Cold and hot data hierarchical processing method and system for industrial internet

Abstract

The invention discloses a hot and cold data hierarchical processing method and system for an industrial internet, and relates to the technical field of industrial internet data processing, and the method comprises the steps: constructing a first-stage hot library through a memory library meeting a first preset access speed, and constructing a second-stage hot library through a time sequence library meeting a second preset access speed, the refrigeration house is constructed through the time sequence library meeting the preset capacity; splitting the industrial internet data according to a preset particle dynamic segmentation strategy to obtain particle segmentation data; judging the particle segmentation data as to-be-reheated data or to-be-cooled data according to a preset access frequency threshold value and the particle segmentation data access frequency; and according to the to-be-reheated data and the to-be-cooled data, data migration in the first-stage heat reservoir, the second-stage heat reservoir and the cold storage is carried out. The method has the beneficial effects that the dynamic cold and hot bidirectional switching of the particle segmentation data is realized, the storage position of the cold and hot data is ensured to adapt to the access requirement of a user, and meanwhile, the data processing efficiency of the industrial internet is ensured without complicated calculation.

Description

Technical Field

[0001] This application relates to the field of industrial internet data processing technology, and in particular to a method and system for layered processing of hot and cold data for industrial internet. Background Technology

[0002] Industrial Internet platforms collect time-series data through data collectors, which is typically stored in time-series databases such as InfluxDB. These databases generally use disk-based storage, meaning data is written to the hard drive immediately after being entered. Therefore, both read and write operations consume system I / O resources. In the context of industrial Internet platforms, systems are typically deployed privately, unlike cloud storage, where I / O resources are limited. Therefore, it is necessary to optimize the I / O pressure during the time-series data storage and retrieval process.

[0003] When time-series databases directly store data, the compression, database sharding, and disk persistence processes result in relatively long storage and retrieval times and higher computational costs. Currently, common methods to improve data access efficiency include adding data indexes and optimizing data sharding.

[0004] The patent "A Distributed Time-Series Database Data Management Method Supporting Hierarchical Management of Hot and Cold Data", publication number CN121070921A, application date: August 20, 2025, specifically discloses a method including: receiving write requests for time-series data and performing preliminary data aggregation and sorting; dynamically identifying hot and cold attributes of data based on a multi-dimensional data hot and coldness calculation model, combined with the write and access behaviors of time-series data; automatically storing time-series data in a hierarchical manner based on the hot and cold attribute identification results and adaptively partitioned hierarchical boundaries; dynamically adjusting the storage hierarchy of time-series data and performing data migration tasks based on a dynamic migration and scheduling mechanism for hot and cold data, according to the access frequency and time decay characteristics of time-series data; and constructing a hierarchical index system adapted to time-series data with different hot and cold attributes, combined with dynamic identification of query frequency and automatic upgrading of the index structure. However, this solution is based entirely on a time-series database, the access speed of hot data is limited by the time-series database architecture, and the hierarchical arrangement depends on the dynamic adjustment of heat scores, requiring a large amount of computing resources for weight optimization, resulting in low computational efficiency.

[0005] The patent, "A Smart Hierarchical Caching Method and System Based on Hyperconverged Architecture," publication number CN120705077A, application date August 27, 2025, specifically discloses a method including converged architecture data acquisition, initial cache hierarchy, three-way load prediction, cache resource scheduling, and smart hierarchical caching. This solution manages data in fixed blocks, but the mixing of low-frequency sub-data into high-frequency data blocks leads to ineffective use of memory resources and also requires significant computing resources, resulting in low computational efficiency. Summary of the Invention

[0006] This application addresses the problems of existing technologies in industrial internet data processing, such as the inability to achieve bidirectional switching between hot and cold data and low processing efficiency. It provides a method and system for layered processing of hot and cold data for industrial internet. By constructing a primary hot storage, a secondary hot storage, and a cold storage, and combining the access frequency calculation of granular segmented data, it realizes multi-layered hot and cold data reheating and cooling judgment for corresponding data blocks, achieving dynamic bidirectional switching between hot and cold data of granular segmented data. This ensures that the storage location of hot and cold data adapts to user access needs, while ensuring the data processing efficiency of industrial internet without complex calculations.

