WO2023046059A1

WO2023046059A1 - Cache warmup method and apparatus, and computer device and storage medium

Info

Publication number: WO2023046059A1
Application number: PCT/CN2022/120815
Authority: WO
Inventors: 谢磊; 王相玲
Original assignee: 中国第一汽车股份有限公司
Priority date: 2021-09-24
Filing date: 2022-09-23
Publication date: 2023-03-30
Also published as: CN113849532A

Abstract

A cache warmup method and apparatus, and a computer device and a storage medium. The cache warmup method comprises: acquiring service collection access information within a historical time period (S110); inputting the service collection access information within the historical time period into a pre-trained service prediction model, so as to obtain service prediction access information within a target time period (S120); inputting the service prediction access information into a pre-trained data prediction model, so as to obtain a plurality of predicted hotspot evaluation values of a plurality of pieces of data in a database within the target time period (S130); and according to the plurality of predicted hotspot evaluation values, selecting hotspot data from the plurality of pieces of data in the database, and performing cache warmup (S140).

Description

Cache preheating method, device, computer equipment and storage medium

This application claims the priority of the Chinese patent application with application number 202111121831.0 filed with the China Patent Office on September 24, 2021, the entire content of which is incorporated herein by reference.

technical field

The embodiments of the present application relate to computer technology, for example, to a cache preheating method, device, computer equipment, and storage medium.

Background technique

Cache applications are essential in large-scale websites or high-concurrency processing systems, and cache warming is one of the important measures to improve data query performance.

In related technologies, cache preheating is realized mainly by fixedly initializing all data of a service according to the number of visits, or by manually selecting data according to service characteristics.

However, the cache warm-up method in the related art has the problems of low cache hit rate, large warm-up range, low cache warm-up accuracy rate and high labor cost caused by a large amount of useless data in the cache.

Contents of the invention

Embodiments of the present application provide a cache preheating method, device, computer equipment, and storage medium, so as to implement cache data preheating, improve cache preheating accuracy, and reduce labor costs.

The embodiment of the present application provides a cache preheating method, including:

Obtain business collection and access information in historical time periods;

Inputting the business collection and access information of the historical time period into the pre-trained business prediction model to obtain the business forecast access information of the target time period;

Inputting the business prediction access information into the pre-trained data prediction model to obtain multiple predicted hot comment values of multiple data in the database of the target time period;

According to the values of the plurality of predicted hotspots, hotspot data is selected from the plurality of data in the database for cache preheating.

The embodiment of the present application also provides a cache preheating device, including:

A business collection and access information acquisition module is configured to obtain business collection and access information in historical time periods;

The business forecast visit information acquisition module is configured to input the business collection visit information of the historical time period into the pre-trained business forecast model to obtain the business forecast visit information of the target time period;

The predicted hot comment value acquisition module is configured to input the business forecast access information into the pre-trained data prediction model to obtain multiple predicted hot comment values of multiple data in the database of the target time period;

The cache preheating module is configured to filter out hot data from the plurality of data in the database for cache preheating according to the values of the plurality of predicted hot comments.

The embodiment of the present application also provides a computer device, and the computer device includes:

one or more processors;

a storage device configured to store one or more programs;

When the one or more programs are executed by the one or more processors, the one or more processors implement the cache warming method provided in the embodiment of the present application.

The embodiment of the present application also provides a storage medium including computer-executable instructions, and the computer-executable instructions are used to execute the cache warming method provided in the embodiment of the present application when executed by a computer processor.

Description of drawings

FIG. 1 is a flow chart of a cache preheating method in Embodiment 1 of the present application;

FIG. 2 is a flow chart of a cache preheating method in Embodiment 2 of the present application;

FIG. 3 is a flow chart of a cache preheating method in Embodiment 3 of the present application;

FIG. 4 is a schematic diagram of an application scenario of a warm-up cache method in Embodiment 4 of the present application;

FIG. 5 is a schematic structural diagram of a cache preheating device in Embodiment 5 of the present application;

FIG. 6 is a schematic structural diagram of a computer device in Embodiment 6 of the present application.

Detailed ways

The application will be described below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, but not to limit the present application. In addition, it should be noted that the drawings only show some structures related to the present application but not all structures.

