A data traceability system and method based on DDS

Abstract

The present application relates to a kind of data traceability system and method based on DDS, the system includes two layers of monitoring layer and terminal layer, terminal layer includes subsystem data collection module and subsystem data stream upload module, subsystem data stream upload module uploads data to monitoring layer by DDS;Monitoring layer includes interactive data acquisition module based on non-intrusive, inter-subject data item correlation analysis module, subject-based data evolution sequence generation module and visual display module;The technical scheme supports real-time data traceability in the process of system operation, not post-analysis, not dependent on system operation log;For the dynamic addition of subject in running state, the subject publishing and subscribing relationship and data flow path can be updated in real time;Using publish / subscribe mechanism, it brings non-intrusive data acquisition mode, so it is not necessary to modify the code of running system, and it is not necessary to set probe, so that the data traceability method has good flexibility and scalability.

Description

Technical Field

[0001] The present invention relates to a DDS-based data tracing system and method, belonging to the technical field of data tracing. Background Art

[0002] Data provenance tracing involves tracing the flow and value changes of faulty data items when a system experiences an operational failure. Most existing data provenance tracing methods rely on log analysis. These methods, however, have drawbacks such as intrusive data acquisition, a lack of real-time data analysis and traceability, and a lack of support for dynamic addition of system modules.

[0003] The Data Distribution Service (DDS) is a data distribution service specification developed by the International Organization for Standardization (OMG) in 2003. Participants exchange data by publishing and subscribing to topics, characterized by temporal and spatial decoupling. Based on an automatic discovery protocol, it supports the dynamic joining and leaving of nodes. Using DDS for inter-subsystem communication is currently a research hotspot, but its application for data traceability has yet to be explored.

[0004] Existing blockchain-based wide-area data interaction log traceability technology utilizes blockchain-based log evidence and traceability methods to collect multi-dimensional log data, parse it into structured log records, and then perform anomaly detection and analysis. This technology avoids manual intervention and has high reliability and authority. However, it has the following shortcomings:

[0005] (1) This method requires setting up a tracking point at the monitoring data location to generate a log, which will modify the code of the running system and is an invasive data acquisition method;

[0006] (2) The data source depends on the system log. In order to reduce the impact on performance, the log module in large systems often adopts an asynchronous implementation method, so the real-time nature of the data cannot be guaranteed. The log record lags behind the actual data interaction. Therefore, the literature cannot support the system to trace the data in real time during the dynamic operation process, and does not have the advantage of real-time analysis.

[0007] (3) The document uses smart contracts to trigger logs and record them on the chain. However, due to the irreversibility of smart contracts, they cannot be revoked or changed after they are generated. Once the system modules are adjusted or modified, it is difficult to adjust the smart contracts to adapt to the dynamic changes, and the maintainability is poor. Therefore, a new solution is urgently needed to solve the above technical problems. Summary of the Invention

[0008] The present application is just aimed at the problems in the prior art, and provides a data tracing system and method based on DDS, which cross uses a Topic-based Relevance Analysis Algorithm (TRAA), a Relevance Enhancement Analysis Algorithm based on Ontology (REAAO) and a Relevance Enhancement Analysis Algorithm based on Breath-First Search and Data Item Dependency (REAABFSDID) to complete relevance analysis of data items among topics; a global interactive data space is further constructed, a data evolution sequence is generated by using a Data Evolution Sequence Generation Algorithm (DESGA), and finally the data evolution sequence is visually displayed, so as to facilitate real-time monitoring and management of operation and maintenance personnel.

[0009] In order to achieve the above-mentioned purpose, the technical scheme of the present application is as follows: a data tracing system based on DDS, the system comprising a monitoring layer and a terminal layer, the terminal layer comprising a subsystem data collection module and a subsystem data stream uploading module, the subsystem data stream uploading module uploading data to the monitoring layer by means of DDS;

[0010] The monitoring layer comprises a non-intrusive interactive data acquisition module, a topic-based data item relevance analysis module, a topic-based data evolution sequence generation module and a visual display module.

