A message pushing method and system model based on an industrial control system

Abstract

The present application relates to the technical field of data processing, and more particularly to a message pushing method and system model based on an industrial control system, comprising the following steps: S1, obtaining a message priority and an urgency priority of a to-be-tested message; S2, inputting the to-be-tested message into a control rule library based on a strategy tree optimization for matching, obtaining a control rule of the to-be-tested message according to the message priority of the to-be-tested message, the control rule comprising a pushing permission configuration and a pushing path configuration; obtaining an emergency pushing configuration according to the urgency priority; S3, performing data preprocessing to obtain a to-be-output to-be-tested message; S4, pushing the to-be-output to-be-tested message and monitoring and managing the state of the pushed to-be-tested message. The present application realizes flexible configuration of message priority and permission setting by using a control rule library algorithm based on strategy tree optimization, which is convenient for adjusting the object range of message pushing according to different actual conditions, makes the message pushing mode more flexible, and the pushing object more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a message pushing method and system model based on an industrial control system. BACKGROUND

[0002] In recent years, with the continuous development of information technology and the strong support of national policy, the process of industrial system automation in China is constantly advancing. There are various kinds of field measurement and control devices in industrial control systems, which can detect the running state and environmental state information of field devices in real time, and push the processed information in the form of messages to the corresponding management platform and management personnel in a timely manner, so as to supervise the field working state in real time and adjust in time. Therefore, one of the most important parts of the industrial control system is message pushing. With the continuous development of industrial control systems, there are more and more digital information that needs to be received and processed by staff in the system, and batch message pushing technology is more and more widely used in the field of industrial control systems.

[0003] Message pushing is to automatically deliver information to users through pushing technology, which is regularly pushed to users according to their needs and permissions, and the messages reach the target personnel of the enterprise in a timely manner through various forms, which can help enterprise personnel to obtain information faster and more efficiently. Through message pushing, on the one hand, staff can monitor equipment flow information in a timely manner and predict industrial field information changes to respond; on the other hand, if a fault occurs, the fault information can be pushed in a timely manner for timely processing; in addition, message pushing can also realize information interaction between different subsystems. And the cost of message pushing is relatively low, and the use of automated message pushing can greatly save the labor cost of enterprises. The main ways of message pushing are network, short message, telephone, broadcast, etc., among which network pushing is the mainstream way at present. For example, in the process of industrial production, if a device fault occurs suddenly, the alarm information is displayed in the form of a pop-up window on the management platform interface, or broadcasted in the form of a broadcast to remind the staff to handle it in a timely manner.

[0004] The current industrial control system needs further improvement and development in message management, message configuration, message pushing and delivery, and there are three problems to be solved:

[0005] (1) The flexibility of user permissions and message pushing priority;

[0006] (2) The efficiency problem of batch pushing of a large number of messages;

[0007] (3) The message pushing problem in response to sudden conditions.

[0008] In order to solve the above problems existing in the information pushing process, different solutions are proposed by many scholars. Some scholars use the role-based access control method (RBAC) to separate users and permissions, so that users are assigned corresponding roles to obtain the permissions of the roles, thereby reducing the complexity of permission management, but the importance of the message to the pushing permission is not considered. Some scholars use RabbitMQ to deliver and deliver confirmation of small messages, which significantly improves the sending rate of persistent small messages. However, RabbitMQ is written based on erlang language, and the research and development investment cost is high, and the throughput is not as good as KafkaMQ and RocketMQ. In addition, when dealing with sudden conditions, the industrial control system also needs more flexible message pushing forms.

[0009] In summary, the existing industrial control system message pushing method still needs to be further improved and optimized in terms of permission control, efficiency improvement and sudden condition response. Therefore, in view of the above problems, a new industrial control system information pushing method is needed to improve the flexibility of authorization control, system running efficiency and the ability to respond to sudden conditions. SUMMARY

[0010] In view of the problems existing in the prior art, the present application provides a message pushing method and system model based on an industrial control system, which solves the problems of the existing industrial control system message pushing method in terms of permission control, efficiency improvement and sudden condition response.

