Fusion system for time-series database and conventional database

Abstract

The present invention relates to the technical field of database fusion. Disclosed in the present invention is a fusion system for a time-series database and a conventional database, which fusion system solves the problems of it being impossible to achieve a good database fusion processing effect, and during an actual fusion process, data being lost during fusion due to the load of a system. In the present invention, write volume data monitored in real time and query volume data are monitored, and on the basis of related parameters monitored in real time, computing power allocation parameters corresponding thereto are identified; and associative allocation is performed on computing power resources, such that corresponding partitions can be in an optimal numerical operation state by means of a specific allocation process, and a better operation standard can be achieved so as to fully ensure the overall storage performance and the read-write performance of corresponding databases which have been subjected to fusion, thereby achieving a good fusion processing effect.

Description

Fusion system of time series database and traditional database Technical Field

[0001] The present invention relates to the technical field of database fusion, and in particular to a fusion system of a time series database and a traditional database. Background Art

[0002] Database convergence is a key development trend in modern data processing, aiming to integrate multiple types of databases, including transactional (OLTP), analytical (OLAP), and big data or data lakes, into a unified platform;

[0003] The application with publication number CN114547168A discloses a method for the fusion and reconstruction of fine chemical production safety data based on a virtual knowledge graph. In view of the characteristics of fine chemical production safety data, namely, a large amount of structured data, multi-source heterogeneous databases, and strong temporal logic, the present invention innovatively proposes the use of a virtual knowledge graph method to complete the fusion and reconstruction of traditional fine chemical databases. Specifically, on the basis of not increasing the storage scale of the original database, the database is reconstructed from a perspective closer to human logic, thereby independently establishing the logical model and storage method of the underlying database, and accessing multi-source databases more conveniently and clearly. For the first time, the present invention integrates the static structured knowledge in the field of fine chemicals with the real-time dynamic database of chemical production safety in the concept of ontology, and organizes time series data in the form of entities. And the mapping rules of the existing OBDA system are improved based on the data set of the present invention.

[0004] In the specific fusion process of its time series database and traditional database, it is generally necessary to fuse and store related data based on the corresponding data features. However, in the actual fusion process, the specific parameter features used by the corresponding database are inconsistent with the parameter features of the original different databases. This will lead to the overall performance of the fused database being worse than that of the previous database in the actual application fusion process, and it cannot achieve a good database fusion processing effect. In addition, in the actual fusion process, data is lost during fusion due to the system load. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a fusion system for time series databases and traditional databases, which solves the problem that a good database fusion processing effect cannot be achieved and that data is lost during fusion due to system load during the actual fusion process.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a fusion system of a time series database and a traditional database, comprising:

[0007] The associated data query end obtains data related to the time-per-unit-time write volume of the time series database to be integrated and the data related to the time-per-unit-time query volume of the traditional database from the historical completed data. The obtained related data is periodic data, and the period of the periodic data is the preset period.

[0008] The computing resource allocation processing end prioritizes determining the standard write volume of the time series database and the standard query volume of the traditional database from the relevant data with correlation in the query. It then simulates the correlation of read, write, and query through the sharding simulation end, and allocates computing resources in a correlated manner, ensuring that the different computing resources corresponding to different shards are in the optimal state. The specific method is as follows:

[0009] The maximum write volume per unit time of the periodic data corresponding to the time series database is used as the standard write volume of the time series database. The maximum query volume of the periodic data of the traditional database is used as the standard query volume of the corresponding traditional database.