[0011] The subsystem data collection module collects data of the subsystem (including business data and publish / subscribe relationship data (empty if no change occurs));

[0012] The non-intrusive interactive data acquisition module acquires interactive data in a completely non-intrusive manner when data distribution based on topics is performed by using the publish / subscribe mechanism of DDS, without modifying the code in the original system or setting a probe, but only needs to subscribe to all topics in the system in the interactive data acquisition module, and therefore, when the topic values obtained by subscription are saved, a time stamp is added,

[0013] The inter-theme data item correlation analysis module is responsible for analyzing the correlation relationship between each theme data item. It specifically includes three submodules: the theme-based correlation analysis module, the theme-based correlation analysis module, and the correlation enhancement analysis module based on data item dependency.

[0014] The data evolution sequence generation module based on the theme is combined with the non-intrusive interactive data acquisition module and the results of the inter-theme data item correlation analysis module to generate the data evolution sequence using DESGA.

[0015] The visualization module displays the data evolution sequence in real time, facilitating the monitoring and management of operation and maintenance personnel and enabling real-time tracing of fault data.

[0016] A data tracing method based on DDS, the method comprising the following steps:

[0017] Step 1: The subsystem data collection module collects business data, triggers the collection of publish / subscribe relationship data, and sends it to the subsystem data stream upload module.

[0018] The terminal layer is composed of various subsystems, and the data interaction between subsystems is completed with the help of DDS. The subsystem data collection module collects business data and sends it to the subsystem data stream upload module.

[0019] If a new node joins or a node exits, the subsystem data collection module triggers the generation of publish / subscribe relationship data and sends it to the subsystem data stream upload module.

[0020] Step 2: Based on the non-invasive interactive data acquisition module, subscribe to all topics in the system in the subsystem data stream upload module to obtain interactive data and add a timestamp when saving.

[0021] Step 3: The inter-theme data item correlation analysis module first models the traceability system, and then cross-uses TRAA, REAABO and REAABBFTDID to complete the correlation analysis of inter-theme data items.

[0022] Step 4: Construct a global interactive data space based on the topic-based data evolution sequence generation module, and then use DESGA to generate data evolution sequences.

[0023] Step 5: Visualize the data evolution sequence to facilitate real-time monitoring and management by operation and maintenance personnel.

[0024] Step 3 is as follows:

[0025] Step 3.1: Model the traceability system.

[0026] Assume that the system S consists of n subsystems, S={s1,s2,……,s n}, n e N, where s1, s2, …, s n For each subsystem, N is a natural number set, let the index set of each subsystem I1 = {1, 2, …, n}, the data interaction between subsystems is based on DDS, that is, each subsystem provides or obtains the required data by publishing or subscribing to the topic,

[0027] Suppose there are m topics in the system S, the set of topics T = {t1, t2, …, tm}, m e N, m}, m e N, where t1, t2, …, tm m For each topic, let the index set of each topic I2 = {1, 2, …, m}, each element t k ,k e I2 will have a publisher subsystem and a subscriber subsystem, which belongs to the subsystem set S,

[0028] Any topic is a structure, which is a combination of multiple data items, suppose there are p data items, let the index set of each data item I3 = {1, 2, …, p}, p e N,

[0029] Suppose t k = {a k,1 , a k,2 , …, a k,p}, k e I2, where a k,r ,k e I2, r e I3 refers to the rth data item in the topic t k

[0030] Data provenance is to obtain the evolution sequence of the specified data item a k,r according to the system running process, let seq = <l1, l2, …, l q >, q e N, where l1, l2, …, l q are each element in the evolution sequence of data item a k,r ;

[0031] Step 3.2:

[0032] Suppose the topic publishing and subscribing relationship between subsystems is known, the data item correlation analysis between topics is carried out. The data item correlation analysis between topics is to establish the data item-topic set correspondence relationship centered on the data item according to the obtained topic information. This analysis can be specifically divided into topic-based correlation analysis and correlation enhancement analysis. Correlation enhancement analysis is to further improve the data item-topic set relationship through ontology and known data item dependency relationship, so as to mine implicit correlation relationship. Correlation enhancement analysis is specifically divided into ontology-based correlation enhancement analysis and data item dependency relationship-based correlation enhancement analysis. The specific steps are as follows,

[0033] ​Step 3.2.1: For the collected data, use TRAA to do relevance analysis, the input of TRAA is topic set T, the output is the associated topic set of each data item, the specific steps are as follows,

[0034] Step 3.2.1.1: Get the first data item a of the first topic 1,1 ; initialize the associated topic set C of data item a 1,1 as an empty set. 1,1

[0035] Step 3.2.1.2: Check whether a 1,1 appears in t2, if it appears, add t2 to C 1,1 ;

[0036] Step 3.2.1.3: Return to step 3.2.1.2 to check the next topic until all topics in T have been checked.