[0007] To achieve the above technical objectives, this application provides a technical solution for layered processing of hot and cold data in the Industrial Internet, comprising the following steps: constructing a primary hot storage using a memory library that meets a first preset access speed, constructing a secondary hot storage using a time-series library that meets a second preset access speed, and constructing a cold storage using a time-series library that meets a preset capacity; wherein the first preset access speed is greater than the second preset access speed; in response to the storage of Industrial Internet data, pre-storing it in the primary hot storage, splitting the Industrial Internet data according to a preset granular dynamic segmentation strategy to obtain granular segmented data; periodically acquiring the access frequency of granular segmented data in the primary hot storage, secondary hot storage, and cold storage, and determining the granular segmented data as data to be reheated or data to be cooled based on a preset access frequency threshold and the granular segmented data access frequency; and migrating data in the primary hot storage, secondary hot storage, and cold storage based on the data to be reheated and the data to be cooled.

[0008] Furthermore, the construction of a secondary hot library for a time-series library to meet the second preset access speed includes: constructing a secondary hot library using a time-series library mounted on an SSD.

[0009] Furthermore, the construction of a cold storage for a time-series library to meet a preset capacity includes: constructing a cold storage for a time-series library mounted on a mechanical hard drive.

[0010] Furthermore, the step of building a first-level hot library using a memory library to meet the first preset access speed includes: building a first-level hot library using a Redis memory library.

[0011] Furthermore, the step of splitting industrial internet data according to a preset dynamic particle segmentation strategy to obtain particle segmentation data includes: using the divisible time period corresponding to a single identifier of industrial internet data as the particle segmentation segment corresponding to that identifier, and obtaining particle segmentation data.

[0012] Furthermore, the step of periodically acquiring the access frequency of particle segmentation data in the primary hot storage, secondary hot storage, and cold storage includes: constructing a time-series particle access statistics table corresponding to the tag number; performing particle segmentation data access statistics in real time according to the preset particle statistics time corresponding to the primary hot storage, secondary hot storage, and cold storage, and recording it in the time-series particle access statistics table; and periodically calculating the particle segmentation data access frequency based on the time-series particle access statistics table.

[0013] Furthermore, the step of periodically calculating the access frequency of granular segmentation data based on the time-series granular access statistics table includes periodically calculating the access frequency of granular segmentation data based on the start time, recent access time, and access count of the granular segmentation data in the time-series granular access statistics table.

[0014] Furthermore, the step of determining whether particle-segmented data is data to be reheated or data to be cooled based on the preset access frequency threshold and the particle-segmented data access frequency includes: if the particle-segmented data access frequency is greater than the maximum preset access frequency threshold, it is determined to be data to be reheated; if the particle-segmented data access frequency is less than the minimum preset access frequency threshold, it is determined to be data to be cooled.

[0015] Furthermore, the real-time execution of particle segmentation data access and statistics based on the preset particle statistics time corresponding to the primary hot storage, secondary hot storage, and cold storage also includes: setting a first preset particle statistics time corresponding to the primary hot storage, a second preset particle statistics time corresponding to the secondary hot storage, and a third preset particle statistics time corresponding to the cold storage; wherein, the first preset particle statistics time is less than the second preset particle statistics time, and the second preset particle statistics time is less than the third preset particle statistics time.

