021] In this embodiment, the secure session channel is a dedicated communication transmission link established between the edge gateway and a single coffee machine, which uses encrypted data transmission and is used only for exchanging batch configuration-related instructions.

[0022] Optionally, the edge gateway pre-stores a unique device key for each coffee machine locally. This device key is synchronously written into the coffee machine's local storage area and the edge gateway's local database during the device's manufacturing process. The key uses a symmetric encryption algorithm for encryption and decryption. For example, the industry-standard AES128 encryption algorithm is used to ensure lightweight encryption operations are compatible with the edge gateway's limited computing power. Before initiating a secure session with any coffee machine, the edge gateway calls its local random number generation module to generate a random number challenge code. Preferably, a 128-bit random number challenge code is generated. The random number generation module uses a cryptographically secure pseudo-random number generator to generate the data, avoiding the replay attack risk associated with predictable random codes. The edge gateway sends the generated random number challenge code to the corresponding coffee machine through a unified control interface. After receiving the random number challenge code, the coffee machine retrieves its locally stored unique device key and uses the same AES128 encryption algorithm as the edge gateway to encrypt the random number challenge code, generating an encrypted response code. The response code is then sent back to the edge gateway.

[0023] After receiving the response code from the coffee machine, the edge gateway retrieves the unique device key matching that coffee machine from its database. It then decrypts the response code using the same AES128 decryption algorithm and compares the decrypted data bit-by-bit with the random challenge code initially issued by the edge gateway. If the two sets of data match perfectly, the coffee machine is deemed to have passed the identity verification between the random challenge code and the response code. If any bit is inconsistent, the verification is immediately deemed to have failed, and subsequent latency deviation comparison steps are not performed.

[0024] Next, after completing the identity verification, the edge gateway extracts the historical average response latency of the corresponding coffee machine from the environmental health vector, calculates the absolute difference between the current communication response latency and the average, and if the absolute difference is less than the preset millisecond-level security threshold, the verification is deemed successful and a secure session channel is established. All subsequent batch configuration-related commands are transmitted through this secure session channel. The specific implementation process of this step will be described in detail later.

[0025] By employing a dual verification method—combining cryptographic random challenge-response encryption verification with secondary verification of communication delay deviation—the identity of the coffee machine device and the security of the communication link are determined, establishing a trusted communication transmission channel for cluster batch configuration operations.

[0026] The edge gateway uses a cross-attention mechanism to match batch configuration tasks, the explicit topology diagram, and the environmental health vector to generate a grayscale sequence, and divides the grayscale test group and the follower group according to the grayscale sequence.

[0027] In this embodiment, the cross-attention mechanism is a vector operation logic that associates and matches batch configuration task characteristics with device topology and health characteristics, automatically calculating device configuration adaptation weights. The grayscale sequence is an ordered list of all compliant coffee machines sorted from highest to lowest grayscale fitness score. The grayscale test group consists of a small number of highly adaptable coffee machines selected from the grayscale sequence, used for trial operation of the new configuration package. The follow-up group consists of the remaining coffee machines in the cluster that have passed verification but have not been included in the grayscale test group; they will be configured after the grayscale test proceeds without any anomalies.

[0028] In one embodiment of this application, firstly, the edge gateway sets the task feature attributes of the batch configuration task as the query vector, concatenates the explicit topology structure diagram with the environmental health vector as the key vector and value vector, calculates the inner product of the query vector and the key vector and obtains the weight allocation through Softmax normalization, and obtains the matching feature vector by weighted summation of the value vector according to the weight allocation. Then, the matching feature vector is processed by the linear mapping layer to output the gray-level fitness score. Finally, the gray-level sequence is generated by sorting the gray-level fitness scores from high to low. The specific implementation process of this step will be described in detail later.

[0029] Next, the edge gateway filters coffee machines whose gray-scale fitness scores are greater than or equal to the preset admission scores within the gray-scale sequence. It then counts the business traffic data corresponding to the filtered coffee machines and removes coffee machines that are in peak business hours. The preset number of coffee machines ranked at the top of the remaining coffee machines are selected as the gray-scale test group, and the remaining coffee machines that have completed the verification are uniformly divided into the follower group. The specific implementation process of this step will also be described in detail later.

[0030] The edge gateway sends a differentiated configuration package to the grayscale test group and issues an execution command after receiving the pre-test ready signal. When the execution status is abnormal, a global circuit breaker is triggered and the configuration is rolled back using the local old parameter baseline.

[0031] In this embodiment, the differentiated configuration package is a unique configuration data package that is individually packaged and generated based on the coffee machine model, hardware version, and current operating parameters.

[0032] Specifically, the edge gateway first sends a differentiated configuration package to the gray-scale test group. Each coffee machine completes parameter range pre-check and hardware aging load capacity assessment. After all passes, the configuration register is locked and a pre-check ready signal is sent back. After receiving all ready signals within the preset waiting time window, the edge gateway sends an execution command. The specific implementation process of this step will be explained in detail later.

[0033] When an exception occurs during configuration execution, a global circuit breaker is triggered, the issuance of follow group instructions is stopped and the original state of the follow group is locked. The edge gateway issues a rollback instruction to the gray-scale test group through a secure session channel, and uses the local old parameter baseline to cover the differentiated configuration package to restore the historically available configuration version. The specific implementation process of this step will also be described in detail later.

[0034] Furthermore, the method provided in this application embodiment includes: Discrete Fourier transform is performed on the time-series data continuously collected from each coffee machine to extract high-frequency harmonic component features; the high-frequency harmonic component features are input into a lightweight multilayer perceptron for feature mapping to obtain an environmental health vector, wherein the environmental health vector is used to reflect the operating status of the coffee machine.

