A system for improving database performance

Abstract

The present invention relates to a system (1) which enables the time series of usage rates in the subsequent time interval to be estimated by determining the characteristics of the relevant database servers (V) by retrieving data in the form of usage rates, number of active queries, prolongation time of queries, total resource ratio parameters consumed by queries in database servers (V) in certain periods and processing them with an artificial intelligence model; the estimation results to be monitored by authorized persons through an interface (2); authorized persons to perform resource recommendations and, if needed, to perform resource (CPU and RAM) optimization on servers.

Description

[0001] A SYSTEM FOR IMPROVING DATABASE PERFORMANCE

[0002] Technical Field

[0003] The present invention relates to a system which enables the time series of usage rates in the subsequent time interval to be estimated by determining the characteristics of the relevant database servers by retrieving data in the form of usage rates, number of active queries, prolongation time of queries, total resource ratio parameters consumed by queries in database servers in certain periods and processing them with an artificial intelligence model; the estimation results to be monitored by authorized persons through an interface; authorized persons to perform resource recommendations and, if needed, to perform resource (CPU and RAM) optimization on servers.

[0004] Background of the Invention

[0005] Today, fast and efficient data management is of critical importance for the success of many businesses. Databases become increasingly larger and more complex depending on the size and complexity of businesses. And this requires effective management and optimization of database resources. In particular, important resources such as CPU and RAM are the determining factors of database performance. One of the key factors of database performance is the fast and efficient execution of database operations. The CPU is the main component that provides processing power in database management systems. When CPU resources are not assigned correctly, processes may slow down, responses may be delayed, and database performance may decrease. On the other hand, RAM is used for storing temporary data used by the database and increases the speed of database processes. When there is not enough RAM, the database system has to access disk frequently, which significantly affects performance.

[0006] For this reason, considering the studies and deficiencies included in the current technique, it is understood that there is a need for a system which enables CPU and RAM resources to be optimized for improving database performance.

[0007] The Turkish patent document no. TR2023 / 009061, an application included in the state of the art, discloses a virtual server traffic transfer system. The said invention relates to a virtual server traffic transfer system which enables instantaneous energy consumption of virtual servers running in physical servers located in data centers to be measured in order to achieve workforce gain and increase efficiency; virtual server traffic to be moved to empty physical servers and virtual server traffic to be directed by estimating the energy expected to be consumed by its virtual server in a future time with the help of artificial intelligence.

[0008] The Turkish patent document no. TR2021 / 017289, another application included in the state of the art, discloses a resource optimization system. The said invention relates to a system which enables resource optimization to be performed automatically during maintenance operations such as routine patching and repairing performed on virtual servers, if resource optimization can be performed on the virtual servers being worked on by taking advantage of being in the maintenance interval.

[0009] The Turkish patent document no. TR2021 / 017189, another application included in the state of the art, discloses a delay reduction optimization system and method in transmission networks. The said invention is a delay reduction optimization system which enables the telecom infrastructure provider to determine which one(s) of the delay types occurring in their routers in the transmission network are higher than the normal level with the help of the data received from the routers and with the machine learning method; and characterized in that it comprises AYS, which transmits information such as processor and memory usage information of all the processes that the router is currently running and how long it takes to complete the processes to the Delay Reduction Optimization System in Transmission Networks by retrieving them via interfaces, and at the same time transmits the optimization decisions coming from the Delay Reduction Optimization System in Transmission Networks to the routers as commands; management interfaces, which transmit information such as processor and memory usage information of all the processes that the router is currently running and how long it takes to complete the processes to AYS by retrieving them, and at the same time transmit the commands coming from AYS to the routers; Delay Reduction Optimization System in Transmission Networks, which finds the best path that will provide the smallest amount of delay according to the information received from the router and gives a command to the AYS to implement this path in the transmission network for the relevant traffic flow, or sends a command to the AYS to change the usage rate that the router resources in the network allocate to the work, or sends a warning to the system for new resource planning; Analytical Module, which is a module of the Delay Reduction Optimization System in Transmission Networks, generates numerical results for all possible different path options in order to find the best path with the least delay by analyzing the numerical data received from the routers, and generates all possible numerical results that will provide the least delay in the network with existing resources and in case of new resource planning, and shares these numerical values with the Controller Module; Controller Module, which detects the best path where the delay will be the lowest according to the numerical data it receives from the Analytical Module, re-plans the resource utilization planning allocated to each process that will reduce the delay time in the devices, generates a warning to the system for new resource planning in case of insufficient resources and transmits these orders to the AYS.