[0011] The present application is realized by the following technical solutions:

[0012] A message pushing method based on an industrial control system, comprising the following steps:

[0013] S1, obtaining the message priority and emergency priority of the to-be-tested message;

[0014] S2, inputting the to-be-tested message into a control rule library based on a strategy tree optimization for matching, and obtaining the control rule of the to-be-tested message according to the message priority of the to-be-tested message, wherein the control rule includes a pushing permission configuration and a pushing path configuration;

[0015] The emergency priority obtains an emergency pushing configuration;

[0016] S3, data preprocessing is performed on the pushing permission configuration, the pushing path configuration and the emergency pushing configuration to obtain the to-be-output to-be-tested message;

[0017] S4, pushing the to-be-output to-be-tested message, and monitoring and managing the state of the pushed to-be-tested message.

[0018] Preferably, the RBF neural network is used to identify the emergency priority of the to-be-tested message.

[0019] Preferably, the establishment step of the control rule library based on the strategy tree optimization is as follows:

[0020] S11, obtaining data fields of a plurality of real-time messages from a real-time database;

[0021] S12, sequentially comparing the real-time data fields of the plurality of real-time messages with model data fields of a data model library to determine the control rules corresponding to the real-time messages according to the determination results, and storing the control rules corresponding to the real-time messages to form a control rule library based on message priority.

[0022] Preferably, in S12, when the priority P of the real-time data field is less than or equal to the priority T of the model data field, the control rule in the model database remains unchanged. n o

[0023] When the priority P of the real-time data field is greater than the priority T of the model data field of the data model library, a new control rule corresponding to the real-time data field is added according to the priority P of the real-time data field, and the control rule library is updated. n o n

[0024] Preferably, in S2, when the priority of the to-be-tested message is less than or equal to the priority T of the model data field, the corresponding control rule is read according to the priority of the to-be-tested message; when the priority of the to-be-tested message is greater than the priority T of the model data field of the data model library and less than the priority P of the real-time data field, the corresponding new control rule is read. o o n

[0025] Preferably, in S3, the preprocessing is performed in the Kafka message queue; and in the preprocessing, a DDG clustering algorithm is used to process the data in the Kafka message queue.

[0026] Preferably, in S4, the emergency priority of the to-be-tested message is combined with the user location information to push the to-be-tested message nearby.

[0027] Preferably, in S3, the processing state of the to-be-tested message after being pushed includes unprocessed, processing, processed, and no need to process, and the to-be-tested message that is unprocessed is repeatedly pushed.

[0028] ​​​​​​​​The system model of the message pushing based on the industrial control system comprises a data acquisition module, a message generation module, a message configuration module, a message pushing module and a message management module connected in sequence, the data acquisition module is used for acquiring data fields of a to-be-tested message, the message generation module is used for acquiring a priority of the to-be-tested message, the message configuration module is used for analyzing the to-be-tested message and acquiring a pushing permission configuration, a path configuration, an emergency priority and an emergency pushing configuration of the to-be-tested message, the message pushing module pushes the to-be-tested message to a receiving end, and the message management module is used for managing and recording the to-be-tested message which has been pushed.

[0029] Compared with the prior art, the present application has the following beneficial effects:

[0030] The message pushing method based on the industrial control system adopted by the present application realizes flexible configuration of message priority and permission setting by using a strategy tree optimization control rule library algorithm, facilitates adjustment of the object range of message pushing according to different actual conditions, makes the message pushing mode more flexible and the pushing object more accurate. Meanwhile, the strategy tree is used to optimize the control rule library, thereby improving the efficiency of permission configuration.

[0031] Further, in response to an emergency situation, the RBF neural network algorithm is used to identify the emergency degree of the situation, and the user is pushed in proximity according to the emergency degree level and the user position information, thereby realizing efficient processing in the emergency situation.

[0032] Further, in the data processing layer, the characteristics of large quantity and multiple types of data acquired by the industrial field device are considered. In the reprocessing process, the data fields of the to-be-tested message are clustered by using the DDG clustering algorithm, and the clustering results are distributed to the partitions corresponding to the Kafka message queue, thereby pushing the to-be-output to-be-tested message. The large quantity of data acquired by the field device is clustered by using the DDG clustering algorithm based on the Kafka message queue, thereby effectively improving the data processing efficiency and improving the timeliness of the message. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 It is a message pushing system model diagram;

[0034] Figure 2 It is a control rule library permission configuration structure diagram based on strategy tree optimization;

[0035] Figure 3 It is a DDG data clustering efficiency improvement structure diagram based on the Kafka message queue;

[0036] Figure 4 It is a structure diagram for realizing the nearby pushing based on the RBF neural network algorithm;

[0037] Figure 5A schematic diagram of a system model based on message pushing of an industrial control system. DETAILED DESCRIPTION

[0038] The application will be further described in connection with specific embodiments, which are intended to explain the application but not to limit it.