Embodiment one

Figure 1 is a flow chart of a cache preheating method provided in Embodiment 1 of the present application. This embodiment is applicable to the situation of cache preheating, and the method can be executed by a cache preheating device, which can be implemented by software and and/or implemented by hardware, and configured in a computer device, the computer device may be a server device and a client device, for example, the client device may be a mobile phone, a tablet computer, a vehicle terminal or a desktop computer, etc. The cache warming method includes the following steps:

S110. Obtain service collection and access information in historical time periods.

The historical time period refers to a time period before the current time, for example, it may be the previous 8 hours, the previous day, or the previous week. Business collected visit information refers to the business type and visit volume of the visit information collected in the historical time period. The business type may be order placing, return and refund, etc., where the order placing business type includes business data such as product information, product images, and recommended product lists. The number of visits refers to the number of visits to data. Exemplarily, when a user opens a web page with detailed content of a product, the number of visits to data such as product information, product images, and recommended product lists displayed on the web page increases. It should be noted that business types can also be divided in more detail. For example, the order business type can be subdivided into business types such as food, clothing, and daily necessities. The level of detail of the division can be set according to actual needs. Do limited.

S120. Input the service collection and access information of the historical time period into the pre-trained service prediction model to obtain the service forecast visit information of the target time period.

The target time period is a time period that needs to be predicted, that is, a period of time after the current time period. Exemplarily, it may be 8 hours, a day, or a week after the current time. The service forecast visit information refers to predicting the visit volume of the service to be visited in a future time period. The business forecasting model refers to a model in which the input is the service collection and access information in the historical time period, and the output is the service forecasting access information in the target time period. Exemplarily, the service prediction model may be a convolutional neural network model, a time series neural network model, an extreme learning machine model, or an autoencoder neural network model. A pre-trained business forecasting model refers to a model that meets the preset prediction accuracy rate conditions obtained by training the model through sample data in advance, wherein meeting the forecast accuracy rate conditions means that the forecast accuracy rate of the business forecasting model is greater than or equal to the preset Accuracy threshold, the preset accuracy threshold is a manually preset value.

S130. Input the service forecast access information into a pre-trained data forecast model to obtain multiple predicted hot comment values of multiple data in the database in the target time period.

The data prediction model refers to a model in which the input is business forecast access information in the target time period, and the output is multiple predicted hot spot values of multiple data in the database in the target time period. Exemplarily, the data prediction model may be a convolutional neural network model, a time series neural network model, an extreme learning machine model, or an autoencoder neural network model. A database refers to a collection of data associated with a business. The hot comment value indicates the degree of attention to the data. Exemplarily, the hot comment value may be the probability value of data being accessed. The hot comment value calculation method can be set artificially, for example, divide the visit volume of a piece of data by the total data visit volume of the business corresponding to the data within a period of time as the hot comment value of the data.

S140. According to the values of the plurality of predicted hotspots, select hotspot data from the plurality of data in the database to preheat the cache.

Hot data refers to data with a high predicted access probability. Cache warming refers to loading the data in the database into the cache in advance. After obtaining multiple predicted hot spot values, determine a numerical range interval for the predicted hot spot value, and use this range interval as a condition for filtering hot spot data. Screening hot data from a plurality of data in the database refers to filtering data in the database whose predicted hot comment value is within the numerical range of the predicted hot comment value, and the filtered data is hot data.

In an optional embodiment, the hot data is selected from multiple data in the database for cache preheating according to the multiple predicted hot comment values, including: multiple data in the database Screen out the data whose predicted hot comment value is greater than or equal to the preset evaluation value threshold, and determine the filtered data as hot data; store the hot data in the cache.

According to the obtained predicted hot comment value, the data corresponding to the predicted hot comment value greater than or equal to the preset evaluation value threshold is taken as hot data. Wherein, the preset evaluation value threshold refers to a value set artificially in advance, which is used to screen hot data, that is, when the predicted hot comment value of the data is greater than or equal to the preset evaluation value threshold, the data is determined to be hot data. Filter out all the data in the database whose value of predicted hot reviews is greater than or equal to the preset evaluation value threshold, and load these data into the cache.

Screen out hot data based on the value of predicted hot reviews, reduce the cache space occupied by invalid data, improve the accuracy of hot data screening, and improve the utilization of cache space. By preheating the cache, when a user requests data, the data can be read directly from the cache without reading from the database, reducing the pressure on the database and saving system performance overhead.