[0010] The database sharding simulation end simulates two groups of storage partitions, and marks one group of storage partitions as time series partitions and the other group of storage partitions as traditional partitions. Time series partitions correspond to the database sharding characteristics of time series databases, and traditional partitions correspond to the database sharding characteristics of traditional databases. Then, the computing power resource allocation processing end associates and allocates the original computing power resources, giving priority to evenly allocating the specified computing power resources to the two groups of storage partitions, and records the write volume of the time series partitions in real time. k And the query volume CX of traditional partitions k , where k represents different moments, the amount written is XR k and CX k Compare with the standard write amount or standard query amount, calibrate the standard write amount as BR, and the standard query amount as BC:

[0011] If XR k <BR and CX k When <BC, a signal indicating insufficient computing power allocation is generated and displayed on the external display terminal;

[0012] When XR k BR and CX k <BC time: allocate computing resources of time series partition to traditional partition, to XR k =BR or CX k =BC stops, records the computing power resources allocated at the current moment, uses them as the execution allocation computing power, and transmits them to the write execution end;

[0013] When XR k <BR and CX k >BC time: allocate computing resources of traditional partitions to time partitions, to XR k =BR or CXk =BC stops, records the computing power resources allocated at the current moment, uses them as the execution allocation computing power, and transmits them to the write execution end;

[0014] When there is XR k =BR or CX k =BC, no processing is required, and the computing power resources allocated at the current moment are recorded as the execution allocation computing power and transmitted to the write execution end;

[0015] The write execution end allocates computing power to the designated partitions based on the execution power allocated to the corresponding partitions. In the actual fusion system, the corresponding computing power resources are allocated to the designated partitions for execution. The write execution end includes a write debugging unit, a load monitoring unit, and a monitoring, analysis, and control unit.

[0016] The write debugging unit performs write debugging for the database integration process. It determines a set of debugging cycles, adjusts the write volume of different partitions within these debugging cycles, records the database load, and locks the associated change intervals of different partitions based on the value changes of different partitions. The specific method is as follows:

[0017] Determine a set of debugging cycles t, where t is a preset value, and divide this test cycle t into two fine-tuning cycles t1 and t2;

[0018] During the fine-tuning period t1, the write volume XL of the associated partition storing time series data is gradually increased, and the numerical change of its system load rate is recorded. The write volume of another partition remains unchanged. Based on the increase value Z1 of the write volume XL per unit time and the increase value Z2 of the system load rate, the associated parameter corresponding to the unit time is determined. This associated parameter = Z2 ÷ Z1. Based on the several sets of associated parameters determined during the fine-tuning period t1 and the minimum and maximum values ​​of the associated parameters, the associated change interval of the associated partition storing time series data is determined.

[0019] During the fine-tuning period t2, the write volume of the associated partition storing traditional data is gradually increased, and the associated change interval of the associated partition storing traditional data is locked by recording the numerical change of its system load rate and keeping the write volume of another partition unchanged;

[0020] Transmitting the determined two sets of associated change intervals to the monitoring, analysis and control unit;

[0021] The load monitoring unit monitors the system load rate in real time during the database fusion process and transmits the real-time monitored system load rate to the monitoring, analysis and control unit in real time;

[0022] The monitoring, analysis, and control unit adjusts the write volume of different partitions in real time based on the system load rate monitored in real time at the current moment, ensuring that the system is not under load while still meeting the maximum write volume. The specific methods are as follows:

[0023] Based on the system load rate Fz monitored at the current moment, the standard check value Hz associated with the system load rate is extracted, where Hz is a preset standard parameter;

[0024] If Fz>Hz, the write amount of different partitions will be reduced in real time until Fz≤Hz;

[0025] If Fz=Hz, the write volume of the original corresponding partition remains unchanged, and its system load rate is monitored in real time;

[0026] If Fz < Hz, then determine the system load rate difference: Cz = Hz - Fz, and then based on the different associated change intervals determined by different partitions [GL1 q , GL2 q ]Add the write amount of different partitions, where q represents different partitions.