[0039] The application discloses a message pushing method based on an industrial control system, referring to Figure 1 , comprising the following steps:

[0040] S1, obtaining the message priority and the emergency priority of the to-be-tested message;

[0041] S2, inputting the to-be-tested message into a control rule library based on a strategy tree optimization to match, obtaining the control rule of the to-be-tested message according to the message priority of the to-be-tested message, and the control rule comprising a pushing permission configuration and a pushing path configuration;

[0042] The control rule library is optimized by using the strategy tree method to obtain an optimized control rule library. The specific steps are as follows:

[0043] S11, obtaining the data fields of a plurality of real-time messages from a real-time database;

[0044] S12, comparing and judging the real-time data fields of the plurality of real-time messages with the model data fields of a data model library in sequence, obtaining the control rule corresponding to the real-time message according to the judgment result, and storing the control rule of the real-time message correspondingly to form a control rule library based on the message priority.

[0045] The permission configuration process is as follows: data acquisition and analysis, establishment of a control rule library and optimization of the rule library based on a strategy tree.

[0046] The steps of data acquisition and analysis are as follows: firstly, a transmission data packet is acquired according to the protocol type of different devices, then the captured data packet is disassembled and analyzed, the byte segment containing data information in the data packet is extracted, and the data information carried in the data packet is acquired by analysis, for subsequent generation of the control rule library.

[0047] The establishment steps of the control rule library are as follows:

[0048] 1) establishing a basic permission control rule library based on the message priority.

[0049] 2) analyzing the data packet in step 2.1 to construct a set of priorities of information contained in the data packet: U p ={P1,P2,P3,...}

[0050] P n represents the priority value of each piece of data.

[0051] 3) Assign an initial value T to the threshold condition based on the staff's experience. o .

[0052] 4) Prioritize information P n With threshold condition T o Comparison:

[0053] When the priority P of the real-time data field n The priority T of the model data field in the data model library is less than or equal to o At that time, the control rules in the model database remain unchanged.

[0054] When the priority P of the real-time data field n The priority T of the model data fields in the data model library is greater than that of the model data fields. o At that time, based on the priority P of the real-time data field n Add corresponding new control rules and update the control rule base to form a control rule base based on message priority.

[0055] The steps for adding new control rules are as follows: Based on the settings of different departments, expand or reduce the scope of the control rules, or notify staff to set new rules and add new threshold conditions T. n and P n Assign the new threshold condition T. n .

[0056] Repeat this process until all data has been processed, resulting in a permission control rule base based on message priority.

[0057] The above-mentioned access control rule base is optimized using the policy tree method. The specific steps are as follows:

[0058] 1) Based on the rule set, select the IP address as the root node, take the data range with the same attribute as the non-leaf nodes in the first layer, and take the corresponding rule operation as the leaf node to build the policy tree.

[0059] 2) After constructing the policy tree, detect adjacent non-leaf nodes with the same attribute. Let two adjacent non-leaf nodes be branches. i .node and branch j .node, whose data value range is branch i .node.value and branch j .node.value, whose subtrees are branches i .node.subtree and branch j .node.subtree. When two nodes simultaneously meet the following conditions:

[0060] branch i .node.value and branch j .node.value is continuous

[0061] branch i .node.subtree = branch j .node.subtree

[0062] Then you can branch i .node and branch j Merging .node nodes reduces leaf nodes, which in turn reduces the number of rules, thus optimizing the rule base.

[0063] The establishment of a comprehensive access control rule base and the optimization of the policy tree rule base enable push permission configuration based on message priority.

[0064] S2, input the message to be tested into the control rule base for matching, obtain the message priority and urgency priority of the message to be tested; then obtain the push permission configuration and push path configuration of the message to be tested based on the message priority and urgency priority of the message to be tested, and reprocess the push permission configuration and push path configuration of the message to be tested to obtain the message to be output.