[0035] In this embodiment, the lightweight multilayer perceptron is a simplified three-layer fully connected neural network program deployed locally on the edge gateway, used to convert the frequency domain operating characteristics of the coffee machine into an environmental health vector that can characterize the operating condition of the equipment.

[0036] Specifically, after the explicit topology diagram is constructed, the edge gateway, through the device's local data acquisition channel, exemplarily, continuously collects the runtime timing data of each coffee machine at a sampling frequency of 20Hz. The timing data includes four operating parameters: extraction pressure, grinder motor speed, heating module power, and water flow rate. Preferably, after collecting 128 consecutive sampling points of timing data, the edge gateway preprocesses the single set of timing data, removing the DC component, and then performs a Discrete Fourier Transform on the timing sequences of the four operating parameters, converting the time-domain operating sequences of each parameter into corresponding frequency-domain signals to obtain the amplitude data corresponding to different frequency components of each parameter. As an optional implementation, the edge gateway pre-sets 6Hz as a high-frequency threshold, filters out frequency bands higher than 6Hz and lower than 10Hz as the high-frequency analysis interval, and extracts three uniformly distributed frequency point amplitudes for the extraction pressure parameter, two uniformly distributed frequency point amplitudes for the grinder motor speed parameter, three uniformly distributed frequency point amplitudes for the heating module power parameter, and two uniformly distributed frequency point amplitudes for the water flow rate parameter. The amplitude data obtained from the extraction of the four types of parameters are sequentially concatenated according to the extraction pressure parameter, grinding motor speed parameter, heating module power parameter, and effluent flow rate parameter to form a high-frequency harmonic component feature with a dimension of 10, thus completing the frequency domain feature extraction of a single set of time-series data. For example, as shown... Figure 2 As shown, after a single set of time-series data undergoes discrete Fourier transform, the amplitude distribution curves of the four operating parameters in the frequency domain are displayed. The horizontal axis represents frequency (Hz), and the vertical axis represents the normalized value of the harmonic amplitude. The figure also marks the high-frequency analysis range from 6Hz to 10Hz, as well as the specific locations of the amplitudes of a total of 10 frequency points extracted from each parameter within this range.

[0037] The lightweight multilayer perceptron employs a three-layer fully connected feedforward network structure: an input layer, a hidden layer, and an output layer. Data transfer between layers is fully connected, and each fully connected layer includes a bias term. The output value of each layer is equal to the product of the input weights of the previous layer plus the corresponding layer's bias value. The input layer has 10 neurons, each corresponding to a dimension of the high-frequency harmonic component features, used to receive the input feature data. The hidden layer has 16 neurons, using a linear rectified function as the activation function for nonlinear transformation. The weights are initialized using a Xavier uniform method, and the bias term is initialized to 0. The output layer has 8 neurons, using a linear activation function to output the original calculated value. The original calculated value is mapped to the 0-1 interval through min-max normalization. The min-max calculation method is that the output value for a single dimension equals the original value minus the minimum value of the training samples in that dimension, divided by the difference between the maximum and minimum values ​​of the training samples in that dimension. The eight output dimensions correspond to eight operational health dimensions: extraction pressure stability, grinding motor stability, heating module operating condition, water circuit unobstructedness, valve body sealing performance, powder hopper compatibility status, temperature control accuracy, and overall machine energy consumption level.

[0038] The lightweight multilayer perceptron is pre-trained offline before deployment to the edge gateway. Prior to training, runtime sequence data from multiple coffee machines in different stores with varying usage durations is collected. High-frequency harmonic component features are obtained through the same frequency domain feature extraction steps as described above and used as training samples. Maintenance personnel, combining equipment maintenance records, fault history, and performance test results, classify the health status of each device into four levels: excellent, good, average, and warning. Then, according to a preset level-vector mapping rule, the health status level is converted into an 8-dimensional health vector as training labels. The specific level-vector mapping rule is shown in Table 1.

[0039] Table 1: Health Status Level - Vector Mapping Rule Table

[0040] The training process divides the samples and labels into training and validation sets in an exemplary 8:2 ratio. Mean squared error is used as the loss function. Preferably, a stochastic gradient descent algorithm with a learning rate of 0.001 is used to iteratively update the network weights, with a batch size of 32 for each iteration. After each iteration using all training samples, a validation loss value is calculated using the validation set samples. As an optional implementation, when the decrease in the validation loss value is less than a preset convergence threshold of 0.001 for five consecutive iterations, the model training is considered complete, and the trained model weight file is deployed to the edge gateway for local execution.

[0041] After the edge gateway extracts the high-frequency harmonic component features of a single coffee machine, it inputs a 10-dimensional feature vector into a locally deployed lightweight multilayer perceptron. The feature data, after being passed through the input layer, undergoes nonlinear feature mapping through the hidden layer, and finally outputs an 8-dimensional environmental health vector from the output layer. The values ​​of each dimension within the environmental health vector correspond to the health status of the coffee machine across eight operational dimensions; values ​​closer to 1 indicate better operational status for the corresponding dimension, while values ​​closer to 0 indicate a higher risk of failure for that dimension.

[0042] By using a frequency domain feature extraction method with parameter matching and a lightweight neural network with a well-defined structure training process, we can achieve clear control of the global topology of the coffee machine cluster and quantitative output of the health status of individual devices, providing stable and reliable data support for subsequent batch configuration and scheduling.

[0043] Furthermore, the method provided in this application embodiment includes: Extract the historical average response latency from the environmental health vector; calculate the absolute difference between the current communication response latency and the historical average response latency; when the absolute difference is less than a preset millisecond-level security threshold, the verification is passed and the secure session channel is established.