[0016] Another technical solution provided in this application is a hot and cold data tiered processing system for the Industrial Internet, used to implement the method described above, including: a primary hot storage, constructed from a memory library that meets a first preset access speed, which, in response to the storage of Industrial Internet data, splits the Industrial Internet data according to a preset dynamic granularity segmentation strategy to obtain granular segmented data; a secondary hot storage, constructed from a time-series library that meets a second preset access speed; a cold storage, constructed from a time-series library that meets a preset capacity; and a hot and cold data switching management module, used to periodically obtain the access frequency of granular segmented data in the primary hot storage, secondary hot storage, and cold storage, determine the granular segmented data as data to be reheated or data to be cooled according to a preset access frequency threshold, and perform data migration in the primary hot storage, secondary hot storage, and cold storage based on the data to be reheated and the data to be cooled.

[0017] The beneficial effects of this application are as follows: 1. By setting up primary hot storage, secondary hot storage, and cold storage separately, industrial internet data can be stored in layers according to different access needs. At the same time, the stored data is granularly segmented using a preset dynamic granular segmentation strategy, and user access behavior is broken down into granular levels. Based on the access frequency of the granular segmented data, the corresponding data blocks are judged for multi-layer hot and cold data reheating and cooling, realizing dynamic bidirectional hot and cold switching of granular segmented data. This ensures that the storage location of hot and cold data adapts to user access needs, improves the flexibility and adaptability of the storage system, and ensures the data processing efficiency of the industrial internet without complex calculations.

[0018] 2. Implement multi-tiered hierarchical storage for hot and cold data, storing less frequently accessed hot data in a faster database and more frequently accessed cold data in a slower database. Based on the time-series storage characteristics of the Industrial Internet, orthogonal tagging, and specific durations, time-series data is granularly segmented. Simultaneously, the size of the segmented time-series data granules is dynamically adjusted according to these characteristics, allowing user access behavior to be broken down into granular levels. Access frequency of different time-series data granules is statistically analyzed based on user access behavior. Based on the time-series data granule statistics table, query behavior is distributed and aggregated across multiple time-series databases. Utilizing the access frequency of time-series data granules, corresponding data blocks are classified as hot or cold data at multiple levels. This enables dynamic, bidirectional switching between hot and cold data based on access frequency, and improves migration efficiency through inter-level transmission at the segmented time-series data granule level. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the cold and hot data stratification processing method for the Industrial Internet used in this application.

[0020] Figure 2 This is a schematic diagram of the migration process of the cold and hot data layering processing method for the Industrial Internet used in this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] like Figure 1 As shown in the first embodiment of this application, a method for layering hot and cold data in the industrial internet includes the following steps: A primary hot library is constructed using a memory library that meets a first preset access speed, a secondary hot library is constructed using a time-series library that meets a second preset access speed, and a cold library is constructed using a time-series library that meets a preset capacity; wherein, the first preset access speed is greater than the second preset access speed. In response to the storage of industrial internet data, the data is pre-stored in the primary hot database and then split into granular segments according to a preset dynamic granular segmentation strategy to obtain granular segmentation data. Regularly obtain the access frequency of particle cutting data in primary heat storage, secondary heat storage and cold storage, and determine the particle cutting data as data to be reheated or data to be cooled based on the preset access frequency threshold and particle cutting data access frequency. Data migration is performed in the primary heat storage, secondary heat storage, and cold storage based on the data to be reheated and the data to be cooled.

[0023] In this embodiment, by setting up primary, secondary, and cold storage facilities, industrial internet data is stored in layers according to different access needs. At the same time, a preset dynamic granular segmentation strategy is used to segment the stored data into granular levels, breaking down user access behavior into granular levels. Based on the access frequency of the segmented data, multi-layer hot and cold data reheating and cooling judgment is implemented for the corresponding data blocks, realizing dynamic bidirectional switching between hot and cold data of the segmented data. This ensures that the storage location of hot and cold data adapts to user access needs, improves the flexibility and adaptability of the storage system, and ensures the data processing efficiency of the industrial internet without complex calculations.

[0024] Specifically, the first-level hot library for building a memory library to meet the first preset access speed includes: Build a first-level hot library using Redis in-memory library.

[0025] Redis uses direct in-memory read / write, eliminating disk I / O overhead, and its access speed meets the needs of fast access to frequently accessed data. Storing frequently accessed data in the first-level hot database minimizes the amount of data stored and maximizes access speed.