In the embodiment of the present application, the business prediction model is used to predict the business prediction access information of the target time period based on the business collection and access information in the historical time period, and the data prediction model is used to predict the value of multiple hot comments of multiple data in the database. The evaluation value screens out hot data for cache preheating, which solves the problems of low cache preheating accuracy and high labor costs caused by fixedly initializing all data of a business or manually selecting data according to business characteristics in related technologies, and realizes data based on historical time The segment business collects access information, predicts hot data, predicts business forecast access information and hot comment value based on two models, refines the prediction process, improves the accuracy of hot data prediction and reduces labor costs.

Embodiment two

FIG. 2 is a flow chart of a cache preheating method provided in Embodiment 2 of the present application. The technical solution of this embodiment is described on the basis of the above technical solution. Before inputting the business forecasting access information into the pre-trained data forecasting model to obtain multiple forecasted hot comment values of multiple data in the database of the target time period, it also includes: obtaining data training samples, the data training The samples include database access information, service access information, and a plurality of detected hot comment values of a plurality of data in the database; the first model is trained according to the data training samples; when the first model training is completed, the The first model at the current moment is determined as a data prediction model. The method includes:

S210. Acquire service collection and access information in historical time periods.

S220. Input the service collection and access information of the historical time period into the pre-trained service prediction model to obtain the service forecast visit information of the target time period.

S230. Acquire a data training sample, where the data training sample includes database access information, service access information, and values of multiple detected hotspots of multiple data in the database.

Data training samples are used to train data prediction models. The data training samples are database access information, business access information, and multiple detection hotspot evaluation values of multiple data in the database within a period of time. The database access information refers to the number of visits to each of the multiple data in the database. Exemplarily, the data in the database includes product information, product images, and recommended product lists, etc.; the detected hot comment value of each data in the database can be Obtained by manual labeling. Business access information refers to the access volume of each data under each business in various businesses.

In an optional embodiment, the data training samples further include cache access information.

The cache access information is the visit volume of each of the multiple data in the cache, and the data in the cache is a source of information for a user when accessing a website.

By using the cache access information as a data training sample, the content of the data training sample can be enriched, the representativeness of the data training sample can be increased, and the prediction accuracy of the data prediction model can be improved.

In an optional embodiment, the acquiring data training samples includes: acquiring cache access information, database access information, and service access information during the sampling period, and determining multiple detection hotspots of multiple data in the database Value: forming a data training sample according to cache access information, data access information, service access information and values of multiple detected hotspots of multiple data in the database during the sampling period.

According to the cache access information, database access information and service access information in the sampling time period, the values of multiple detection hotspots of multiple data in the database can be calculated. Exemplarily, according to the visit volume of data associated with service access information, the visit volume of data associated with cache access information, and the visit volume of data associated with data access information, the access of each of the multiple data in the database can be counted amount, and normalize the access amount of each data to obtain the detection hotspot value of each data in the multiple data in the database.

Using cache access information, data access information, business access information, and multiple detection hot comment values pre-marked for multiple data in the database as training samples, increasing the coverage of data training sample data and increasing the coverage of data training sample data Representativeness can improve the accuracy of the trained data prediction model.

S240. Train the first model according to the data training samples.

The first model is an intelligent algorithm model. Exemplarily, the first model may include a linear model, a decision tree model, or a deep learning model. The data training samples are used as input data to train the first model, and the training is a process in which the first model continuously adjusts the parameters of the first model according to the training results.

In an optional embodiment, the determining the values of multiple detected hotspots of multiple data in the database includes: inputting data access information and business access information in the sampling time period into the first model In the method, a plurality of detection hotspot evaluation values of a plurality of data in the database are determined.

Before determining the value of multiple detection hot spots of multiple data in the database, first construct a model and train the model to obtain the first model. Exemplary constructed models may include linear models, decision tree models or deep learning models, etc. . The training samples are constructed manually, and the training samples are used to train the constructed model, and the detection hot spot value of each data in the multiple data in the training samples is manually marked. The training samples are divided into training set and validation set. Exemplarily, 80% of the training samples may be used as a training set, and 20% of the samples may be used as a verification set. The training samples are used to train the built model, and the verification set is used to verify whether the output accuracy of the built model has reached the expected effect. For example, the output accuracy threshold of the model can be preset, and the output accuracy threshold is used to measure the built model Whether to complete the training, when the output accuracy of the model is greater than or equal to the preset model output accuracy threshold, the model training is completed to obtain the first model; when the output accuracy of the data prediction model is less than the preset model output accuracy threshold , continue to train the model until the output accuracy of the model is greater than or equal to the preset output accuracy threshold of the model, then end the training and obtain the first model. After the first model is obtained, the data access information and service access information of the sampling time period are input into the first model, and the first model outputs multiple detection hotspot evaluation values of multiple data in the database.