[0044] Specifically, after the edge gateway completes the identity verification operation of the random challenge code and response code, it retrieves the locally stored environment health vector. The environment health vector pre-stores response latency data generated by each communication interaction of a single coffee machine, that is, the millisecond-level time record of the time consumed from the edge gateway sending the random challenge code to the coffee machine to receiving the corresponding response code. Subsequently, the edge gateway uses the arithmetic mean calculation method to extract all historical communication response latency data, add them together, and divide by the total number of data entries to obtain the historical average response latency for that coffee machine.

[0045] The edge gateway records the system timestamp of the random number challenge code issued this time, and at the same time records the system timestamp of the response code received from the corresponding coffee machine. The difference between the two timestamps is calculated to obtain the duration of the complete interaction. This duration is the current communication response latency, and the unit of time calculation is uniformly set to milliseconds.

[0046] Next, the edge gateway uses an absolute value calculation method, subtracting the historical average response latency from the current communication response latency, and taking the absolute value of the result to obtain the absolute difference between the current communication response latency and the historical average response latency. The edge gateway locally stores a preset millisecond-level security threshold. For example, the preset millisecond-level security threshold ranges from 20 milliseconds to 50 milliseconds, adapting to stable communication scenarios in store LANs. Specific threshold values ​​can be customized by those skilled in the art through configuration commands issued by the cloud-based operations and maintenance platform according to actual needs.

[0047] Finally, the edge gateway performs a size comparison operation between the absolute difference and the preset millisecond-level security threshold. If the absolute difference is less than the preset millisecond-level security threshold, it is determined that the coffee machine has passed the identity verification and the communication link verification has also passed. The edge gateway immediately opens a dedicated data transmission link and establishes a secure session channel between the edge gateway and the coffee machine. All subsequent batch configuration-related control commands and parameter data packets can only be transmitted through the secure session channel.

[0048] By extracting historical latency averages, calculating the difference between real-time latency and the average, and comparing the difference with security thresholds, a link verification process is added after identity verification to determine the stability of the communication link, ensuring that there are no security risks such as hijacking or lag in the communication link for establishing a secure session channel.

[0049] Furthermore, the method provided in this application embodiment includes: When a coffee machine fails verification, it is automatically moved to the edge isolation re-authentication queue. After the gray-scale test group or the follower group completes the configuration task, the edge gateway reduces the authentication initiation cycle of the edge isolation re-authentication queue and re-initiates the verification.

[0050] In this embodiment, the edge isolation re-authentication queue is a first-in-first-out data storage list stored locally by the edge gateway. It is used to uniformly store information of coffee machine devices that have failed security verification, perform re-authentication verification cyclically, and count the number of consecutive device failures.

[0051] Optionally, after the edge gateway completes the entire verification process for the random challenge code and response code of a single coffee machine, if the randomly obtained data from the decryption comparison does not match, or if the absolute difference in communication response latency is greater than or equal to a preset millisecond-level security threshold, the coffee machine is determined to have failed the overall verification. The edge gateway immediately stops establishing a secure session channel with the coffee machine and simultaneously retrieves the locally maintained edge isolation re-authentication queue. It then writes the corresponding device number, physical deployment path, and timestamp of the verification failure for this coffee machine into the edge isolation re-authentication queue to complete the move-in operation. The edge isolation re-authentication queue adopts a first-in, first-out data storage structure. Each storage entry in the queue is bound to a unique device identifier for the corresponding coffee machine. For example, the edge gateway defaults to setting the initial authentication initiation cycle to 60 seconds. It retrieves device entries from the queue sequentially according to this cycle and repeats the entire verification process of random challenge code issuance and response code verification.

[0052] The edge gateway continuously monitors the execution progress of the cluster batch configuration tasks and obtains real-time feedback signals on the completion of configuration tasks for all coffee machines in the gray-scale test group and all coffee machines in the follow group. When all coffee machines in the gray-scale test group or the follow group complete the execution of the differentiated configuration package, the edge gateway triggers the queue period adjustment logic. For example, the authentication initiation period of the edge isolation re-authentication queue is reduced from 60 seconds to 20 seconds, shortening the retry verification interval of the devices in the queue, and then the device entries in the queue are retrieved again to complete the full set of verification processes of random number challenge code and response code.

[0053] By using a coordinated scheduling mechanism that automatically queues and stores devices that fail verification and shortens the re-authentication cycle after batch configuration tasks are completed, the system continuously performs identity verification on coffee machines that fail security checks, thereby improving the efficiency of restoring faulty devices to the cluster management link.

[0054] Furthermore, the method provided in this application embodiment includes: The cumulative number of failures for a single coffee machine in the edge isolation re-authentication queue is counted. When the cumulative number of failures is greater than or equal to a preset blocking threshold, the edge gateway blocks communication with the control interface of the coffee machine and reports an alarm log containing the physical path of the coffee machine to the cloud.

[0055] Specifically, each time the edge gateway retrieves a coffee machine from the edge isolation re-authentication queue to perform the random number challenge code and response code verification process, it simultaneously reads the cumulative failure count field bound to that coffee machine. If the current verification process determines that the coffee machine still fails verification, the cumulative failure count value corresponding to that coffee machine is incremented by 1 and updated in real time to the storage entry of that device in the edge isolation re-authentication queue. The preset blocking threshold value is obtained completely through the offline device security testing process. During the offline testing phase, multiple coffee machines with different service years and different store scenarios are selected, and continuous random number challenge code verification tests are conducted. The critical number of times forged instructions or illegal hijacking of communication interfaces occurs after continuous verification failures is recorded. The average value of the critical number values ​​obtained from multiple tests is calculated, and two redundant verification counts are added as a security buffer. For example, a fixed value of 5 is finally obtained as the preset blocking threshold. This preset blocking threshold is synchronously and permanently stored in the local configuration file of the edge gateway and can be modified and adjusted as needed through the operation and maintenance platform.