[0037] Step 3.2.1.4: Return to step 3.2.1.1 to check the next data item until all data items in T have been checked.

[0038] Step 3.2.2: For the collected data, use REAAO to do relevance enhancement analysis, the input of REAAO is topic set T, the output is the associated topic set of each data item. REAAO is modified relative to TRAA, the step 3.2.1.2 is modified and the ontology library is added as input, each record in the ontology library is represented by a two-tuple, that is, <data item, synonym set>, the specific steps are as follows,

[0039] Step 3.2.2.1: Get the first data item a of the first topic 1,1 ; initialize C 1,1 as an empty set.

[0040] Step 3.2.2.2: Find a 1,1 in the ontology library to get all synonyms of the data item.

[0041] Step 3.2.2.3: Check whether a 1,1 appears in t2, if it appears, add t2 to C 1,1 ;

[0042] Step 3.2.2.4: Check the first synonym of a 1,1 , execute step 3.2.2.3;

[0043] Step 3.2.2.5: Execute step 3.2.2.4 until all synonyms are checked,

[0044] ​Step 3.2.2.6: return to step 3.2.2.2, check the next topic until all topics in T have been checked;

[0045] Step 3.2.2.7: return to step 3.2.2.1, check the next data item until all data items in T have been checked,

[0046] Step 3.2.3: perform relevance enhancement analysis on the collected data using REAABFSDID, the input of REAABFSDID is topic set T, and the output is the set of relevant topics for each data item. REAABFSDID is based on REAAO, and adds a data item dependency table to complete relevance enhancement analysis and results based on data item dependency. The data item dependency is represented by a two-dimensional table, each row includes two fields of data item 1 and data item 2, indicating that data item 2 depends on data item 1, that is, data item 2 is obtained by related data processing based on data item 1, the specific steps are as follows,

[0047] Step 3.2.3.1: initialize parameters deep = 0, tag = deep, get the first data item a of the first topic 1,1 ; initialize C 1,1 as an empty set;

[0048] Step 3.2.3.2: check if a 1,1 appears in t2, if it appears, add <t2, tag> to C 1,1

[0049] Step 3.2.3.3: find a 1,1 as data item 1 in the data item dependency table to obtain dependent data item a 1,1 ', and add it to the data item dependency search queue Q 1,1

[0050] Step 3.2.3.4: repeat step 3.2.3.3 for a 1,1 until the data item dependency table is fully searched.

[0051] Step 3.2.3.5: tag = deep + 1, dequeue elements of Q 1,1 , repeat step 3.2.3.2 until Q 1,1 is empty,

[0052] Step 3.2.3.6: return to step 3.2.3.2, check the next topic until all topics in T have been checked;

[0053] ​​Step 3.2.3.7: Return to step 3.2.3.1 and check the next data item until all data items in T have been checked.

[0054] It should be noted that TRAA, REAAO and REAABFSDID can be used interchangeably and iteratively.

[0055] Among them, step 4: the specific steps are as follows,

[0056] Step 4.1: Based on the received subscribed topics and the known publish-subscribe relationship between subsystems, a global interaction data space is constructed to form a global interaction data sequence Interaction. The publishing topic and subscription topic events are arranged in order from early to late. The sequence consists of a group of events, each of which is represented by a five-tuple, namely (t i ,p / s,s j ,Tvalue,timestamp),i∈I2,j∈I1, where t i Indicates the topic name, p / s is a Boolean variable, indicating whether to publish or subscribe, s j Represents the subsystem name, Tvalue represents the topic value, and timestamp represents the timestamp when the event occurs.