[0027] Preferably, when Fz>Hz, the specific method of reducing the writing amount of different partitions in real time is:

[0028] Record the newly added write amount from the previous moment to the current moment of the corresponding partition, and mark the different newly added write amounts of different partitions as Xq, where q represents different partitions, and then determine the associated change interval based on the corresponding partition [GL1 q , GL2 q ] to determine the system load change value that can be generated when the corresponding partition write amount changes: GL1 q ×Xq=BHmin and GL2 q ×Xq=BHmax, to determine the system load change value range [BHmin, BHmax] that can be generated by the corresponding partition;

[0029] Then, the middle value Zq of the interval is locked from the determined load change value interval, and the middle values ​​Zq of the two intervals are ratio-processed to determine their ratio sequence. Then, based on this ratio sequence, the write amount of different partitions is reduced, and the specific ratio of the specific reduced values ​​is consistent with this ratio sequence.

[0030] Preferably, when Fz < Hz, the specific method of increasing the write amount of different partitions is:

[0031] Use (Cz÷2)÷GL1 q =XZqmax and (Cz÷2)÷GL2 q=XZqmin, obtain the maximum value XZqmax and minimum value XZqmin of the write amount that can be added to the corresponding partition in unit time, and increase the write amount of the specified partition in the next unit time. The added value is the minimum value XZqmin. If the system load rate still does not reach Fz=Hz, then based on the original write amount, adjust XZqmin gradually towards XZqmax until Fz=Hz or Fz>Hz.

[0032] The present invention provides a fusion system for time series databases and traditional databases. Compared with the existing technology, it has the following advantages:

[0033] The present invention initially identifies the sub-library features of the time series database and the corresponding sub-library of the traditional database, and then determines the relevant numerical parameters based on the initially identified relevant features, and subsequently determines the standard values ​​to facilitate subsequent correlation verification;

[0034] Subsequently, the real-time monitoring of write and query data is conducted, and based on the relevant parameters of real-time monitoring, the corresponding computing power allocation parameters are identified. Then, by correlating and allocating computing power resources, through a specific allocation process, the corresponding partitions are kept in the optimal numerical operating state, achieving better operating standards, and fully guaranteeing the overall storage performance and read and write performance of the corresponding database after fusion, achieving better fusion processing results.

[0035] In the actual database fusion process, the load rates of different databases during reading, writing and downloading are monitored in a correlated manner, and comprehensive adjustments are made based on the specific related values ​​monitored. This allows the corresponding fusion system to not only ensure the maximum read, write and download volume when performing database fusion, but also simultaneously ensure that its database will not be overloaded, thereby achieving better load and read and write control effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] FIG1 is a schematic diagram of the principle framework of the present invention;

[0037] FIG2 is a schematic diagram of the system load monitoring process flow of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] Please refer to Figure 1. This application provides a fusion system of a time series database and a traditional database, including a shard database simulation end, a related data query end, a computing power resource allocation processing end, a write execution end, and a shard database data verification end. The related data query end is electrically connected to the input node of the computing power resource allocation processing end, and the shard database simulation end is bidirectionally connected to the computing power resource allocation processing end. The computing power resource allocation processing end is electrically connected to the input node of the write execution end, and the write execution end is electrically connected to the input node of the shard database data verification end.

[0040] The write execution end includes a write debugging unit, a load monitoring unit, and a monitoring analysis and control unit, wherein the write debugging unit and the load monitoring unit are both electrically connected to the input node of the monitoring analysis and control unit;

[0041] The associated data query end obtains data related to the unit time write volume of the time series database to be integrated and the unit time query volume of the traditional database from the historical completed data. The obtained related data is periodic data, and the period of the periodic data is a preset period, which is prepared in advance by relevant operators based on experience. The technical starting point of the periodic data is the current moment, and the preset period is T. That is, starting from the current moment, tracing back to the corresponding period T, the relevant data belonging to this period T is confirmed;

[0042] The computing resource allocation processing end prioritizes determining the standard write volume of the time series database and the standard query volume of the traditional database from the relevant data with correlation in the query. It then simulates the correlation of reads, writes, and queries through the sharding simulation end, and allocates computing resources in a correlated manner, ensuring that the different computing resources corresponding to different shards are in the optimal state. The specific method for correlation allocation is as follows:

[0043] The maximum write volume per unit time of the corresponding time series database in the periodic data is used as the standard write volume of the time series database (that is, based on the corresponding minimum and maximum values, the write characteristics of the corresponding time series database in a period are determined and the relevant characteristic interval is locked). The maximum query volume of the periodic data of the traditional database is used as the standard query volume of the corresponding traditional database;

[0044] The sharding simulation end simulates two groups of storage partitions, and marks one group of storage partitions as time series partitions, and the other group of storage partitions as traditional partitions. The time series partition corresponds to the sharding characteristics of the time series database (its sharding characteristics can be understood as some operating parameters associated with the corresponding database. In order to ensure the fusion effect, two associated partitions with different characteristics are set in the corresponding system to associate and store different types of data). The traditional partition corresponds to the sharding characteristics of the traditional database. Then, the computing power resource allocation processing end associates and allocates the original computing power resources, and prioritizes evenly distributing the specified computing power resources to the two groups of storage partitions (for example, when the original computing power resources are 300, 150 computing power resources are reasonably evenly distributed to different storage partitions, and associated simulation is performed to identify whether the write volume or query volume of the corresponding partition meets the standard). The write volume XR of the time series partition is recorded in real time. k And the query volume CX of traditional partitions k , where k represents different moments, the amount written is XR k and CX k Compare with the standard write amount or standard query amount, calibrate the standard write amount as BR, and the standard query amount as BC:

[0045] If XR k <BR and CX k When the value is less than BC (indicating insufficient computing power resources and the relative capabilities of the two partitions are relatively weak, resulting in the inability to meet the response requirements of the corresponding databases after database fusion), it indicates that the original computing power allocation is insufficient and the computing power allocation needs to be re-confirmed. A computing power allocation shortage signal is directly generated and displayed on the external display terminal, and the relevant operators are required to reallocate the computing power.

[0046] When XR k BR and CX k <BC time: allocate computing resources of time series partition to traditional partition, to XR k =BR or CX k =BC (the stop operation is executed first when the first partition arrives. The original method is to divide the resources equally, but after the equal distribution, some partitions will have insufficient computing power resources and some partitions will have excessive computing power resources. In this case, a secondary adjustment can be performed to ensure the reasonable distribution of computing power resources). The computing power resources allocated at the current moment are recorded as the execution allocation computing power and transmitted to the write execution end.

[0047] When XR k <BR and CX k >BC time: allocate computing resources of traditional partitions to time partitions, to XR k =BR or CX k =BC stops, records the computing power resources allocated at the current moment, uses them as the execution allocation computing power, and transmits them to the write execution end;

[0048] When there is XR k =BR or CX k =BC, no processing is required. The computing power resources allocated at the current moment are recorded as the execution allocation computing power and transmitted to the write execution end.

[0049] The write execution end allocates computing power to the designated partition within the actual fusion system based on the execution power allocated to the corresponding partition. (That is, if there are two partitions in the fusion system, one for storing traditional data and the other for storing time series data, the computing power allocated to each partition is different.) The write execution end includes a write debugging unit, a load monitoring unit, and a monitoring, analysis, and control unit.

[0050] The write debugging unit performs write debugging for the database fusion process: by determining a set of debugging cycles, adjusting the write volume of different partitions within these debugging cycles, recording the database load, and locking the associated change intervals of different partitions based on the value changes of different partitions (the write here refers to the database fusion stage, the process of two different partitions writing and storing related data stored in the original database. It has the same meaning as the write mentioned above, but different application scenarios. The above write is for the corresponding write and storage process of real-time time series data). The specific method for determining the partition associated change interval is:

[0051] Determine a set of debugging cycles t, where t is a preset value, and its specific value is determined by the operator based on experience. Divide this test cycle t into two fine-tuning cycles t1 and t2;

[0052] During the fine-tuning period t1, gradually increase the write volume XL of the associated partition storing time series data, record the numerical changes in its system load rate, and keep the write volume of the other partition unchanged. Based on the increase value Z1 of the write volume XL per unit time and the increase value Z2 of the system load rate, determine the associated parameter corresponding to the unit time. This associated parameter = Z2 ÷ Z1 (as the write volume increases, the load rate of the entire system will also increase synchronously, because the write volume of the other partition remains unchanged). Based on the several sets of associated parameters determined during the fine-tuning period t1, and based on the minimum and maximum values ​​of the associated parameters, determine the associated change range of the associated partition storing time series data.