By inputting the data access information and business access information in the sampling period into the first model, it is possible to determine the value of multiple detection hotspots of multiple data in the database, and obtain multiple data in the database through the first model Multiple detected hot spot values, improving the accuracy of hot spot value detection for each of the multiple data in the database, thereby improving the accuracy of hot spot value prediction.

S250. When the training of the first model is completed, determine the first model at the current moment as the data prediction model.

The completion of the first model training means that the prediction accuracy of the first model is greater than or equal to the preset accuracy threshold or the set rounds of training have been completed. Wherein, the preset accuracy rate threshold may be a value set manually, and the number of setting rounds refers to the number of times set manually. The current moment refers to the moment when the training of the first model is completed, and the first model at the current moment is saved and determined as the data prediction model.

S260. Input the service forecast access information into a pre-trained data forecast model to obtain multiple predicted hot comment values of multiple data in the database in the target time period.

S270. According to the values of the plurality of predicted hotspots, select hotspot data from multiple pieces of data in the database to preheat the cache.

In this embodiment of the present application, the first model is trained to determine the data prediction model by acquiring database access information, business access information, and multiple detection hot comment value data training samples of multiple data in the database. The data training samples include multiple This information increases the coverage of data training samples, increases the representativeness of data training samples, improves the prediction accuracy of data prediction models, improves the accuracy of cached data predictions, and saves labor costs.

Embodiment three

FIG. 3 is a flow chart of a cache preheating method provided in Embodiment 3 of the present application. The technical solution of this embodiment is described on the basis of the above technical solution. Before inputting the business collection and access information of the historical time period into the pre-trained business forecast model to obtain the business forecast visit information of the target time period, it also includes: obtaining the business collection and visit information of the first day, and the second day’s business collection and visit information. The business collects access information, and generates business training samples; trains the second model according to the business training samples; and determines the second model at the current moment as the business prediction model when the second model training is completed. The method includes:

S310. Acquire service collection and access information in historical time periods.

S320. Obtain the service collection visit information of the first day and the service collection visit information of the second day, and generate service training samples.

The business collection and access information on the first day is the input data in the training sample, the business collection and access information on the second day is the output data in the training sample, and the business collection and access information on the first day and the business collection and access information on the second day are formed Business training samples.

S330. Train the second model according to the service training samples.

The second model is an intelligent algorithm model. Exemplarily, the second model may be a convolutional neural network model, a time series neural network model, an extreme learning machine model, or an autoencoder neural network model. The business training samples are used as input to train the second model, and the training is the process in which the second model continuously adjusts the parameters of the second model according to the training results.

S340. When the training of the second model is completed, determine the second model at the current moment as the service prediction model.

The completion of the second model training means that the prediction accuracy of the second model is greater than or equal to the preset accuracy threshold or the set rounds of training have been completed. Wherein, the preset accuracy rate threshold may be a value set manually, and the number of setting rounds refers to the number of times set manually. The current moment refers to the moment when the second model finishes training, and the second model at the current moment is saved and determined as the service prediction model.

S350. Input the service collection and access information of the historical time period into the pre-trained service prediction model to obtain the service forecast visit information of the target time period.

S360. Input the service forecast access information into a pre-trained data forecast model to obtain multiple predicted hot comment values of multiple data in the database in the target time period.

S370. According to the multiple predicted hot comment values, select hot data from multiple data in the database to preheat the cache.

In this embodiment of the present application, the business collection and visit information of the first day and the business collection and visit information of the second day are used as business training samples to train the second model and determine the business prediction model, so that the next day can be predicted based on the business collection and visit information of the previous day The business collects and visits information, improves the accuracy of the business forecasting model and saves labor costs.