[0057] Step 4.2: Use DESGA to generate data evolution sequence. The input of the algorithm is data item, interaction, and topic association set C. The output is data evolution sequence. Data evolution sequence is a sequence of evolution events and results of a specified data item during a specific operating period of the system. The sequence consists of a set of events, each of which is represented by a five-tuple, namely (t i ,p / s,s j ,Avalue,timestamp), where t i Indicates the topic name, p / s is a Boolean variable, indicating whether to publish or subscribe, s j Indicates the subsystem name, Avalue indicates the data item value, and timestamp indicates the timestamp when the event occurs. For specific steps, see steps 4.2.1 to 4.2.4.

[0058] Step 4.2.1: According to dataitem, find the corresponding subject association set C i,j ;

[0059] Step 4.2.2: If C i,j Not empty, get an element <t v ,tag>. Find the first one containing t in Interaction v Event u, generate (Event u .t i ,Event u .p / s,Event u .s j ,Avalue,Event u .timestamp,tag), where Event u .t i Represents an event u The subject name information contained in Event u .p / s is a Boolean variable representing an event u Contains publishing or subscription information, Event u .s j Represents an event u Contains subsystem name information, Event u .timestamp indicates the event u The timestamp of the occurrence; repeat step 4.2.2 until all events in the interaction have been found, forming a sequence fragment frag1;

[0060] Step 4.2.3: Repeat step 4.2.2 until all sequence fragments are generated;

[0061] Step 4.2.4: Based on the time order and data item dependencies, concatenate all sequence fragments to generate a data evolution sequence.

[0062] Compared with the existing technology, the present invention has the following advantages: 1) This technical solution uses a publish / subscribe mechanism and adopts a non-invasive data acquisition method, which does not require modifying the code of the running system or setting probes, making the data tracing method have the advantages of good flexibility and scalability; 2) This technical solution supports real-time data tracing during system operation, rather than post-analysis, and does not rely on system operation logs; 3) Using the dynamic subscription mechanism of DDS, topics in the running state are dynamically added, and the topic publish-subscribe relationship and data flow path can be updated in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a framework diagram of a data traceability system and method based on DDS;

[0064] Figure 2 This is a flow chart of a data traceability system and method based on DDS;

[0065] Figure 3 This is a scenario diagram of a data traceability system and method based on DDS;

[0066] Figure 4 A global interactive data space graph for a DDS-based data traceability system and method. DETAILED DESCRIPTION

[0067] In order to deepen the understanding of the present application, the present embodiment will be described in detail below with reference to the accompanying drawings.

[0068] Embodiment 1: see Figures 1-4 A DDS-based data traceability system, the system comprising two layers of a monitoring layer and a terminal layer,

[0069] Terminal layer: comprising a subsystem data collection module and a subsystem data stream uploading module, the subsystem data stream uploading module uploading data to the monitoring layer by means of DDS;

[0070] Monitoring layer: comprising a non-intrusive interactive data acquisition module, a topic-to-topic data item correlation analysis module, a topic-based data evolution sequence generation module, and a visualization display module.

[0071] The subsystem data collection module collects data of the subsystem (including business data, and publish / subscribe relationship data (empty if no change occurs));

[0072] The non-intrusive interactive data acquisition module acquires interactive data in a completely non-intrusive manner without modifying the code in the original system or setting probes, by subscribing to all topics in the system when the DDS publish / subscribe mechanism is used for topic-based data distribution. The topic values obtained by subscription are time-stamped when saved,

[0073] The topic-to-topic data item correlation analysis module is responsible for analyzing the correlation between various topic data items, and specifically comprises three sub-modules, namely a topic-based correlation analysis module, a topic-based correlation analysis module, and a data item dependency-based correlation enhancement analysis module,

[0074] The topic-based data evolution sequence generation module generates data evolution sequences using DESGA, in combination with the results of the non-intrusive interactive data acquisition module and the topic-to-topic data item correlation analysis module,

[0075] The visualization display module displays the data evolution sequences in real time, facilitating monitoring and management by operation and maintenance personnel, and realizing real-time traceability of fault data.

[0076] Embodiment 2: see Figures 1-4 A DDS-based data traceability method, the method comprising the following steps:

[0077] Step 1: The subsystem data collection module collects business data, triggers the collection of publish / subscribe relationship data, and sends it to the subsystem data stream upload module.

[0078] The terminal layer is composed of various subsystems, and the data interaction between subsystems is completed with the help of DDS. The subsystem data collection module collects business data and sends it to the subsystem data stream upload module.