[0026] The construction of a secondary hot library for a time-series library to meet the second preset access speed includes: A secondary hot library is built using a timing library mounted on an SSD.

[0027] Leveraging the high-speed I / O capabilities of SSDs (Solid State Drives), the timing data in the timing library is adapted, with SSDs, balancing access speed and storage capacity, serving as a transitional layer to meet the needs of mid-frequency data. The secondary hot storage stores the cooled data from the primary hot storage. The amount of data stored in the secondary hot storage is greater than that in the primary hot storage but less than that in the cold storage, and the access speed is greater than that in the cold storage but less than that in the primary hot storage.

[0028] The construction of a cold storage facility to meet the preset capacity of a time-series library includes: A cold storage is constructed using a mechanical hard drive as the mounted time-series library.

[0029] The large capacity of hard disk drives (HDDs) meets the storage needs for long-term archiving of cold data, and the time-series library is mounted independently on the HDDs, requiring no additional resources. The cold storage, as the final data storage location, has the slowest access speed and the largest data volume.

[0030] In this embodiment, a combined architecture is used to construct a primary hot database using a Redis in-memory library, a secondary hot database using an SSD-mounted time-series library, and a cold database using a mechanical hard drive. This achieves the technical effects of extremely fast response to hot data, balanced adaptation of mid-frequency data, and low-cost archiving of cold data. Furthermore, the data formats of Redis, the SSD time-series library, and the mechanical hard drive time-series library are compatible, eliminating the need for complex format conversions or index adjustments during the hot data cooling migration and cold data reheating migration processes, thus improving migration efficiency.

[0031] In the Industrial Internet, time-series data is characterized by high write frequency and strong time-series correlation. However, due to susceptibility to network fluctuations and equipment failures, the data is prone to out-of-order processing. Therefore, this embodiment utilizes a Redis in-memory database to construct a primary hot database based on the characteristics of Industrial Internet data. This adapts to the high write frequency and strong time-series correlation of the Industrial Internet. Furthermore, Redis in-memory database uses a key-value storage model and does not have strict sequential dependencies, thus preserving the existing data storage order. The time-sharded storage of the time-series database allows out-of-order data to be written to the corresponding time shards. Even if data becomes out of order, the original time shard can be located using the timestamp, without affecting the overall time order of the data block. This achieves targeted storage and processing of Industrial Internet data, improving the adaptability of Industrial Internet applications.

[0032] In this embodiment, when industrial internet data is stored, it is first stored in a primary hot database. The industrial internet data is then split according to a preset dynamic granularity partitioning strategy to obtain granular partitioned data. The stored time-series data is initially managed as hot data. Redis's high write efficiency is used to achieve rapid writing of the overall data. Then, the industrial internet data is split according to the preset dynamic granularity partitioning strategy to obtain granular partitioned data. Since the granular partitioned data is used for periodic access frequency calculations, it can be executed asynchronously with the writing phase. By first storing the original data completely in the primary hot database and then asynchronously executing the partitioning operation, the risk of data loss due to anomalies during the partitioning process is avoided. Redis's multi-threading capabilities ensure real-time data writing efficiency and partitioning integrity.

[0033] Specifically, industrial internet data is split according to a preset dynamic granularity segmentation strategy to obtain granular segmented data, including: The particle segmentation data is obtained by using the divisible time period corresponding to a single tag in the industrial internet data as the particle segmentation segment corresponding to that tag.

[0034] The tag number is the smallest category of collected data in the database, representing a time-distributed dataset uploaded by a data acquisition unit, such as temperature information transmitted by a sensor over a certain period. Data is segmented into granular parts according to individual tag numbers to avoid mixing data with different tag numbers and to ensure the accuracy of industrial internet data segmentation.