Embodiment four

FIG. 4 is a schematic diagram of an application scenario of a warm-up cache method provided in Embodiment 4 of the present application. The technical solution of this embodiment is applicable to the application scenario of the cache preheating method, and the schematic diagram includes:

The database access information module 410, the service access information module 420 and the cache access information module 430 are respectively configured to acquire database access information, service access information and cache access information in the sampling period.

The data training sample module 460 is configured to use the acquired database access information, service access information and cache access information data to form a data training sample, and the data training sample is used to train the data prediction model 470 .

The service access information acquired by the service access information module 420 in the sampling period is used to train the service prediction model 440 . The business forecast model 440 is used to obtain business forecast visit information according to the input business visit information. Exemplarily, the business forecast model 440 can predict the business visit information of D+1 day according to the business visit information, wherein D+1 represents the number of days plus one, That is, the service prediction model 440 is used to predict service forecast visit information 24 hours after the input service visit information.

The business forecast visit information module 450 is configured to input the business forecast visit information obtained from the business forecast model 440 into the data forecast model 470 as input data.

The data prediction model 470 is used to obtain the predicted hot comment value according to the service forecast access information. Exemplarily, the data forecast model 470 predicts the data access information and hot comment value of D+1 day according to the business forecast visit information of D+1 day. The input of the data prediction model 470 is business forecast access information, and the output is data access information and hot comment value. The value of hot reviews is normalized so that the evaluation values of hot spots are concentrated between 0 and 1.

The predicted hot comment value module 480 is configured to screen hot data according to the obtained hot comment value. Exemplarily, the predicted hot comment value greater than or equal to 0.7 is selected from the predicted hot comment value, and the data corresponding to the filtered predicted hot comment value is used as hot data.

The cache preheating module 490 is configured to load hotspot data into the cache, perform cache preheating, and set the valid time to 24 hours.

The cache destruction module 411 is set to destroy the cache regularly, that is, the cache data is cleared when the storage time in the cache is equal to the valid time set, so as to load new hot data, and load the clear content into the data training sample module 460 for training The data prediction model is to update the data prediction model 470 in time to improve the accuracy of predicting the value of hot reviews in the next time period.

The functional modules of this embodiment can be developed through Java code.

The embodiment of the present application predicts the value of multiple predicted hot comments of multiple data in the database through the obtained business forecast access information, and filters out the hot data, and sets the valid time of the cache. The time for storing the cached data in the cache is equal to the set valid time Then use the cached data for data prediction model training and update the cached data. Realize timely update of cache data and improve the accuracy of data prediction model, improve the accuracy and real-time performance of cache preheating.

Embodiment five

FIG. 5 is a schematic structural diagram of a cache preheating device provided in Embodiment 5 of the present application. Embodiment 5 is a corresponding device for implementing the cache preheating method provided in the above embodiments of the present application. The device can be implemented in software and/or hardware, and can generally be integrated into a computer device. The buffer preheating device includes: a business collection and access information acquisition module 510, which is configured to obtain business collection and access information in historical time periods; a service forecast access information acquisition module 520, configured to input the business collection and access information in the historical time period into the pre-set The trained business prediction model obtains the business forecast visit information of the target time period; the forecasted hot comment value acquisition module 530 is configured to input the business forecast visit information into the pre-trained data forecast model to obtain the target time period database multiple predicted hot comment values of multiple data; the cache warm-up module 540 is configured to filter out hot data from the multiple data in the database to perform cache warm-up according to the multiple predicted hot comment values .

Optionally, the cache preheating device further includes: a training sample acquisition module configured to acquire data training samples, the data training samples including database access information, business access information and multiple detections of multiple data in the database hot comment value; the first model training module is set to train the first model according to the data training samples; the data prediction model determination module is set to use the first model at the current moment when the first model training is completed Determine the model for data prediction.

Optionally, the data training samples further include cache access information.

Optionally, the training sample acquisition module is configured to: acquire cache access information, database access information, and business access information in the sampling time period, and determine the value of multiple detection hotspots of multiple data in the database; A data training sample is formed according to cache access information, data access information, service access information, and multiple detected hot comment values of multiple data in the database during the sampling period.

Optionally, the training sample acquisition module determines the value of multiple detected hotspot reviews of multiple data in the database in the following manner: input data access information and service access information in the sampling time period into the first In the model, a plurality of detection hot spot values of a plurality of data in the database are determined.