[0065] The message priority identification process is as follows: when the priority of the message under test is less than or equal to the priority T of the model data field... o When the priority of the message to be tested is greater than the priority T of the model data field in the data model library, the corresponding control rules are read. o And its priority is lower than that of the real-time data field P. n At that time, read the corresponding new control rules.

[0066] The urgency priority of the message to be tested is identified using an RBF neural network. When the priority of the message to be tested is greater than the priority T of the model data field in the data model library, the priority is determined. o At that time, the priority of the message to be tested is combined with the user's location information, and the message to be tested is pushed to the nearest location.

[0067] The specific implementation is based on the RBF neural network algorithm to learn and model multidimensional data information, obtain a model to classify the importance of information, and assign the classification results to different location ranges. Then, user location information is obtained and messages are pushed to users within the corresponding range.

[0068] The implementation steps include establishing and preprocessing the original dataset and building and testing the RBF neural network model.

[0069] The specific steps of establishing the original data set and the pretreatment are as follows:

[0070] Firstly, a multi-dimensional attribute data set related to the emergency degree of the message provided by the industrial field device is acquired, including the peak value of certain data, the amplitude exceeding the normal value, the time exceeding the normal range, the mutation rate, etc., to constitute an original multi-dimensional data set, and each group of data is given a corresponding message pushing range value.

[0071] Then, the original multi-dimensional data set is dimensionally reduced and pretreated. The equal distance feature mapping algorithm is used to reduce the dimension of the original multi-dimensional data set.

[0072] Finally, the pretreated data set is randomly divided into a training set and a test set according to the sample quantity size in a ratio of n1:n2.

[0073] The specific steps of establishing and testing the RBF neural network model are as follows:

[0074] The RBF neural network is also called a radial basis function neural network, and the model structure thereof is divided into three layers: an output layer, a hidden layer and an output layer. The classification idea is to directly map the input vector into the hidden layer space, transform the linearly inseparable low-dimensional data into high-dimensional space, and make it linearly separable, so that the mapping from the hidden layer space to the output layer space is linear. The nonlinear transformation formula of the input layer and the hidden layer is as follows:

[0075]

[0076] Wherein, x is an n-dimensional network input vector; represents the radial basis function of the hidden layer neuron; c j is the data center value of the jth hidden layer neuron, which is a vector with the same dimension as the input vector x; σ j represents the width of the jth hidden layer neuron; ||x-c j || represents the norm of the vector x-c j , representing the distance between the vector x and c j .

[0077] The linear transformation formula of the hidden layer and the output layer is as follows:

[0078]

[0079] Wherein, y k represents the output value of the kth output neuron in the network output layer; W is an m*l (m represents the number of hidden layer neurons, and l represents the number of output layer nodes) dimensional output weight matrix; represents the connection weight value between the jth hidden layer neuron and the kth output layer neuron.

[0080] In the present application, the function of the input layer is the information group in the pre-processed data set, the hidden layer is the model to be built, and the output layer is the corresponding message push range.

[0081] The RBF neural network is modeled using the training set and the test set, and the specific steps are as follows:

[0082] First, the input layer, hidden layer and output layer network structure of the RBF neural network are constructed, and the center vector, variance and weight parameters of the RBF neural network are initialized;

[0083] Then, the training set data obtained in the previous step is substituted, and the model is trained by the adaptive momentum estimation algorithm to obtain the corresponding parameters;

[0084] After that, the test set data is used to simulate the RBF neural network model trained, and the setting parameters are adjusted;

[0085] Finally, the accuracy, recall rate and F1 value of the prediction results of the model are calculated to obtain the performance evaluation of the model.

[0086] Through the RBF neural network model trained, the industrial system can perform message push in the corresponding range according to the urgency of the information, so as to effectively deal with the sudden situation in the production process.

[0087] During the reprocessing, the DDG clustering algorithm is used to cluster the data fields of the to-be-tested message, and the clustering results are distributed to the partitions corresponding to the Kafka message queue.

[0088] The implementation steps include generating a device dependency graph and DDG clustering and result distribution.