[0056] The edge gateway reads the cumulative number of failures for each coffee machine in the edge isolation re-authentication queue in real time. It compares this cumulative number of failures with a preset blocking threshold. When the cumulative number of failures is greater than or equal to the preset blocking threshold of 5, the edge gateway immediately invokes the control interface blocking module to sever all data interaction links between itself and the unified control interface of that coffee machine, ceasing to send any random challenge codes, configuration commands, or upgrade commands to that coffee machine. Simultaneously, the edge gateway reads the pre-stored physical deployment path of the coffee machine from the queue entry. The physical deployment path includes the store address, device placement location, and gateway communication port number. It integrates the physical deployment path information, device number, cumulative number of failures, and blocking timestamp to generate a standardized alarm log. This log is then uploaded to the cloud operations and maintenance platform via a persistent communication link between the edge gateway and the cloud operations and maintenance platform for persistent storage and pop-up alarm display.

[0057] By continuously tracking the number of failed verifications of isolated queue devices, determining blocking thresholds through offline testing, blocking communication when limits are exceeded, and uploading complete alarm logs, we can achieve the isolation and control of risky coffee machines that cannot pass security verification for a long time, thus avoiding the security risks of command intrusion caused by unauthorized devices accessing the cluster.

[0058] Furthermore, the method provided in this application embodiment includes: The task feature attributes of the batch configuration task are used as the query vector. The explicit topology diagram and the environmental health vector are concatenated to form the key vector and value vector. The inner product of the query vector and the key vector is calculated and subjected to Softmax normalization to obtain the weight assignment. The value vector is weighted and summed according to the weight assignment to obtain the matching feature vector. The matching feature vector is processed through a linear mapping layer to output the corresponding gray-scale fitness score. The gray-scale sequence is generated in descending order of gray-scale fitness score.

[0059] Specifically, the edge gateway extracts the task feature attributes of batch configuration tasks, including four core attributes: configuration package type, task priority, estimated bandwidth usage, and estimated execution time. The categorical attributes are converted into standardized values ​​through integer encoding, and the numerical attributes are mapped to the 0 to 1 range through linear normalization. They are then concatenated in a fixed order to form a query vector with a dimension of 4, where the fixed order is configuration package type, task priority, estimated bandwidth usage, and estimated execution time. The query vector serves as the unified task benchmark feature for matching calculation.

[0060] For each coffee machine in the cluster that has passed security verification, the edge gateway extracts three topological attributes from the explicit topology diagram: communication response latency, link packet loss rate, and node online status. After normalizing the values ​​of these three attributes to the range of 0 to 1, they are concatenated with the corresponding 8-dimensional environmental health vector, one end of each dimension, to form a 11-dimensional comprehensive feature vector for the device. From this comprehensive feature vector, the edge gateway selects four features strongly correlated with task resource requirements—node online status, communication response latency, link packet loss rate, and overall energy consumption—as the key vector for the device, ensuring that the key vector dimension is consistent with the query vector (both are 4-dimensional). The complete 11-dimensional comprehensive feature vector is then used as the value vector for the device.

[0061] The edge gateway calculates the inner product of the query vector and the corresponding coffee machine key vector for each coffee machine. The inner product operation involves multiplying the values ​​of the corresponding dimensions of the two vectors one by one and then summing the results to obtain the original matching score for each coffee machine. The edge gateway collects the original matching scores of all coffee machines, calculates the natural index for each original matching score, and then divides the index result of a single device by the sum of the index results of all devices to obtain the weight allocation value corresponding to each coffee machine. Softmax normalization is then performed, and the sum of the weight allocation values ​​of all devices after processing is 1. A higher value indicates a higher priority for the device in adapting to the batch configuration task.

[0062] Next, the edge gateway multiplies the value vector corresponding to a single coffee machine with the weight allocation value of the device dimension by dimension, and then sums all the multiplied dimensional values ​​to obtain the matching feature vector corresponding to a single coffee machine. The matching feature vector is a single-dimensional numerical vector that integrates comprehensive information such as task adaptability, topology status and device health.

[0063] Furthermore, the edge gateway is pre-configured with a linear mapping layer, which is a single-input, single-output linear transformation structure. The operation rule is that the output value equals the weight coefficient multiplied by the input value, plus the bias value. The input dimension and the matching feature vector dimension are kept consistent at 1, and the output dimension is 1. The weight coefficients and bias values ​​of the linear mapping layer are obtained through offline supervised training. Specifically, device samples with verified configuration execution effects from historical batch configuration tasks are selected. Each sample contains the corresponding device's matching feature vector value as the input feature, and the actual configuration execution success rate of the device is used as the label. All samples together form the training dataset. The training process uses the least squares method for fitting and solving. The sum of squared errors between the predicted output values ​​and the labeled success rates of all training samples is used as the loss function. The partial derivatives of the loss function with respect to the weight coefficients and bias values ​​are calculated separately and set to zero to derive the analytical calculation formulas for the weight coefficients and bias values. Substituting all the sample values ​​in the training dataset yields the optimal weight coefficients and bias values. After training, the weight coefficients and bias values ​​are stored locally on the edge gateway. The edge gateway inputs the matching feature vector of a single coffee machine into the linear mapping layer. After the linear operation of multiplying the weight coefficients and superimposing the bias values, the grayscale fitness score corresponding to the coffee machine is output. The higher the score, the more suitable the device is to be included in the grayscale test range.

[0064] Finally, the edge gateway collects the grayscale fitness scores of all coffee machines that have passed security verification, i.e., coffee machines that have passed the entire process of random number challenge code and response code verification and communication response delay deviation verification. Then, all coffee machines are arranged in descending order of score value to generate the final grayscale sequence.