Optionally, the cache preheating device also includes: a business training sample generation module, configured to obtain the business collection and access information of the first day and the business collection and visit information of the second day, and generate business training samples; the second model training A module configured to train the second model according to the business training samples; a business prediction model determining module configured to determine the second model at the current moment as the business prediction model when the training of the second model is completed.

Optionally, the cache preheating module is set to: filter out the data whose value of the predicted hot comment is greater than or equal to the preset evaluation value threshold from the plurality of data in the database, and filter out the data determined as hotspot data; storing the hotspot data in the cache.

The above-mentioned device can execute the cache warm-up method provided in the embodiment of the present application, and has corresponding functional modules and effects for executing the cache warm-up method.

Embodiment six

FIG. 6 is a schematic structural diagram of a computer device provided in Embodiment 6 of the present application. FIG. 6 shows a block diagram of an exemplary computer device 12 suitable for implementing embodiments of the present application. The computer device 12 shown in FIG. 6 is only an example, and should not limit the functions and scope of use of the embodiment of the present application.

As shown in FIG. 6, computer device 12 takes the form of a general-purpose computing device. Components of computer device 12 may include, but are not limited to: one or more processors or processing units 16 , system memory 28 , bus 18 connecting various system components including system memory 28 and processing unit 16 . The computer device 12 may be a bus-attached device.

Bus 18 represents one or more of a variety of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus structures. For example, these bus structures include but are not limited to Industry Standard Architecture (Industry Standard Architecture, ISA) bus, Micro Channel Architecture (Micro Channel Architecture, MCA) bus, Enhanced ISA bus, Video Electronics Standards Association (Video Electronics Standards Association, VESA) local bus and Peripheral Component Interconnect (PCI) bus.

Computer device 12 includes a variety of computer system readable media. These media can be any available media that can be accessed by computer device 12 and include both volatile and nonvolatile media, removable and non-removable media.

System memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM) 30 and/or cache memory 32 . Computer device 12 may include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be configured to read and write to non-removable, non-volatile magnetic media (not shown in FIG. 6, commonly referred to as a "hard drive"). Although not shown in Figure 6, the present application may provide a disk drive configured to read and write to a removable non-volatile disk (such as a "floppy disk"), as well as to a removable non-volatile disk (such as a compact disk ROM (Compact Disc Read-Only Memory, CD-ROM), Digital Video Disc (Digital Video Disc-Read Only Memory, DVD-ROM) or other optical media) CD-ROM drive. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (eg, at least one) of program components configured to perform the functions of various embodiments of the present application.

Programs/utilities 40 may be stored, for example, in system memory 28 as a set (at least one) of program components 42 including, but not limited to, an operating system, one or more application programs, other program components, and program data, each or a combination of these examples may include implementations of the network environment. Program components 42 generally perform the functions and/or methodologies of the embodiments described herein.

The computer device 12 may also communicate with one or more external devices 14 (e.g., a keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with the computer device 12, and/or with Any device (eg, network card, modem, etc.) that enables the computing device 12 to communicate with one or more other computing devices. This communication can be performed through an input/output (Input/Output, I/O) interface 22 . And, computer device 12 can also communicate with one or more networks (such as local area network (Local Area Network, LAN), wide area network (Wide Area Network, WAN)) by network adapter 20. As shown in the figure, network adapter 20 communicates with by bus 18 other components of computer device 12. Although not shown in FIG. 6, computer device 12 may be used in conjunction with other hardware and/or software components, including but not limited to: microcode, device drivers, Additional processing units, external disk drive array (Redundant Arrays of Inexpensive Disks, RAID) systems, tape drives, and data backup storage systems, etc.

The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28 , such as implementing the cache warming method provided in any embodiment of the present application.

Embodiment seven

Embodiment 7 of the present application provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the cache warming method provided in all embodiments of the present application is implemented: that is, the program is processed The following methods are implemented when the controller is executed: acquiring the business collection and access information of the historical time period; inputting the business collection and access information of the historical time period into the pre-trained business forecasting model to obtain the business forecasting visit information of the target time period; The business forecast access information is input into the pre-trained data prediction model, and multiple predicted hot comment values of multiple data in the database in the target time period are obtained; according to the multiple predicted hot comment values, filter out from the database Hotspot data is cached and warmed up.