[0010] The Chinese patent document no. CN112286903, another application included in the state of the art, discloses a relational database optimization method. The said invention describes relational database optimization method and device. The method comprises the steps of obtaining a database access request, carrying out the judgment based on the database access request, and carrying out database optimization through a database buffer according to a judgment result, and monitoring whether the memory or CPU utilization rate of the database buffer is higher than a preset threshold value or not in real time, recording a log and sending a log notification when the memory or CPU utilization rate is higher than the preset threshold value, receiving the log notification and analyzing the log to judge changed data, and performing transverse / longitudinal expansion on the database buffer according to a preset value in configuration.

[0011] The international patent document no. W02023108800, another application included in the state of the art, discloses a performance analysis method based on a CPU-GPU heterogeneous architecture, and a device and a storage medium. In the said invention a performance analysis method based on a CPU-GPU heterogeneous architecture, and a device and a storage medium are provided. The method comprises the steps of acquiring a preset zero-knowledge proof performance analysis model, and calibrating coefficient values in the zeroknowledge proof performance analysis model; inputting, into the zero-knowledge proof performance analysis model, a first data volume, and a plurality of parameter combinations of window sizes and actual thread numbers; selecting the smallest total execution time from a plurality of total execution times output by the zero-knowledge proof performance analysis model, and selecting an optimal parameter combination, which corresponds to the smallest total execution time; and setting parameters of a CPU-GPU heterogeneous architecture on the basis of the optimal parameter combination, wherein the number of parameter combinations is the same as the number of total execution times. In this way, the performance analysis method of the said application performs performance prediction on a zero -knowledge proof algorithm by means of providing a zeroknowledge proof performance optimization analysis model, such that an optimal parameter combination influencing the performance of a CPU-GPU heterogeneous architecture is screened out.

[0012] The Chinese patent document no. CN105701185, another application included in the state of the art, discloses an optimization method for improving database performance. The said invention describes an optimization method for improving database performance. The said method comprises the steps of setting a value of a statistics_level and initializing a parameter; displaying a size of a shared pool suggested by oracle; determining a value of reload-to-pins to determine whether to add the shared pool; determining a value of a miss-ratio to determine whether to add the shared pool; and placing an often accessed object into a buffer cache of a keep type, and placing a seldom accessed large table into a buffer cache of a recycle type, and placing the others into a buffer cache of a default type. In an online state, the method disclosed by the said invention improves an SQL hit rate, improves a cache utilization rate, reduces a CPU utilization rate and generally improves performance of databases.

[0013] Summary of the Invention

[0014] An object of the present invention is to realize a system developed for enabling the time series of usage rates in the subsequent time interval to be estimated by determining the characteristics of the relevant database servers by retrieving data in the form of usage rates, number of active queries, prolongation time of queries, total resource ratio parameters consumed by queries in database servers in certain periods and processing them with an artificial intelligence model; the estimation results to be monitored by authorized persons through an interface; authorized persons to perform resource recommendations and, if needed, to perform resource (CPU and RAM) optimization on servers.

[0015] Detailed Description of the Invention “A System for Improving Database Performance” realized to fulfil the objective of the present invention is shown in the figure attached, in which:

[0016] Figure 1 is a schematic view of the inventive system.

[0017] The components illustrated in the figures are individually numbered, where the numbers refer to the following:

[0018] 1. System

[0019] 2. Interface

[0020] 3. Database

[0021] 4. Server

[0022] V. Database Server

[0023] The inventive system (1) developed for improving database performance comprises at least one interface (2) which is configured to display estimation results in visual / written form to authorized persons; at least one database (3) which is configured to keep a record of RAM and CPU data and detail information collected by using the connection resources of all database servers (V) therein; at least one server (4) which is configured to establish communication with the interface (2) and the database (3) by using any communication protocol; to detect all database servers (V); to collect RAM and CPU data at 10-minute periods; to generate estimations of resource recommendations in the form of “decrease”, “increase”, “no change” for the relevant servers through created model.

[0024] The interface (2) included in the inventive system (1) is configured to establish connection with the server (4). The interface (2) is configured to display resource recommendations for the relevant servers that are transmitted through the server (4).