[0089] The specific steps of generating the device dependency graph are as follows:

[0090] First, the device data byte stream is deserialized to obtain an original data set S, which contains multiple fields, and a field name A is assumed. For field A, the information gain ratio is calculated, which represents the information gain ratio of data set S when A field is used as the division basis, and the calculation formula is as follows:

[0091]

[0092] Where G(S, A) represents the information gain of data set S using A field as the division basis, and the calculation formula is as follows: A (S) is the information entropy of data set S with respect to A, and the calculation formula is as follows:

[0093] G(S, A) = H(S) - H(S|A)

[0094]

[0095] In the formula, H(S|A) represents the empirical conditional entropy of feature A to data set S, and the calculation formula is as follows:

[0096]

[0097] Recursive calculation is performed for each division, for each recursion, feature A is taken, and division is performed according to the value of A until the subclass cannot be divided; finally, the clustering result of the data is obtained, and the result is sent to all computing nodes, which are responsible for similarity calculation, DDG generation and clustering operation, wherein the similarity calculation method is as follows, wherein D A and D B represent the data of device A and device B respectively:

[0098]

[0099] In the formula, the Jaccard(O A , O B ) calculation method is as follows:

[0100]

[0101] Thus, the dependency relationship between each device can be calculated, and a device dependency graph can be generated according to the calculated result.

[0102] The specific steps of DDG clustering and result distribution are as follows:

[0103] First, the master node obtains the calculation result of the device dependency graph, and merges the leaves connected by the sub- DDGs with the same root;

[0104] Then, the number n of connected subgraphs of the merged DDG is obtained, if n is less than the minimum number of subclasses, each connected graph is directly generated into a subgraph and returned; otherwise, each graph node is initialized as an independent subset, and iteration is started, each iteration obtains a pair of subsets with the maximum similarity, if the subsets exist in the DDG, the union set is performed according to the dependency relationship.

[0105] Through the above operation, the clustering result can be obtained, and then the result is distributed. In the result distribution, the Kafka message queue partitioner is defined according to the device type, and the clustering result data is distributed to the corresponding partition of the Kafka message queue, so that the data processing efficiency is improved and the timeliness of the message is improved.

[0106] S3, pushing the to-be-output test message, and monitoring and managing the state of the pushed to-be-output test message.

[0107] The processing state of the pushed to-be-tested message is recorded, and the processing state of the to-be-tested message includes unprocessed, processing, processed and no need to process, and the to-be-tested message that is unprocessed is repeatedly pushed.

[0108] The application further discloses a system model based on message pushing of an industrial control system, referring to Figure 5 , comprising a data acquisition module, a message generation module, a message configuration module, a message pushing module and a message management module.

[0109] The data acquisition module is used for acquiring data fields of a to-be-tested message. The data acquisition module uploads information collected by an industrial field device to a real-time database in a timing polling manner, and the acquired data fields include event time, subsystem ID, device ID, message coding and message record value.

[0110] The message generation module is used for acquiring a priority of a to-be-tested message. The message generation module compares the acquired data fields with a data model library, obtains the type and priority of the message, stores the message in a message database, and assembles complete pushing text content according to a message information flow table format stored in the message database.

[0111] The message configuration module is used for analyzing a to-be-tested message and acquiring pushing permission configuration, path configuration, emergency priority and emergency pushing configuration of the to-be-tested message. The message configuration module mainly includes permission configuration and pushing path configuration of a message pushing object. The pushing permission configuration of different messages is determined according to the priority of the message. The pushing path configuration of the message is determined according to the importance and timeliness requirement of the message.

[0112] The message pushing module pushes a to-be-tested message to a receiving end. The message pushing module queries the configuration result of the message configuration module, and pushes a specific message to a configured receiving user through a corresponding path.

[0113] The message management module is used for managing and recording a to-be-tested message that has been pushed. The message management module acquires processing state information of the pushed message, divides the processing state into "unprocessed", "processing", "processed" and "no need to process", and selects repeated pushing or no longer pushing according to the processing state.

[0114] Firstly, field industrial devices collect various data, and upload the data to a real-time database in a timing polling manner through calling a serial port, a network port or a WIFI path.

[0115] In the real-time database, the real-time data fields of the message are compared with model data fields of a data model library to determine the type and importance level of the real-time data fields, and the real-time data fields are stored in a to-be-tested message library.

[0116] Secondly, the control rule base is optimized according to the priority of the message, and the control rule base is optimized using a strategy tree, and then the message push permission configuration is obtained, and the push path is determined according to the message priority.