[0065] By using a cross-attention mechanism to associate three types of information—batch configuration tasks, explicit topology diagrams, and environmental health vectors—the device configuration adaptability is adaptively calculated and an ordered grayscale sequence is generated. This achieves precise matching between configuration tasks and terminal devices, ensuring operational stability and configuration execution reliability during the grayscale testing phase.

[0066] Furthermore, the method provided in this application embodiment includes: Coffee machines with a grayscale fitness score greater than or equal to a preset admission score in the grayscale sequence are selected; the business traffic data associated with the selected coffee machines are statistically analyzed, and coffee machines that are in peak business hours are removed; a preset number of coffee machines ranked at the top of the remaining coffee machines are selected as the grayscale test group, and the remaining coffee machines that have passed the verification are selected as the follower group.

[0067] In one embodiment, the edge gateway first reads the grayscale fitness score corresponding to each coffee machine within the complete grayscale sequence, pre-stores a fixed preset admission score locally, and combines it with the normalized value range of the grayscale fitness score from 0 to 1. For example, the preset admission score is typically set to 0.6, and can be flexibly modified by the cloud-based operation and maintenance platform according to the risk level of the batch configuration task. The edge gateway compares the grayscale fitness score of each coffee machine with the preset admission score, retains coffee machines with a grayscale fitness score greater than or equal to the preset admission score, and removes coffee machines with a grayscale fitness score less than the preset admission score, completing the first round of device screening.

[0068] Next, the edge gateway retrieves the sales traffic data associated with each coffee machine that passed the first round of screening. This sales traffic data records the number of orders generated by the corresponding device at different times each day. Using a specific store as an example, the edge gateway statistically defines peak sales periods based on the historical sales traffic data for the past thirty days. Specifically, the edge gateway calculates the average and standard deviation of order traffic for the same time period within that store over the past thirty days. The average plus twice the standard deviation is used as the peak threshold. Periods where order traffic consistently exceeds this threshold are defined as peak sales periods, and these peak sales periods are stored locally for each store. The edge gateway then compares the current operating time of each coffee machine with its peak sales period. If the current operating time falls within the peak sales period, the coffee machine is determined to be in a peak sales period, and it is removed from the screened device list, thus completing the second round of device removal.

[0069] The edge gateway then reads the preset quantity parameters stored locally. These preset quantity parameters can be customized and adjusted according to the scale of a single batch configuration task, through configuration commands issued by the cloud-based operations and maintenance platform. The edge gateway maintains the original descending sort order of the grayscale sequence for the coffee machines remaining after two rounds of screening, and selects the corresponding preset number of coffee machines sequentially starting from the first position in the sort, thus aggregating all selected devices to form a grayscale test group.

[0070] Finally, the edge gateway iterates through all coffee machines within the grayscale sequence that have completed security verification but have not been assigned to a grayscale test group, and groups these devices into a follower group. All coffee machines in the follower group are in a pending configuration state. Only after all coffee machines in the grayscale test group have completed the execution of the differentiated configuration package will the edge gateway send the corresponding batch configuration package to the follower group. The grayscale test group executes the batch configuration package first, while the follower group waits for all coffee machines in the grayscale test group to complete their configuration tasks before executing its corresponding batch configuration package.

[0071] By employing a hierarchical process of score-based filtering, removal of devices during peak business hours, and grouping a fixed number of devices by sorting, the gray-scale test group and the follow-up group are distinguished. This reduces the interference of batch configuration operations on normal store operations and lowers the cluster operation risks caused by configuration anomalies.

[0072] Furthermore, the method provided in this application embodiment includes: After receiving the differentiated configuration package, each coffee machine in the grayscale test group performs a parameter range pre-check and evaluates its load capacity in conjunction with the current hardware aging parameters. When the pre-check and load capacity evaluation pass, the local configuration register is locked and a pre-check ready signal is returned to the edge gateway. When the edge gateway receives the pre-check ready signals returned by all coffee machines in the grayscale test group within a preset waiting time window, it issues the execution command.

[0073] In this embodiment, the local configuration register is an embedded storage register inside the coffee machine used to persistently store various operating control parameters, supporting parameter writing, locking, and overwriting operations.

[0074] Optionally, the edge gateway pre-stores a full baseline configuration template. The template includes three types of basic data: a set of general configuration items, standard value ranges for each configuration item, and a hardware compatibility table for different device models. The specific data collection steps are as follows: For each coffee machine in the gray-scale test group, the edge gateway first extracts the device's model identifier and hardware version number, searches the model hardware compatibility table, filters out the configuration items supported by the model and hardware version, and removes incompatible configuration items, forming a set of compatible configuration items specific to that device. The edge gateway reads the coffee machine's currently active operating parameters, uses these parameters as a baseline, and, combined with the target adjustment direction of the batch configuration task, calculates the target parameter value within the legal value range of the corresponding configuration item, ensuring the parameter adjustment range is within the range of smooth device adaptation. The edge gateway encapsulates the adapted target configuration parameters, the legal upper and lower limits of each parameter, and the model-specific hardware compatibility verification rules in a fixed data format, generating a differentiated configuration package uniquely bound to each device identifier, and distributes it to each machine through a secure session channel. Upon receiving the differentiated configuration package, the coffee machine temporarily stores it in its local cache area, without directly overwriting the original operating parameters.

[0075] Each coffee machine reads the target configuration parameters from the differentiated configuration package and simultaneously retrieves the legal upper and lower limits of each parameter's value from the local storage. It then performs a value range comparison operation on each target parameter. If all target parameters fall within their corresponding legal value ranges, the parameter range pre-check is considered successful. If any parameter exceeds its legal value range, the pre-check is considered unsuccessful, the configuration package is discarded, and a pre-check failure signal is returned to the edge gateway.