[0079] If a new node joins or a node exits, the subsystem data collection module triggers the generation of publish / subscribe relationship data and sends it to the subsystem data stream upload module.

[0080] Step 2: Based on the non-invasive interactive data acquisition module, subscribe to all topics in the system in the subsystem data stream upload module to obtain interactive data and add a timestamp when saving.

[0081] Step 3: The inter-theme data item correlation analysis module first models the traceability system, and then cross-uses TRAA, REAABO and REAABBFTDID to complete the correlation analysis of inter-theme data items.

[0082] Step 4: Construct a global interactive data space based on the topic-based data evolution sequence generation module, and then use DESGA to generate data evolution sequences.

[0083] Step 5: Visualize the data evolution sequence to facilitate real-time monitoring and management by operation and maintenance personnel.

[0084] Step 3 is as follows:

[0085] Step 3.1: Model the traceability system.

[0086] Assume that the system S consists of n subsystems, S={s1,s2,……,s n},n∈N, where s1,s2,……,s n For each subsystem, N is a natural number set, and the index set of each subsystem is I1 = {1, 2, ..., n}. The subsystems interact with each other based on DDS, that is, each subsystem provides or obtains the required data by publishing or subscribing to topics.

[0087] Assume that there are m topics in the system S, and the set of topics T = {t1, t2, ..., t m},m∈N, where t1,t2,……,t m For each topic, let the index set of each topic I2 = {1, 2, ..., m}, each element t in the T set k ,k∈I2 will have a publisher subsystem and a subscriber subsystem, which belongs to the subsystem set S,

[0088] Any one topic is a structure, which is a combination of multiple data items, and has p data items, and let the index set of each data item be I3={1, 2, …, p}, p∈N,

[0089] Let t k ={a k,1 ,a k,2 ,……,a k,p}, k∈I2, where a k,r ,k∈I2, r∈I3 refers to the rth data item in the topic t k ,

[0090] Data provenance is to obtain the evolution sequence of the specified data item a k,r according to the system running process, which is set as seq=<l1, l2, …, l q >, q∈N, where l1, l2, …, l q are each element in the evolution sequence of the data item a k,r ;

[0091] Step 3.2:

[0092] Given the topic publishing and subscribing relationship between the subsystems, the data item correlation analysis between topics is carried out. The data item correlation analysis between topics is to establish the data item-topic set correspondence relationship based on the obtained topic information. This analysis can be specifically divided into topic-based correlation analysis and correlation enhancement analysis. Correlation enhancement analysis is to further improve the data item-topic set relationship through ontology and known data item dependency relationship, so as to mine the implicit correlation relationship. Correlation enhancement analysis is specifically divided into ontology-based correlation enhancement analysis and data item dependency relationship-based correlation enhancement analysis. The specific steps are as follows,

[0093] Step 3.2.1: Use TRAA to analyze the correlation of the collected data. The input of TRAA is the topic set T, and the output is the correlation topic set of each data item. The specific steps are as follows,

[0094] Step 3.2.1.1: Obtain the first data item a 1,1 of the first topic; initialize the correlation topic set C 1,1 of the data item a 1,1 as an empty set.

[0095] Step 3.2.1.2: Check whether a 1,1 appears in t2. If it appears, add t2 to C 1,1 ;

[0096] Step 3.2.1.3: Return to step 3.2.1.2 to check the next topic until all topics in T have been checked.

[0097] Step 3.2.1.4: Return to step 3.2.1.1 and check the next data item until all data items in T have been checked.

[0098] Step 3.2.2: Use REAAO to perform relevance enhancement analysis on the collected data. The input of REAAO is the topic set T, and the output is the topic set associated with each data item. Compared with TRAA, REAAO modifies step 3.2.1.2 and adds an ontology library as input. Each record in the ontology library is represented by a two-tuple, i.e. <data item, synonym set>. The specific steps are as follows:

[0099] Step 3.2.2.1: Get the first data item a of the first subject 1,1 ; Initialize C 1,1 An empty collection.

[0100] Step 3.2.2.2: Find a in the ontology library 1,1 , get all synonyms for the data item.