[0035] In the data storage process, data is segmented into granular units of x hours for each individual data point. The starting point for segmentation is a timestamp that is divisible by the number of seconds corresponding to x. For example, 1758610800 (2025-09-23 15:00:00) is divisible by 1 hour, so this time point is considered a valid starting point for granular segmentation (corresponding to x=1). The setting of the starting point is to conform to the general syntax rules of time series database queries in order to improve the efficiency of time series database retrieval.

[0036] In this embodiment, the frequency of periodically acquiring particle segmentation data from primary heat storage, secondary heat storage, and cold storage includes: A time-series particle access statistics table is constructed corresponding to the tag number. Based on the preset particle statistics time corresponding to the primary hot storage, secondary hot storage and cold storage, the particle segmentation data access statistics are performed in real time and recorded in the time-series particle access statistics table. Periodically calculate the access frequency of granular segmentation data based on the time-series granular access statistics table.

[0037] As shown in Table 1, a time-series granular access statistics table is constructed for each bit number.

[0038] Table 1. Statistics on Time-Sequence Granularity Access

[0039] Specifically, based on the preset particle statistical time corresponding to the primary, secondary, and cold storage facilities, the particle segmentation data access statistics are performed in real time and recorded in the time-series particle access statistics table, including: In response to a data request, update the last access time and access count of the granular segmentation data in the time-series granular access statistics table according to the data request.

[0040] In this embodiment, the time-series granular access statistics table records a single data request, i.e., a single query (with conditions including the query bit number and query time) that covers the bit number and time period of the granular segmented data. The query time is recorded, and the access count is increased. This signifies the access status of data within the granular segment by external queries.

[0041] In response to a data request, based on the coverage bit and time period specified in the query statement, the system checks the Level 3 time-series granularity access statistics table to see if the corresponding hot data granules are covered. For granular rows that are covered, the data is extracted from the corresponding level library.

[0042] The time-series granular access statistics table records the start time, most recent access time, and number of accesses for the granular data. The start time of the granular data is the initial time of the granular data segmentation, which can be used as the basis for calculations and also serves as the identifier for the granular data segmentation. The most recent access time indicates the time when the granular data segmentation was last accessed. The number of accesses indicates the total number of times the granular data segmentation was accessed.

[0043] Periodically calculate the access frequency of granular segmentation data based on the time-series granular access statistics table, including: Periodically calculate the access frequency of granular segmentation data based on the start time, recent access time, and access count of the granular segmentation data in the time-series granular access statistics table.

[0044] In this embodiment, the periodic time period can be set according to the switching requirements of hot and cold data.

[0045] Based on the start time, recent access time, and access frequency of the granular segmentation data in the time-series granular access statistics table, the granular segmentation data access frequency is calculated as follows: ; ; in, Here, m represents the access frequency of the granular data segmentation, and m represents the number of times the granular data segmentation is accessed. This indicates the current time during calculation, which is the starting time of the particle segmentation data. Cooling time ( ) indicates the recent visit time ( Distance calculation is performed at the current time. (The duration of the duration).

[0046] Furthermore, based on the preset access frequency threshold and the access frequency of the particle segmentation data, the particle segmentation data is determined to be either data to be reheated or data to be cooled, including: If the access frequency of the particle segmentation data is greater than the maximum preset access frequency threshold, it is determined to be data to be reheated. If the access frequency of the particle segmentation data is less than the minimum preset access frequency threshold, it is determined to be data to be cooled.

[0047] In this embodiment, the preset access frequency threshold is a threshold range, and the range exists. When the frequency of accessing granular segmented data is less than Data deemed to require cooling is identified when the frequency of particle segmentation data access is greater than [a certain value]. This data is identified as data awaiting reheating. The preset access frequency threshold can be set according to data retrieval needs.

[0048] The process of real-time access to and statistics of particle segmentation data based on the preset particle statistics time corresponding to primary, secondary, and cold storage facilities also includes: Set a first preset particle counting time for a first-level hot storage, a second preset particle counting time for a second-level hot storage, and a third preset particle counting time for a cold storage; wherein the first preset particle counting time is less than the second preset particle counting time, and the second preset particle counting time is less than the third preset particle counting time.