[0076] After the parameter range pre-check passes, the coffee machine reads the locally stored hardware aging parameters, including three core aging data categories: cumulative runtime of the grinder motor, cumulative number of heating module operations, and cumulative number of valve actions. The coffee machine pre-stores the load-bearing threshold for each hardware component. This threshold is calibrated based on the hardware's factory-rated lifespan parameters and historical data from store equipment maintenance. Preferably, it is conventionally set to 80% of the corresponding component's rated lifespan. For example, for the grinder motor, its load-bearing threshold is set to 80% of its factory-rated total runtime. When the cumulative runtime reaches this threshold, the load-bearing capacity is considered close to its limit, and no further configuration write operations are accepted. The coffee machine compares the current aging parameter values ​​with the corresponding load-bearing thresholds. If all hardware aging parameters are below their respective thresholds, the hardware load-bearing capacity assessment is considered successful; if any parameter exceeds the threshold, the assessment is considered unsuccessful, the configuration process is terminated, and a failure signal is returned.

[0077] Once both the parameter range pre-check and hardware capacity assessment are passed, the coffee machine performs a write-protection lock operation on its local configuration register, prohibiting any parameter modification operations other than the current configuration write. After locking is complete, a pre-check ready signal is generated and transmitted back to the edge gateway through the established secure session channel. The signal carries the unique device identifier of the corresponding coffee machine and the lock status verification code.

[0078] While the edge gateway distributes the differentiated configuration package, it starts a preset waiting window timer. For example, the preset waiting window is typically set to 30 seconds. During the timer, the edge gateway continuously receives signals from each coffee machine in the gray-scale test group and counts the number of devices that have returned a pre-test ready signal in real time. When all coffee machines in the gray-scale test group have returned valid pre-test ready signals before the timer ends, the edge gateway immediately generates a unified execution command and synchronously distributes it to all coffee machines in the gray-scale test group through a secure session channel. After receiving the execution command, each coffee machine in the gray-scale test group releases the write-protected lock state of its local configuration register, writes the locally cached differentiated configuration package into the configuration register, and completes the update and activation of the target configuration parameters. If any device fails to return a ready signal after the waiting window timer ends, the gray-scale configuration pre-verification is deemed a failure, and the configuration distribution process is terminated. This termination operation occurs before the configuration package is written to the configuration register; the configuration register retains its original parameter state, and no configuration rollback operation is required.

[0079] By employing a step-by-step verification process that includes parameter range pre-check, hardware aging capacity assessment, configuration register locking, and unified triggering of the waiting window, dual verification of legality and hardware compatibility is completed before the configuration is officially written. This avoids the risk of invalid parameter writing and hardware overload, ensuring the stability and security of gray-scale configuration execution.

[0080] Furthermore, the method provided in this application embodiment includes: When the execution status is abnormal, stop issuing instructions to the follower group and lock the follower group to the current unconfigured state; the edge gateway issues a rollback instruction to the gray-scale test group through the secure session channel, overwrites the differentiated configuration package with the old parameter baseline temporarily stored locally by the edge gateway, and restores it to the historically available version.

[0081] In one embodiment, during the configuration process of the gray-scale test group, the edge gateway continuously receives execution status messages from each coffee machine. These messages contain three types of data: parameter writing results, device operating status codes, and heartbeat identifiers. The edge gateway pre-sets anomaly judgment rules; for example, the heartbeat detection cycle is set to 5 seconds. When any coffee machine is detected to have failed parameter writing, an abnormal operating status code, or no response for three consecutive heartbeat cycles, the execution status is determined to be abnormal, and a global circuit breaker mechanism is immediately triggered. After the circuit breaker is triggered, the edge gateway first clears the instruction delivery queue of the follower group, terminates all pending configuration and execution instructions, and simultaneously marks the device status of the follower group as locked and unconfigured, prohibiting any configuration writing operations for the follower group, ensuring that the devices in the follower group always maintain their original available parameters.

[0082] Before initiating the batch configuration task, the edge gateway extracts the currently effective operating parameters from each coffee machine in the gray-scale test group and temporarily stores them as the old parameter baseline in a dedicated local storage partition. Each old parameter baseline is uniquely bound to a device identifier, and the storage partition is write-protected, disallowing modification during configuration execution. After triggering the circuit breaker and before issuing the rollback command, the edge gateway first verifies the integrity of the local old parameter baseline. Verification and comparison confirm that the baseline data is not corrupted, ensuring the reliability of the rollback data source.

[0083] Next, the edge gateway sends a rollback command to each coffee machine in the gray-scale test group through the established secure session channel. The command carries the old parameter baseline data packet for the corresponding device. Upon receiving the rollback command, the coffee machine immediately terminates the current process of the differentiated configuration package, clears the locally cached differentiated configuration parameters, and then writes the old parameter baseline data into the local configuration register, overwriting the differentiated configuration parameters updated in this update, thus completing the parameter rollback operation and restoring to the historically available version before the configuration execution.

[0084] After the rollback operation is completed, each coffee machine automatically performs a parameter validity self-check, verifying the value range of all operating parameters, the hardware driver adaptation status, and the operating status of the core functional modules. If all parameters fall within the legal value range, driver adaptation is normal, and the core functional module self-check shows no errors, the self-check is considered successful. Once the device's operating status is confirmed to be normal, a rollback completion signal is sent back to the edge gateway. The edge gateway compiles the rollback results for all devices in the gray-scale test group. After all devices have completed the rollback and passed the self-check, the global circuit breaker is maintained, awaiting intervention from maintenance personnel to investigate the cause of the anomaly. During this period, the batch configuration task is not automatically restarted.

[0085] By executing a process that triggers a global circuit breaker to lock the following group and rolls back the gray-scale test group configuration based on the local old parameter baseline, the risk propagation is quickly terminated when an anomaly occurs during configuration execution, ensuring the operational stability of most devices in the cluster and reducing the business impact of batch configuration failures.