[0053] During the fine-tuning period t2, the write volume of the associated partition storing traditional data is gradually increased, and the associated change interval of the associated partition storing traditional data is locked by recording the numerical change of its system load rate and keeping the write volume of another partition unchanged;

[0054] Transmitting the determined two sets of associated change intervals to the monitoring, analysis and control unit;

[0055] Specifically, in order to avoid system load during the database fusion process, once system load occurs, it will cause congestion during writing, resulting in data loss during writing. In that case, when the data is supplemented later, the amount of data will be huge. Therefore, in order to avoid system load, it is necessary to perform correlation monitoring and ensure that the system is in a normal writing process through real-time control of the writing amount.

[0056] The load monitoring unit monitors the system load rate in real time during the database fusion process and transmits the real-time monitored system load rate to the monitoring, analysis and control unit in real time;

[0057] In conjunction with Figure 2, the monitoring, analysis, and control unit adjusts the write volume of different partitions in real time based on the system load rate monitored in real time at the current moment, ensuring that the system is not in a loaded state while meeting the maximum write volume. The specific method of performing real-time adjustment is as follows:

[0058] Based on the system load rate Fz monitored at the current moment, the standard check value Hz associated with the system load rate is extracted, where Hz is a preset standard parameter, which is prepared in advance by relevant operators based on experience;

[0059] If Fz>Hz, the write volume of different partitions will be reduced in real time until Fz≤Hz:

[0060] Record the newly added write amount from the previous moment to the current moment of the corresponding partition, and mark the different newly added write amounts of different partitions as Xq, where q represents different partitions, and then determine the associated change interval based on the corresponding partition [GL1 q , GL2 q ] to determine the system load change value that can be generated when the corresponding partition write amount changes: GL1 q ×Xq=BHmin and GL2 q ×Xq=BHmax, to determine the system load change value range [BHmin, BHmax] that can be generated by the corresponding partition;

[0061] Then, the middle value Zq of the interval is locked from the determined load change value interval, and the middle values ​​Zq of the two intervals are compared to determine their ratio sequence. Then, based on this ratio sequence, the write amount of different partitions is reduced, and the specific ratio of the specific value of the reduction is consistent with this ratio sequence;

[0062] If Fz=Hz, the write volume of the original corresponding partition remains unchanged, and its system load rate is monitored in real time;

[0063] If Fz < Hz, then determine the difference in system load rate: Cz = Hz - Fz, and then based on the different associated change intervals [GL1 q , GL2 q , where q represents different partitions, and use (Cz÷2)÷GL1 q = XZqmax and (Cz÷2)÷GL2 q = XZqmin, to obtain the maximum value XZqmax and the minimum value XZqmin that can be newly added to the write volume per unit time for the corresponding partition, and add the newly added value, which is the minimum value XZqmin, to the write volume of the specified partition in the next unit time. If the system load rate still does not reach Fz = Hz, then on the basis of the original write volume (that is, the corresponding moment when Fz < Hz, such as the situation determined at the 5th moment, the determined interval is [1, 5], and the newly added amount used in the next unit time is 1, so add 1 to the original write volume. When monitoring, if there is no abnormality, then 2 can be added to the original write volume. Relative to the previous moment of the corresponding moment, it is an increase of 1, because 1 has already been added in the previous moment, and so on, to adjust the write volume of the relevant partition in real time), gradually adjust XZqmin to XZqmax until Fz = Hz or Fz > Hz;

[0064] [[ID=**12**]]Example: Monitor the system load rate (Fz): First, it is necessary to monitor the load rate of the current system in real time.