[0049] In this embodiment, the time-series granular access statistics tables in different databases have different access frequency requirements. The time-series granular access statistics table in the primary hot storage has a higher frequency requirement than that in the secondary hot storage, while the time-series granular access statistics table in the cold storage has the lowest requirement, thus limiting the amount of primary hot data. This tiered variation in statistical time ensures that the statistical precision matches the access requirements of the storage levels, avoiding resource waste caused by fine-grained statistics across all levels.

[0050] Among them, the first preset particle counting time, the second preset particle counting time, and the third preset particle counting time increase exponentially in sequence.

[0051] like Figure 2 As shown, the data migration between the primary heat storage, secondary heat storage, and cold storage based on the data to be reheated and the data to be cooled includes: Migrate the data to be reheated to the upper-level database, migrate the data to be cooled to the lower-level database, and simultaneously adjust the time-series particle access statistics table corresponding to the data.

[0052] Understandably, if the access frequency of granular data in the primary hot storage is greater than the maximum preset access frequency threshold, migration will not occur; similarly, if the access frequency of granular data in the cold storage is less than the minimum preset access frequency threshold, migration will also not occur. In some cases, it may be possible to determine whether the access frequency of granular data in the primary hot storage is less than the minimum preset access frequency threshold, and whether the access frequency of granular data in the cold storage is greater than the maximum preset access frequency threshold, thereby reducing the number of determinations.

[0053] In this embodiment, hot and cold data are stored in a multi-level hierarchical manner, with less hot data stored in a faster database and more cold data stored in a slower database. Based on the time-series storage characteristics of the Industrial Internet, orthogonal identifiers, and specific durations, the time-series data is granularly segmented. Simultaneously, the size of the segmented time-series data granules is dynamically adjusted according to these characteristics, allowing user access behavior to be broken down into granular levels. The access frequency of different time-series data granules is statistically analyzed based on user access behavior. The query behavior is then distributed and aggregated across multiple time-series databases based on the time-series data granule statistics table. By utilizing the access frequency of the time-series data granules, corresponding data blocks are classified as hot or cold data at multiple levels. This enables dynamic, bidirectional switching between hot and cold data based on access frequency, and improves migration efficiency through inter-level transmission at the segmented time-series data granule level.

[0054] As a second embodiment of this application, a cold and hot data tiered processing system for the Industrial Internet includes: The first-level hot library is built from a memory library that meets the first preset access speed. In response to the storage of industrial Internet data, it splits the industrial Internet data according to the preset dynamic granular segmentation strategy and obtains granular segmented data. The secondary hot library is constructed from a time-series library that meets the second preset access speed; The cold storage is constructed from time-series storage that meets the preset capacity. The hot and cold data switching management module is used to periodically obtain the access frequency of particle segmentation data in the primary hot storage, secondary hot storage and cold storage, determine the particle segmentation data as data to be reheated or data to be cooled according to the preset access frequency threshold, and perform data migration in the primary hot storage, secondary hot storage and cold storage according to the data to be reheated and data to be cooled.

[0055] In this embodiment, the cold and hot data switching management module is responsible for storing the corresponding interval data in the hot storage into the secondary database in whole particles, and moving the corresponding row in the time-series particle access statistics table to the next level statistics table for the data to be cooled; the cold and hot data switching management module is responsible for storing the corresponding interval data in the secondary cold storage into the upper level database in whole particles, and moving the corresponding row in the time-series particle access statistics table back to the upper level statistics table for the data to be reheated.