[0086] In summary, the secure batch configuration method for a coffee machine cluster using an edge gateway proxy provided in this application has the following technical effects: This application utilizes a cross-attention mechanism to match batch configuration tasks, generates grayscale sequences from explicit topology diagrams and environmental health vectors, hierarchically filters and divides grayscale test groups and follower groups, distributes differentiated configuration packages to grayscale test groups, and issues execution instructions after multiple pre-checks. In case of anomalies, a circuit breaker is triggered, and the configuration is rolled back based on the local old parameter baseline. This achieves safe and orderly batch configuration of coffee machine clusters, reduces the risk of configuration failure propagation, improves the stability of cluster configuration operation, and achieves linkage verification between configuration operations and equipment safety status. Furthermore, it performs hierarchical distribution and abnormal circuit breaker rollback based on the real-time status of the coffee machine cluster, thereby improving the technical effects of security, consistency, and reliability of the configuration process.

[0087] Example 2, as Figure 3 As shown, based on the same inventive concept as the aforementioned Embodiment 1, this application provides a secure batch configuration system for a coffee machine cluster using an edge gateway proxy, the system comprising: The environmental health vector generation module 1 is used by the edge gateway to construct an explicit topology diagram of multiple coffee machines in the coffee machine cluster, and to collect time-series data of each coffee machine during operation to generate an environmental health vector.

[0088] Secure session channel construction module 2 is used by the edge gateway to verify the response codes returned by multiple coffee machines according to the random number challenge code, and to establish a secure session channel when the verification is successful and the corresponding communication response delay meets the preset deviation.

[0089] The grayscale sequence generation module 3 is used by the edge gateway to match the batch configuration task, the explicit topology diagram and the environmental health vector using the cross attention mechanism to generate a grayscale sequence, and to divide the grayscale test group and the follower group according to the grayscale sequence.

[0090] The signal receiving and execution module 4 is used by the edge gateway to send differentiated configuration packages to the grayscale test group, and to issue execution instructions after receiving the pre-detection ready signal. When the execution status is abnormal, it triggers a global circuit breaker and performs configuration rollback using the local old parameter baseline.

[0091] Furthermore, the environmental health vector generation module 1 is used to perform the following steps: Discrete Fourier transform is performed on the time-series data continuously collected from each coffee machine to extract high-frequency harmonic component features; the high-frequency harmonic component features are input into a lightweight multilayer perceptron for feature mapping to obtain an environmental health vector, wherein the environmental health vector is used to reflect the operating status of the coffee machine.

[0092] Furthermore, the secure session channel construction module 2 is used to perform the following steps: Extract the historical average response latency from the environmental health vector; calculate the absolute difference between the current communication response latency and the historical average response latency; when the absolute difference is less than a preset millisecond-level security threshold, the verification is passed and the secure session channel is established.

[0093] Furthermore, the grayscale sequence generation module 3 is used to perform the following steps: The task feature attributes of the batch configuration task are used as the query vector. The explicit topology diagram and the environmental health vector are concatenated to form the key vector and value vector. The inner product of the query vector and the key vector is calculated and subjected to Softmax normalization to obtain the weight assignment. The value vector is weighted and summed according to the weight assignment to obtain the matching feature vector. The matching feature vector is processed through a linear mapping layer to output the corresponding gray-scale fitness score. The gray-scale sequence is generated in descending order of gray-scale fitness score.

[0094] Furthermore, the grayscale sequence generation module 3 is used to perform the following steps: Coffee machines with a grayscale fitness score greater than or equal to a preset admission score in the grayscale sequence are selected; the business traffic data associated with the selected coffee machines are statistically analyzed, and coffee machines that are in peak business hours are removed; a preset number of coffee machines ranked at the top of the remaining coffee machines are selected as the grayscale test group, and the remaining coffee machines that have passed the verification are selected as the follower group.

[0095] Furthermore, the signal receiving and execution module 4 is used to perform the following steps: After receiving the differentiated configuration package, each coffee machine in the grayscale test group performs a parameter range pre-check and evaluates its load capacity in conjunction with the current hardware aging parameters. When the pre-check and load capacity evaluation pass, the local configuration register is locked and a pre-check ready signal is returned to the edge gateway. When the edge gateway receives the pre-check ready signals returned by all coffee machines in the grayscale test group within a preset waiting time window, it issues the execution command.

[0096] Furthermore, the signal receiving and execution module 4 is used to perform the following steps: When the execution status is abnormal, stop issuing instructions to the follower group and lock the follower group to the current unconfigured state; the edge gateway issues a rollback instruction to the gray-scale test group through the secure session channel, overwrites the differentiated configuration package with the old parameter baseline temporarily stored locally by the edge gateway, and restores it to the historically available version.

[0097] Furthermore, the secure session channel construction module 2 is used to perform the following steps: When a coffee machine fails verification, it is automatically moved to the edge isolation re-authentication queue. After the gray-scale test group or the follower group completes the configuration task, the edge gateway reduces the authentication initiation cycle of the edge isolation re-authentication queue and re-initiates the verification.

[0098] Furthermore, the secure session channel construction module 2 is used to perform the following steps: The cumulative number of failures for a single coffee machine in the edge isolation re-authentication queue is counted. When the cumulative number of failures is greater than or equal to a preset blocking threshold, the edge gateway blocks communication with the control interface of the coffee machine and reports an alarm log containing the physical path of the coffee machine to the cloud.