[0065] [[ID=**15**]]Extract the standard check value (Hz): Extract the preset standard parameter Hz according to the system load rate, which is a threshold set by the operator based on experience.

[0066] [[ID=**18**]]Compare Fz and Hz:

[0067] [[ID=**21**]]If Fz > Hz, then reduce the write volume until Fz ≤ Hz.

[0068] [[ID=**24**]]If Fz = Hz, then keep the current write volume unchanged and continue to monitor.

[0069] [[ID=**27**]]If Fz < Hz, then the write volume can be increased.

[0070] [[ID=**30**]]Adjust the write volume:

[0071] [[ID=**33**]]When Fz > Hz, calculate the influence interval [BHmin, BHmax] of the newly added write volume (Xq) for each partition on the system load, determine the intermediate value Zq, and reduce the write volume according to this ratio.

[0072] Note: There seems to be a minor issue in the original text where the closing bracket for the interval in line 5 is missing. It should be something like [GL1, GL2]. Also, in the translation, the line breaks are maintained as per the instructions. The text in bold in the translation is for better readability to show the corresponding parts in the original text.When Fz < Hz, calculate the difference Cz in the system load rate, and determine the maximum and minimum write amounts XZqmax and XZqmin that can be increased in each partition per unit time according to the change interval [GL1q, GL2q] of different partitions. Then start increasing gradually from the minimum value until reaching Hz or exceeding it.

[0073] Some of the data in the above formula are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0074] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A fusion system of time series database and traditional database, characterized by: include: The associated data query end obtains data related to the time-per-unit-time write volume of the time series database to be integrated and the data related to the time-per-unit-time query volume of the traditional database from the historical completed data. The obtained related data is periodic data, and the period of the periodic data is the preset period. The computing resource allocation processing end prioritizes determining the standard write volume of the time series database and the standard query volume of the traditional database from the relevant data with correlation in the query. It then simulates the correlation of read, write, and query through the sharding simulation end, and allocates computing resources accordingly, ensuring that the different computing resources corresponding to different shards are in the optimal state. The write execution end allocates computing power to the designated partitions based on the execution power allocated to the corresponding partitions. In the actual fusion system, the corresponding computing power resources are allocated to the designated partitions for execution. The write execution end includes a write debugging unit, a load monitoring unit, and a monitoring, analysis, and control unit. The write debugging unit performs write debugging for the database integration process. It determines a set of debugging cycles, adjusts the write volume of different partitions within these debugging cycles, records the database load, and locks the associated change intervals of different partitions based on the value changes of different partitions. The specific method is as follows: Determine a set of debugging cycles t, where t is a preset value, and divide this test cycle t into two fine-tuning cycles t1 and t2; During the fine-tuning period t1, the write volume XL of the associated partition storing time series data is gradually increased, and the numerical change of its system load rate is recorded. The write volume of another partition remains unchanged. Based on the increase value Z1 of the write volume XL per unit time and the increase value Z2 of the system load rate, the associated parameter corresponding to the unit time is determined. This associated parameter = Z2 ÷ Z1. Based on the several sets of associated parameters determined during the fine-tuning period t1 and the minimum and maximum values ​​of the associated parameters, the associated change interval of the associated partition storing time series data is determined. During the fine-tuning period t2, the write volume of the associated partition storing traditional data is gradually increased, and the associated change interval of the associated partition storing traditional data is locked by recording the numerical change of its system load rate and keeping the write volume of another partition unchanged; Transmitting the determined two sets of associated change intervals to the monitoring, analysis and control unit; The load monitoring unit monitors the system load rate in real time during the database fusion process and transmits the real-time monitored system load rate to the monitoring, analysis and control unit in real time; The monitoring, analysis, and control unit adjusts the write volume of different partitions in real time based on the system load rate monitored in real time at the current moment, ensuring that the system is not under load while meeting the maximum write volume.