[0025] The database (3) included in the inventive system (1) is configured to establish connection with the server (4). The database (3) is configured to keep a record of RAM and CPU data therein. The database (3) is configured to keep a record of the resource recommendations for the relevant servers transmitted through the server (4), depending on the date the model is running. The database (3) is configured to keep a record of the detail information collected by using the connection resources of all database servers (V) therein.

[0026] The server (4) included in the inventive system (1) is configured to establish communication with the interface (2) and the database (3) by using any communication protocol included in the state of the art. The server (4) is configured to access data on the database (3). The server (4) is configured to detect all database servers (V). The server (4) is configured to collect RAM and CPU data at 10-minute periods. The server (4) is configured to split the data into three parts as test data, model data and validation data within a certain time interval of one and / or two months in order to create the model that will be estimated. The server (4) is configured to train the model with the model data and to perform behaviour classification. The server (4) is configured to test the model created with the test data before the final stage and to provide relevant estimation recommendations. The server (4) is configured to analyze the RAM and CPU data and to transmit these data to the model structure on a weekly basis. The server (4) is configured to generate estimations of resource recommendations in the form of “decrease”, “increase”, “no change” for the relevant servers through created model; to transmit these estimation data to the database (3), depending on the date the model is running; and to provide them on the interface (2).

[0027] Industrial Application of the Invention In the inventive system (1), the server (4) detects all database servers (V); collects RAM and CPU data at 10-minute periods; and splits the data into three parts as test data, model data and validation data within a certain time interval of one and / or two months in order to create the model that will be estimated. The server (4) trains the model with the model data and performs behaviour classification. The server (4) tests the model created with the test data before the final stage and provides relevant estimation recommendations. The server (4) analyzes the RAM and CPU data and transmits these data to the model structure on a weekly basis. The server (4) generates estimations of resource recommendations in the form of “decrease”, “increase”, “no change” for the relevant servers through created model; transmits these estimation data to the database (3), depending on the date the model is running, and provides them on the interface (2). In this way, the estimation results are enabled to be monitored by authorized persons through an interface; and authorized persons are enabled to perform resource recommendations and, if needed, to perform resource (CPU and RAM) optimization on servers.

[0028] Within these basic concepts; it is possible to develop various embodiments of the inventive “A System (1) for Improving Database Performance”; the invention cannot be limited to examples disclosed herein and it is essentially according to claims.

Claims

CLAIMS1. A system (1) developed for improving database performance comprising at least one interface (2) which is configured to display estimation results in visual / written form to authorized persons; at least one database (3) which is configured to keep a record of RAM and CPU data and detail information collected by using the connection resources of all database servers (V) therein; and characterized by at least one server (4) which is configured to establish communication with the interface (2) and the database (3) by using any communication protocol; to detect all database servers (V); to collect RAM and CPU data at 10-minute periods; to generate estimations of resource recommendations in the form of “decrease”, “increase”, “no change” for the relevant servers through created model.

2. A system (1) according to Claim 1; characterized by the interface (2) which is configured to establish connection with the server (4).

3. A system (1) according to Claim 1 or 2; characterized by the interface (2) which is configured to display resource recommendations for the relevant servers that are transmitted through the server (4).

4. A system (1) according to any one of the preceding claims; characterized by the database (3) which is configured to establish connection with the server (4).

5. A system (1) according to any one of the preceding claims; characterized by the database (3) which is configured to keep a record of RAM and CPU data therein.

6. A system (1) according to any one of the preceding claims; characterized by the database (3) which is configured to keep a record of the resource recommendations for the relevant servers transmitted through the server (4), depending on the date the model is running.

7. A system (1) according to any one of the preceding claims; characterized by the database (3) which is configured to keep a record of the detail information collected by using the connection resources of all database servers (V) therein.

8. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to establish communication with the interface (2) and the database (3) by using any communication protocol9. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to access data on the database (3).

10. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to detect all database servers (V).

11. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to collect RAM and CPU data at 10-minute periods.

12. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to split the data into three parts as test data, model data and validation data within a certain time interval of one and / or two months in order to create the model that will be estimated.

13. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to train the model with the model data and to perform behaviour classification.

14. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to test the model created with the test data before the final stage and to provide relevant estimation recommendations.

15. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to analyze the RAM and CPU data and to transmit these data to the model structure on a weekly basis.

16. A system (1) according to any one of the preceding claims; characterized by the server (4) which is configured to generate estimations of resource recommendations in the form of “decrease”, “increase”, “no change” for the relevant servers through created model; to transmit these estimation data to the database (3), depending on the date the model is running; and to provide them on the interface (2).

Images