[0101] Step 3.2.2.3: Check a 1,1 Does it appear in t2? If so, add t2 to C. 1,1 ;

[0102] Step 3.2.2.4: Check a 1,1 For the first synonym of , execute step 3.2.2.3;

[0103] Step 3.2.2.5: Execute step 3.2.2.4 until all synonyms are checked.

[0104] Step 3.2.2.6: Return to step 3.2.2.2 and check the next topic until all topics in T have been checked;

[0105] Step 3.2.2.7: Return to step 3.2.2.1 and check the next data item until all data items in T have been checked.

[0106] Step 3.2.3: For the collected data, correlation enhancement analysis is performed using REAABFSDID, the input of which is the topic set T, and the output is the correlation topic set of each data item. REAABFSDID is based on REAAO, and a data item dependency table is added to complete the correlation enhancement analysis and results based on data item dependency. The data item dependency is represented by a two-dimensional table, each row including two fields of data item 1 and data item 2, indicating that data item 2 depends on data item 1, that is, data item 2 is obtained by related data processing based on data item 1, and the specific steps are as follows,

[0107] Step 3.2.3.1: initialize parameters deep = 0, tag = deep, and get the first data item a of the first topic 1,1 ; initialize C 1,1 to be an empty set;

[0108] Step 3.2.3.2: check whether a 1,1 appears in t2, if it appears, add <t2, tag> to C 1,1 ;

[0109] Step 3.2.3.3: take a 1,1 as data item 1 to search the data item dependency table to obtain dependent data item a 1,1 ', and add it to the data item dependency search queue Q 1,1 ;

[0110] Step 3.2.3.4: repeat step 3.2.3.3 for a 1,1 until the data item dependency table is completely searched.

[0111] Step 3.2.3.5: tag = deep + 1, dequeue the elements of Q 1,1 , and repeat step 3.2.3.2 until Q 1,1 is empty,

[0112] Step 3.2.3.6: return to step 3.2.3.2 to check the next topic until all topics in T have been checked.

[0113] Step 3.2.3.7: return to step 3.2.3.1 to check the next data item until all data items in T have been checked.

[0114] It should be noted that TRAA, REAAO and REAABFSDID can be used iteratively.

[0115] Step 4: The specific steps are as follows,

[0116] Step 4.1: Based on the received subscribed topics, construct the global interaction data space according to the known inter-subsystem publishing-subscribing relationship, as shown in Figure 4 Figure 1. A global interaction data sequence Interaction is formed, in which publishing and subscribing topic events are arranged in chronological order from early to late. The sequence is composed of a set of events, each of which is represented by a five-tuple, i.e., (t i ,p / s,s j ,Tvalue,timestamp), i∈I2, j∈I1, where t i represents the topic name, p / s is a Boolean variable indicating whether it is publishing or subscribing, s j represents the subsystem name, Tvalue represents the topic value, and timestamp represents the time stamp when the event occurs.

[0117] Step 4.2: Generate data evolution sequence using DESGA. The input of the algorithm is dataitem, Interaction, and topic association set C, and the output is data evolution sequence. The data evolution sequence is a sequence of evolution events and results of a specified data item within a specific system runtime period. The sequence is composed of a set of events, each of which is represented by a five-tuple, i.e., (t i ,p / s,s j ,Avalue,timestamp), where t i represents the topic name, p / s is a Boolean variable indicating whether it is publishing or subscribing, s j represents the subsystem name, Avalue represents the data item value, and timestamp represents the time stamp when the event occurs. The specific steps are shown in steps 4.2.1 to 4.2.4,

[0118] Step 4.2.1: According to dataitem, find the corresponding topic association set C i,j ;

[0119] Step 4.2.2: If C i,j is not empty, get an element <t v ,tag>. Find the first event Event v containing t u in Interaction, and generate (Event u .t i ,Event u .p / s,Event u .s j ,Avalue,Event u .timestamp,tag), where Event u .ti Event u Event u Event u Event u Event j Event u Event u Event u timestamp; repeat step 4.2.2 until all events in Interaction have been found, forming a sequence fragment fragl;

[0120] Step 4.2.3: repeat step 4.2.2 until all sequence fragments are generated;

[0121] Step 4.2.4: splice all sequence fragments according to time order and data item dependency, generating a data evolution sequence.