2. The fusion system of time series database and traditional database according to claim 1 is characterized in that: The specific method of the computing power resource allocation processing end for associating and allocating computing power resources is as follows: The maximum write volume per unit time of the periodic data corresponding to the time series database is used as the standard write volume of the time series database. The maximum query volume of the periodic data of the traditional database is used as the standard query volume of the corresponding traditional database. The database sharding simulation end simulates two groups of storage partitions, and marks one group of storage partitions as time series partitions and the other group of storage partitions as traditional partitions. Time series partitions correspond to the database sharding characteristics of time series databases, and traditional partitions correspond to the database sharding characteristics of traditional databases. Then, the computing power resource allocation processing end associates and allocates the original computing power resources, giving priority to evenly allocating the specified computing power resources to the two groups of storage partitions, and records the write volume of the time series partitions in real time. k And the query volume CX of traditional partitions k , where k represents different moments, the amount written is XR k and CX k Compare with the standard write amount or standard query amount, calibrate the standard write amount as BR, and the standard query amount as BC: When XR k BR and CX k <BC time: allocate computing resources of time series partition to traditional partition, to XR k =BR or CX k =BC stops, records the computing power resources allocated at the current moment, uses them as the execution allocation computing power, and transmits them to the write execution end; When XR k <BR and CX k >BC time: allocate computing resources of traditional partitions to time partitions, to XR k =BR or CX k =BC stops, records the computing power resources allocated at the current moment, uses them as the execution allocation computing power, and transmits them to the write execution end; When there is XR k =BR or CX k =BC, no processing is required. The computing power resources allocated at the current moment are recorded as the execution allocation computing power and transmitted to the write execution end.

3. The fusion system of time series database and traditional database according to claim 2 is characterized in that: If XR k <BR and CX k When <BC, a signal of insufficient computing power allocation is directly generated and displayed on the external display terminal.

4. The fusion system of time series database and traditional database according to claim 1 is characterized in that: The specific method of the monitoring, analysis and control unit to adjust the write volume of different partitions in real time is as follows: Based on the system load rate Fz monitored at the current moment, the standard check value Hz associated with the system load rate is extracted, where Hz is a preset standard parameter; If Fz>Hz, the write amount of different partitions will be reduced in real time until Fz≤Hz; If Fz=Hz, the write volume of the original corresponding partition remains unchanged, and its system load rate is monitored in real time; If Fz < Hz, then determine the system load rate difference: Cz = Hz - Fz, and then based on the different associated change intervals determined by different partitions [GL1 q , GL2 q ]Add the write amount of different partitions, where q represents different partitions.

5. The fusion system of time series database and traditional database according to claim 4 is characterized in that: When Fz>Hz, the specific method of reducing the write amount of different partitions in real time is: Record the newly added write amount from the previous moment to the current moment of the corresponding partition, and mark the different newly added write amounts of different partitions as Xq, where q represents different partitions, and then determine the associated change interval based on the corresponding partition [GL1 q , GL2 q ] to determine the system load change value that can be generated when the corresponding partition write amount changes: GL1 q ×Xq=BHmin and GL2 q ×Xq=BHmax, to determine the system load change value range [BHmin, BHmax] that can be generated by the corresponding partition; Then, the middle value Zq of the interval is locked from the determined load change value interval, and the middle values ​​Zq of the two intervals are ratio-processed to determine their ratio sequence. Then, based on this ratio sequence, the write amount of different partitions is reduced, and the specific ratio of the specific reduced values ​​is consistent with this ratio sequence.

6. The fusion system of time series database and traditional database according to claim 4 is characterized in that: When Fz is less than Hz, the specific method for increasing the write amount of different partitions is as follows: Use (Cz÷2)÷GL1 q =XZqmax and (Cz÷2)÷GL2 q =XZqmin, obtain the maximum value XZqmax and minimum value XZqmin of the write amount that can be added to the corresponding partition in unit time, and increase the write amount of the specified partition in the next unit time. The added value is the minimum value XZqmin. If the system load rate still does not reach Fz=Hz, then based on the original write amount, adjust XZqmin gradually towards XZqmax until Fz=Hz or Fz>Hz.

Images