The computer storage medium in the embodiments of the present application may use any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared semiconductor system, device, or device, or any combination thereof. Computer-readable storage media can include (a non-exhaustive list) electrical connections with one or more conductors, portable computer disks, hard disks, RAM, Read Only Memory (ROM), erasable programmable only Read memory (Erasable Programmable Read Only Memory, EPROM), flash memory, optical fiber, portable CD-ROM, optical storage device, magnetic storage device, or any combination of the above. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

A computer-readable signal medium may include a data signal in baseband or as a carrier wave, and the computer-readable signal medium carries computer-readable program code. Such propagated data signals may take many forms, including - but not limited to - electromagnetic signals, optical signals, or any combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which can be sent, propagated, or transported for use by or in conjunction with an instruction execution system, apparatus, or device. Programs used in conjunction with the device.

The program code contained on the computer readable medium can be transmitted by any medium, including—but not limited to—wireless, wire, optical cable, radio frequency (Radio Frequency, RF), etc., or any combination of the above.

Computer program code for carrying out the operations of the present application may be written in one or more programming languages, including object-oriented programming languages—such as Java, Smalltalk, C++, as well as conventional procedural programming languages Design Language—such as "C" or a similar programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. Where a remote computer is involved, the remote computer can be connected to the user computer through any kind of network, including a LAN or WAN, or, alternatively, can be connected to an external computer (eg, via the Internet using an Internet service provider).

Claims

A cache warming method, comprising:

Obtain business collection and access information in historical time periods;

Inputting the business collection and access information of the historical time period into the pre-trained business prediction model to obtain the business forecast access information of the target time period;

Inputting the business prediction access information into the pre-trained data prediction model to obtain multiple predicted hot comment values of multiple data in the database of the target time period;

According to the values of the plurality of predicted hotspots, hotspot data is selected from the plurality of data in the database for cache preheating.
The method according to claim 1, before inputting the service forecast access information into the pre-trained data forecasting model to obtain multiple forecasted hot comment values of multiple data in the database of the target time period, further comprising:

Acquiring data training samples, the data training samples including database access information, business access information and multiple detection hot spot values of multiple data in the database;

training the first model according to the data training samples;

When the training of the first model is completed, the first model at the current moment is determined as the data prediction model.
The method according to claim 2, wherein the data training samples further include cache access information.
The method according to claim 3, wherein said acquiring data training samples comprises:

Obtain cache access information, database access information, and service access information in the sampling time period, and determine the value of multiple detection hotspots of multiple data in the database;

A data training sample is formed according to cache access information, data access information, service access information, and multiple detected hot comment values of multiple data in the database during the sampling period.
The method according to claim 4, wherein said determining a plurality of detection hot spot values of a plurality of data in the database comprises:

Inputting the data access information and service access information of the sampling time period into the first model, and determining the value of a plurality of detected hotspots of a plurality of data in the database.
According to the method according to claim 1, before inputting the business collection and access information of the historical time period into the pre-trained business forecasting model to obtain the business forecasting visit information of the target time period, further comprising:

Obtain the business collection and visit information of the first day and the business collection and visit information of the second day, and generate business training samples;

training the second model according to the business training sample;

When the training of the second model is completed, the second model at the current moment is determined as the service prediction model.
The method according to any one of claims 1-6, wherein, according to the plurality of predicted hot comment values, selecting hot data from the plurality of data in the database for cache preheating includes :

Screening out data whose predicted hot comment value is greater than or equal to a preset evaluation value threshold from the plurality of data in the database, and determining the filtered data as hot data;

Store the hotspot data in the cache.
A cache preheating device, comprising:

A business collection and access information acquisition module is configured to obtain business collection and access information in historical time periods;

The business forecast visit information acquisition module is configured to input the business collection visit information of the historical time period into the pre-trained business forecast model to obtain the business forecast visit information of the target time period;

Prediction hot comment value acquisition module is configured to input the business forecast access information into the pre-trained data prediction model to obtain multiple forecast hot comment values of a plurality of data in the database of the target time period;

The cache preheating module is configured to filter out hot data from the plurality of data in the database for cache preheating according to the values of the plurality of predicted hot comments.
A computer device, comprising a memory, a processor, and a computer program stored on the memory and operable on the processor, when the processor executes the program, the cache pre-preservation according to any one of claims 1-7 is realized heat method.
A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the cache warming method according to any one of claims 1-7 is implemented.