[0122] It should be noted that the above examples are not intended to limit the scope of the present application, and any equivalent transformations or substitutions made on the basis of the above technical solutions fall within the scope of the claims of the present application.

Claims

1. A data tracing method based on DDS, characterized in that: Adopting a data traceability system based on DDS, the system consists of two layers: monitoring layer and terminal layer. Terminal layer: includes subsystem data collection module and subsystem data stream upload module. The subsystem data stream upload module uploads data to the monitoring layer with the help of DDS. Monitoring layer: includes a non-invasive interactive data acquisition module, an inter-theme data item correlation analysis module, a theme-based data evolution sequence generation module, and a visualization display module; The method comprises the following steps: Step 1: The subsystem data collection module collects business data, triggers the collection of publish / subscribe relationship data, and sends it to the subsystem data stream upload module. Step 2: Based on the non-invasive interactive data acquisition module, subscribe to all topics in the system in the subsystem data stream upload module to obtain interactive data and add a timestamp when saving. Step 3: The inter-theme data item correlation analysis module first models the traceability system, and then cross-uses TRAA, REAABO and REAABBFTDID to complete the correlation analysis of inter-theme data items. Step 4: Construct a global interactive data space based on the topic-based data evolution sequence generation module, and then use DESGA to generate data evolution sequences. Step 5: Visualize the data evolution sequence to facilitate real-time monitoring and management by operation and maintenance personnel; Step 3 is as follows: Step 3.1: Model the traceability system. Set up system Depend on subsystems, S={s1,s2,……,s n },n∈N, where s1,s2,……,s n For each subsystem, is a set of natural numbers, and the index set of each subsystem is I1 = {1, 2, ..., n}. The subsystems interact with each other based on DDS, that is, each subsystem provides or obtains the required data by publishing or subscribing to topics. Set up system There are Themes, the theme set T = {t1, t2, ..., t m },m∈N, where t1,t2,……,t m For each topic, let the index set of each topic I2 = {1, 2, ..., m}, Each element in the collection There will be a publisher subsystem and a subscriber subsystem, which belong to the subsystem collection , Any topic is a structure, which is a combination of multiple data items. data items, let the index set of each data item I3={1,2,……,p},p∈N, Assume t k ={a k,1 ,a k,2 ,……,a k,p },k∈I2, where Pointing to the subject Middle data items, Data traceability is to obtain specified data items according to the system operation process The evolution sequence is set as seq=<l1,l2,……,l q >,q∈N, where l1,l2,……,l q For data items the individual elements of the evolutionary sequence; Step 3.2: Assuming the topic publishing and subscription relationship between subsystems is known, the correlation analysis of data items between topics is carried out. The correlation analysis of data items between topics is to establish a data item-topic set correspondence relationship centered on the data item based on the acquired topic information. This analysis is specifically divided into topic-based correlation analysis and correlation enhancement analysis. The correlation enhancement analysis is to further improve the data item-topic set relationship through the ontology and known data item dependencies, thereby mining implicit correlation relationships. The correlation enhancement analysis is specifically divided into ontology-based correlation enhancement analysis and data item dependency-based correlation enhancement analysis. The specific steps are as follows: Step 3.2.1: Perform correlation analysis on the collected data using TRAA. The input of TRAA is the theme set. , the output is a collection of topics associated with each data item. The specific steps are as follows: Step 3.2.1.1: Get the first data item of the first subject ; Initialize data items Related topic collection is an empty collection, Step 3.2.1.2: Check Does it appear in If it appears, join in ; Step 3.2.1.3: Return to step 3.2.1.2 and check the next topic until All the topics in have been checked; Step 3.2.1.4: Return to step 3.2.1.1 and check the next data item until All data items in have been checked. Step 3.2.2: Use REAAO to perform correlation enhancement analysis on the collected data. The input of REAAO is the theme set. The output is the subject set associated with each data item. Compared with TRAA, REAAO modifies step 3.2.1.2 and adds the ontology library as input. Each record in the ontology library is represented by a tuple, i.e. <data item, synonym set>. The specific steps are as follows: Step 3.2.2.1: Get the first data item of the first subject ;initialization is an empty collection, Step 3.2.2.2: Search in the ontology library , get all synonyms of the data item, Step 3.2.2.3: Check Does it appear in If it appears, join in ; Step 3.2.2.4: Check For the first synonym of , execute step 3.2.2.3; Step 3.2.2.5: Execute step 3.2.2.4 until all synonyms are checked. Step 3.2.2.6: Return to step 3.2.2.2 and check the next topic until All the topics in have been checked; Step 3.2.2.7: Return to step 3.2.2.1 and check the next data item until All data items in have been checked. Step 3.2.3: Use REAABFSDID to conduct correlation enhancement analysis on the collected data. The input of REAABFSDID is the theme set. The output is a collection of topics associated with each data item. REAABFSDID adds a data item dependency table based on REAAO to complete the correlation enhancement analysis and results based on data item dependency. The data item dependency is represented by a two-dimensional table. Each row includes two fields, data item 1 and data item 2, indicating that data item 2 depends on data item 1. In other words, data item 2 is the result obtained after relevant data processing based on data item 1. The specific steps are as follows. Step 3.2.3.1: Initialize parameters , , get the first data item of the first topic ;initialization is an empty collection; Step 3.2.3.2: Check Does it appear in If it appears, join in ; Step 3.2.3.3: Put As data item 1, search the data item dependency table and obtain the dependent data items , join the data item dependency search queue ; Step 3.2.3.4: Repeat steps 3.2.3.3 until all data item dependency tables have been searched. Step 3.2.3.5: , out of the team Repeat step 3.2.3.2 until is empty, Step 3.2.3.6: Return to step 3.2.3.2 and check the next topic until All the topics in have been checked; Step 3.2.3.7: Return to step 3.2.3.1 and check the next data item until All data items in have been checked; Among them, step 4: the specific steps are as follows, Step 4.1: Based on the received subscribed topics and the known publish-subscribe relationships between subsystems, a global interaction data space is constructed to form a global interaction data sequence. , the publishing topic and subscription topic events are arranged in order from early to late. The sequence consists of a set of events, each of which is represented by a five-tuple, namely ,in Indicates the subject name, Is a Boolean variable indicating whether to publish or subscribe. Indicates the subsystem name, Represents the subject value, Indicates the timestamp when the event occurs. Step 4.2: Use DESGA to generate data evolution sequence. The input of the algorithm is data item , , subject-related collection , the output is a data evolution sequence, which is a sequence of evolution events and results of a specified data item within a specific operating period of the system. The sequence consists of a set of events, each of which is represented by a five-tuple, namely ,in Indicates the subject name, Is a Boolean variable indicating whether to publish or subscribe. Indicates the subsystem name, Represents the value of a data item, Indicates the timestamp when the event occurs. For specific steps, see steps 4.2.1 to 4.2.