[0117] Thirdly, the message push permission configuration and path configuration are queried and obtained, and the message is pushed to the corresponding user, the DDG clustering algorithm based on the Kafka message queue is used to cluster the to-be-tested message, and the clustering result is sent to the partitioner, and the data set is distributed to the partition corresponding to the message queue for processing, so as to avoid the problem of uneven data load distribution in each node in the message queue, and improve the data processing efficiency.

[0118] At the same time, for the message push in emergency, the RBF neural network is used to identify the emergency degree of the message and the nearest push is carried out combined with the user location information.

[0119] Finally, the processing state information of the pushed message is recorded through the message management module, and important information is repeatedly pushed.

Claims

1. A message push method based on an industrial control system, characterized by, The method comprises the following steps: S1, obtaining the message priority and the emergency priority of the to-be-tested message; S2, inputting the to-be-tested message into the control rule library based on the strategy tree optimization to match, obtaining the control rule of the to-be-tested message according to the message priority of the to-be-tested message, and the control rule comprising the push permission configuration and the push path configuration; obtaining the emergency push configuration according to the emergency priority; The establishment step of the control rule library based on the strategy tree optimization is as follows: S11, obtaining the data fields of a plurality of real-time messages from a real-time database; S12, compare the real-time data fields of the plurality of real-time messages with the model data fields of the data model library in sequence to determine, according to the determination result, the control rules corresponding to the real-time messages, and store the control rules of the real-time messages correspondingly to form a control rule library based on message priority; when the priority of the real-time data field is less than or equal to the priority of the model data field , the control rule in the model database remains unchanged . When the priority of the real-time data field is greater than the priority of the model data field of the data model library , the corresponding new control rule is added according to the priority of the real-time data field , and the control rule library is updated . When the priority of the message to be tested is less than or equal to the priority of the model data field , the corresponding control rule is read according to the priority of the message to be tested; when the priority of the message to be tested is greater than the priority of the model data field of the data model library and less than the priority of the real-time data field , the corresponding new control rule is read; S3, performing data preprocessing on the push permission configuration, the push path configuration and the emergency push configuration to obtain the to-be-output to-be-tested message; wherein the preprocessing process is performed in the Kafka message queue; during preprocessing, the data in the Kafka message queue is processed by using the DDG clustering algorithm; S4, pushing the to-be-output to-be-tested message, and monitoring and managing the state of the pushed to-be-tested message.

2. The message push method based on an industrial control system according to claim 1, characterized by, The RBF neural network is used to identify the emergency priority of the to-be-tested message.

3. The industrial control system based message push method of claim 1, wherein, In S4, the emergency priority of the to-be-tested message is combined with the user location information during pushing, and the to-be-tested message is pushed nearby.

4. The industrial control system based message push method of claim 1, wherein, In S3, the processing state of the pushed to-be-tested message comprises unprocessed, processing, processed and no need to process, and the to-be-tested message is repeatedly pushed.

5. A system model for push-based messaging for industrial control systems, for implementing the push-based messaging method for industrial control systems according to any one of claims 1 to 4, characterized in that The method comprises the following steps: S1, obtaining the message priority and the emergency priority of the to-be-tested message; S2, inputting the to-be-tested message into the control rule library based on the strategy tree optimization to match, obtaining the control rule of the to-be-tested message according to the message priority of the to-be-tested message, and the control rule comprising the push permission configuration and the push path configuration; obtaining the emergency push configuration according to the emergency priority; The establishment step of the control rule library based on the strategy tree optimization is as follows: S11, obtaining the data fields of a plurality of real-time messages from a real-time database; S3, performing data preprocessing on the push permission configuration, the push path configuration and the emergency push configuration to obtain the to-be-output to-be-tested message; wherein the preprocessing process is performed in the Kafka message queue; during preprocessing, the data in the Kafka message queue is processed by using the DDG clustering algorithm; S4, pushing the to-be-output to-be-tested message, and monitoring and managing the state of the pushed to-be-tested message. The RBF neural network is used to identify the emergency priority of the to-be-tested message. In S4, the emergency priority of the to-be-tested message is combined with the user location information during pushing, and the to-be-tested message is pushed nearby. In S3, the processing state of the pushed to-be-tested message comprises unprocessed, processing, processed and no need to process, and the to-be-tested message is repeatedly pushed.

Images