[0099] The edge gateway proxy coffee machine cluster security batch configuration system provided in this embodiment of the invention can execute the edge gateway proxy coffee machine cluster security batch configuration method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0100] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0101] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A secure batch configuration method for a coffee machine cluster using an edge gateway proxy, characterized in that, The method includes: The edge gateway constructs an explicit topology diagram of multiple coffee machines in the coffee machine cluster and collects time-series data of each coffee machine during operation to generate an environmental health vector. The edge gateway verifies the response codes returned by multiple coffee machines based on random number challenge codes, and establishes a secure session channel when the verification is successful and the corresponding communication response delay meets the preset deviation. The edge gateway uses a cross-attention mechanism to match batch configuration tasks, the explicit topology diagram, and the environmental health vector to generate a grayscale sequence, and divides the grayscale test group and the follower group according to the grayscale sequence; The edge gateway sends a differentiated configuration package to the grayscale test group and issues an execution command after receiving the pre-detection ready signal. When the execution status is abnormal, a global circuit breaker is triggered and the configuration is rolled back using the local old parameter baseline. The edge gateway utilizes a cross-attention mechanism to match batch configuration tasks, the explicit topology diagram, and the environmental health vector to generate a grayscale sequence, including: The task feature attributes of the batch configuration task are used as query vectors, and the explicit topology diagram and the environmental health vector are concatenated to form key vectors and value vectors. The inner product of the query vector and the key vector is calculated and then normalized using Softmax to obtain the weight allocation. The value vectors are then weighted and summed according to the weight allocation to obtain the matching feature vector. The matching feature vector is processed by a linear mapping layer to output the corresponding gray fitness score, and the gray sequence is generated in descending order of gray fitness score. The edge gateway sends a differentiated configuration package to the grayscale test group and, upon receiving a pre-detection ready signal, issues an execution command, including: After receiving the differentiated configuration package, each coffee machine in the grayscale test group performs a parameter range pre-check and assesses its load-bearing capacity in conjunction with the current hardware aging parameters. Once the pre-inspection and load capacity assessment are passed, the local configuration register is locked, and a pre-inspection ready signal is returned to the edge gateway; When the edge gateway receives the pre-test ready signal returned by all coffee machines in the grayscale test group within the preset waiting time window, it issues the execution command.

2. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 1, characterized in that, Collect time-series data of each coffee machine during operation to generate an environmental health vector, including: Discrete Fourier transform was performed on the time-series data continuously collected from each coffee machine to extract high-frequency harmonic component characteristics; The high-frequency harmonic component features are input into a lightweight multilayer perceptron for feature mapping to obtain an environmental health vector, which is used to reflect the operating status of the coffee machine.

3. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 1, characterized in that, A secure session channel is established when verification is successful and the corresponding communication response delay meets the preset deviation, including: Extract the historical average response delay from the environmental health vector; Calculate the absolute difference between the current communication response delay and the historical average response delay; When the absolute difference is less than a preset millisecond-level security threshold, the verification is passed and the secure session channel is established.

4. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 1, characterized in that, The grayscale sequence is divided into a grayscale test group and a follower group, including: Coffee machines whose grayscale fitness score is greater than or equal to a preset admission score in the grayscale sequence are selected; The sales volume data associated with the selected coffee machines was analyzed, and coffee machines operating during peak hours were removed. Select the top-ranked number of coffee machines from the remaining coffee machines as the grayscale test group, and the remaining verified coffee machines as the follow-up group.

5. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 1, characterized in that, When an execution state is abnormal, a global circuit breaker is triggered, and a configuration rollback is performed using the local old parameter baseline, including: When the execution status is abnormal, stop issuing instructions to the follower group and lock the follower group to the current unconfigured state; The edge gateway sends a rollback command to the gray-scale test group through the secure session channel, overwriting the differentiated configuration package with the old parameter baseline temporarily stored locally on the edge gateway, and restoring it to a previously available version.

6. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 1, characterized in that, The edge gateway verifies the response codes returned by multiple coffee machines based on a random number challenge code, and also includes: When a coffee machine fails authentication, it is automatically moved to the edge isolation re-authentication queue. Once the gray-scale test group or the follower group has completed its configuration task, the edge gateway reduces the authentication initiation cycle of the edge isolation re-authentication queue and re-initiates the verification.

7. The secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in claim 6, characterized in that, The cumulative number of failures for a single coffee machine in the edge-isolated re-authentication queue is recorded. When the cumulative number of failures is greater than or equal to the preset blocking threshold, the edge gateway blocks communication with the control interface of the coffee machine and reports an alarm log containing the physical path of the coffee machine to the cloud.

8. A secure batch configuration system for a coffee machine cluster using an edge gateway proxy, characterized in that, A system for implementing a secure batch configuration method for a coffee machine cluster using an edge gateway proxy as described in any one of claims 1-7, the system comprising: The environmental health vector generation module is used by the edge gateway to construct an explicit topology diagram of multiple coffee machines in a coffee machine cluster, and to collect time-series data of each coffee machine during operation to generate an environmental health vector. The secure session channel construction module is used by the edge gateway to verify the response codes returned by multiple coffee machines based on random number challenge codes, and to establish a secure session channel when the verification is successful and the corresponding communication response delay meets the preset deviation. The grayscale sequence generation module is used by the edge gateway to match batch configuration tasks, the explicit topology diagram and the environmental health vector using a cross-attention mechanism to generate a grayscale sequence, and to divide the grayscale test group and the follower group according to the grayscale sequence. The signal receiving and execution module is used by the edge gateway to send differentiated configuration packages to the grayscale test group, and to issue execution instructions after receiving the pre-detection ready signal. When the execution status is abnormal, it triggers a global circuit breaker and performs configuration rollback using the local old parameter baseline.

Citation Information

Patent Citations

  • Power data transmission node dynamic control method based on gray release algorithm

    CN120416044A

  • Software automatic distribution and deployment method oriented to cloud edge collaboration

    CN122018938A