4. Step 4.2.1: According to , find the corresponding subject association set ; Step 4.2.2: If Not empty, get an element ,exist Find the first one containing Events ,generate ,in Representing an event The subject name information contained in A Boolean variable representing an event Contains publish or subscribe information, Representing an event Contains subsystem name information, Representing an event The timestamp of the occurrence; repeat step 4.2.2 until All events in have been found to form a sequence segment ; Step 4.2.3: Repeat step 4.2.2 until all sequence fragments are generated; Step 4.2.4: Based on the time order and data item dependencies, concatenate all sequence fragments to generate a data evolution sequence.

2. The data tracing method based on DDS according to claim 1, characterized in that: The subsystem data collection module collects subsystem data including business data and publish / subscribe relationship data.

3. The data tracing method based on DDS according to claim 2 is characterized in that: When the interactive data acquisition module uses the DDS publish / subscribe mechanism to distribute data based on topics, there is no need to modify the code in the original system or set up probes. It only needs to subscribe to all topics in the system in the interactive data acquisition module. Therefore, this is a completely non-invasive way to obtain interactive data. The subscribed topic value is timestamped when it is saved. The inter-theme data item correlation analysis module is responsible for analyzing the correlation relationship between each theme data item. The data evolution sequence generation module based on the theme is combined with the non-intrusive interactive data acquisition module and the results of the inter-theme data item correlation analysis module to generate the data evolution sequence using DESGA. The visualization module displays the data evolution sequence in real time, facilitating the monitoring and management of operation and maintenance personnel and enabling real-time tracing of fault data.

Images