[0056] The specific embodiments described above are preferred embodiments of the cold and hot data layering processing method and system for the Industrial Internet of Things, and are not intended to limit the specific scope of this application. The scope of this application includes but is not limited to these specific embodiments. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A method for layered processing of hot and cold data for the Industrial Internet, characterized in that: Includes the following steps: A primary hot library is constructed using a memory library that meets a first preset access speed, a secondary hot library is constructed using a time-series library that meets a second preset access speed, and a cold library is constructed using a time-series library that meets a preset capacity; wherein, the first preset access speed is greater than the second preset access speed. In response to the storage of industrial internet data, the data is pre-stored in the primary hot database and then split into granular segments according to a preset dynamic granular segmentation strategy to obtain granular segmentation data. Regularly obtain the access frequency of particle cutting data in primary heat storage, secondary heat storage and cold storage, and determine the particle cutting data as data to be reheated or data to be cooled based on the preset access frequency threshold and particle cutting data access frequency. Data migration is performed in the primary heat storage, secondary heat storage, and cold storage based on the data to be reheated and the data to be cooled.

2. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 1, characterized in that: The construction of a secondary hot library based on a time-series library to meet a second preset access speed includes: A secondary hot library is built using a timing library mounted on an SSD.

3. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 2, characterized in that: The construction of the cold storage facility to meet the preset capacity includes: A cold storage is constructed using a mechanical hard drive as the mounted time-series library.

4. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 3, characterized in that: The construction of a first-level hot library for the memory library to meet the first preset access speed includes: Build a first-level hot library using Redis in-memory library.

5. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 1, characterized in that: The step of splitting industrial internet data according to a preset dynamic granular segmentation strategy to obtain granular segmentation data includes: The particle segmentation data is obtained by using the divisible time period corresponding to a single tag in the industrial internet data as the particle segmentation segment corresponding to that tag.

6. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 1, characterized in that: The frequency of periodically acquiring granular segmented data from primary, secondary, and cold storage facilities includes: A time-series particle access statistics table is constructed based on the corresponding tag number. Particle segmentation data access statistics are performed in real time according to the preset particle statistics time corresponding to the first-level hot storage, second-level hot storage and cold storage, and recorded in the time-series particle access statistics table. Periodically calculate the access frequency of granular segmentation data based on the time-series granular access statistics table.

7. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 6, characterized in that: The periodic calculation of the access frequency of granular segmentation data based on the time-series granular access statistics table includes: Periodically calculate the access frequency of granular segmentation data based on the start time, recent access time, and access count of the granular segmentation data in the time-series granular access statistics table.

8. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 1, characterized in that: The step of determining whether particle-segmented data is to be reheated or cooled based on a preset access frequency threshold and the access frequency of particle-segmented data includes: If the access frequency of the particle segmentation data is greater than the maximum preset access frequency threshold, it is determined to be data to be reheated. If the access frequency of the particle segmentation data is less than the minimum preset access frequency threshold, it is determined to be data to be cooled.

9. The method for layered processing of hot and cold data for the Industrial Internet as described in claim 6, characterized in that: The real-time execution of particle segmentation data access and statistics based on the preset particle statistics time corresponding to the primary heat storage, secondary heat storage, and cold storage also includes: Set a first preset particle counting time for a first-level hot storage, a second preset particle counting time for a second-level hot storage, and a third preset particle counting time for a cold storage; wherein the first preset particle counting time is less than the second preset particle counting time, and the second preset particle counting time is less than the third preset particle counting time.

10. A cold and hot data tiered processing system for the Industrial Internet, used to implement the method as described in any one of claims 1 to 9, characterized in that: include: The first-level hot library is built from a memory library that meets the first preset access speed. In response to the storage of industrial Internet data, it splits the industrial Internet data according to the preset dynamic granular segmentation strategy and obtains granular segmented data. The secondary hot library is constructed from a time-series library that meets the second preset access speed; The cold storage is constructed from time-series storage that meets the preset capacity. The hot and cold data switching management module is used to periodically obtain the access frequency of particle segmentation data in the primary hot storage, secondary hot storage and cold storage, determine the particle segmentation data as data to be reheated or data to be cooled according to the preset access frequency threshold, and perform data migration in the primary hot storage, secondary hot storage and cold storage according to the data to be reheated and data to be cooled.

Images