TR201907249A2

TR201907249A2 - A LOAD DISTRIBUTION SYSTEM

Info

Publication number: TR201907249A2
Application number: TR2019/07249A
Authority: TR
Inventors: Alkent Samet; Yüksel Ahmet; Sayi Berna; Kalkan Yunus
Original assignee: Turkcell Technology Research And Development Co; Turkcell Teknoloji̇ Araştirma Ve Geli̇şti̇rme Anoni̇m Şi̇rketi̇
Priority date: 2019-05-14
Filing date: 2019-05-14
Publication date: 2019-06-21

Abstract

Bu buluş, veri tabanı seviyesinde çalıştırılan işlerde çeşitli makine öğrenme algoritmaları ile işlenerek gerçekleştirilen işlerin sistem üzerinde oluşturduğu yükün tahminlenmesi ile yükün zaman bazında dağıtılarak sistem performansının artırılması ve tıkanmanın önlenmesini sağlayan bir sistem (1) ile ilgilidir.This invention relates to a system (1) that increases system performance and prevents blockage by estimating the load created on the system by the works performed by processing with various machine learning algorithms in jobs operated at database level, and distributing the load on time basis.

Description

TARIFNAME BIR YÜK DAGITIM SISTEMI Teknik Alan Bu bulus, veri tabani seviyesinde çalistirilan islerde çesitli makine ögrenme algoritmalari ile islenerek gerçeklestirilen islerin sistem üzerinde olusturdugu yükün tahminlenmesi ile yükün zaman bazinda dagitilarak sistem performansinin artirilmasi ve tikanmanin önlenmesini saglayan bir sistem ile ilgilidir. Önceki Teknik Bir anlik veri toplama sisteminde belirli zaman dilimlerinde gerçeklestirilen veri toplama islemi sistem üzerinde bir yük yaratmaktadir. Yük yaratan islemler sistem üzerinde tikaniklik ve performans problemlerine yol açmaktadir. Teknigin bilinen durumunda yük yaratan islemlerin yapay zekâ teknolojisi kullanilarak tespit edilerek sistemin daha performansli çalismasi ve tikanikliklarin engellenmesini saglamak üzere olusan yükün zaman dilimlerine dagitilmasini saglayan bir sistem bulunmamaktadir. DESCRIPTION A LOAD DISTRIBUTION SYSTEM Technical Field This invention applies to various machine learning tasks that run at the database level. The work carried out by processing with algorithms creates on the system By estimating the load, the system performance is increased by distributing the load on a time basis. It is about a system that increases the flow rate and prevents clogging. Prior Art Data collected at specific time periods in an instantaneous data collection system The collection process creates a burden on the system. Load creating operations system It causes congestion and performance problems. Your technique is known The processes that create a burden in the event of a situation are detected using artificial intelligence technology. by ensuring that the system operates more efficiently and blockages are prevented. A system that allows the load to be distributed over time periods to ensure There is no.

Mevcut teknikte bulunan çalismalar ve eksiklikler göz önünde bulunduruldugunda veri tabani seviyesinde çalistirilan islerde çesitli makine ögrenme algoritmalari ile islenerek gerçeklestirilen islerin sistem üzerinde olusturdugu yükün tahminlenmesi ile yükün zaman bazinda dagitilarak sistem performansinin artirilmasi ve tikanmanin önlenmesini saglamak amaciyla yapilandirilan bir sisteme ihtiyaç duyuldugu anlasilmaktadir. Considering the studies and deficiencies in the current technique with various machine learning algorithms in jobs run at the database level. The load created by the work carried out on the system system performance by distributing the load on a time basis by estimating A system designed to increase the flow rate and prevent clogging. It is clear that the system is needed.

Teknigin bilinen durumunda yer alan USlOO78569 sayili Birlesik Devletler patent dokümaninda bahsi geçen bulus I/O istatistikleri için ARIMA (auto-regressive- integrated-moving average) modeli kullanilmasi ile elde edilen tahmini verilerin kullanilmasini gerçeklestiren bir veri depolama optimizasyon yönteminden bahsetmektedir. USlOO78569 sayili Birlesik Devletler patent dokümaninda bahsi geçen ARIMA modeli gelecekteki yük talebinin tahminlenmesi için I/O erisiminin periyodik örüntülerini veya is yükü trendlerini yakalamak amaciyla kullanilmaktadir. USlOO78569 sayili Birlesik Devletler patent dokümaninda bahsi geçen I/O istatistikleri anlik zaman periyodunda toplanmaktadir, güncel veri grubu ve en az bir ARIMA modeli kullanilarak bir sonraki zaman periyodu için bir tahminlenmis I/O istatistikleri veri grubu belirlenmektedir. USlOO78569 sayili Birlesik Devletler patent dokümaninda bahsi geçen I/O istatistikleri anlik zaman periyodunda toplanmaktadir, güncel veri grubu ve en az bir ARIMA modeli kullanilarak bir sonraki zaman periyodu için bir tahminlenmis I/O istatistikleri veri grubu belirlenmektedir. United States patent number US10078569, which is in the state of the art The invention mentioned in the document is ARIMA (auto-regressive-) for I/O statistics. Estimated data obtained by using the integrated-moving average) model A data storage optimization method that uses mentions. Mentioned in United States patent document No. US10078569 The previous ARIMA model uses I/O access to predict future load demand. in order to capture periodic patterns or workload trends is used. Mentioned in United States patent document No. US10078569 Last I/O statistics are collected in an instantaneous time period, current data for the next time period using the group and at least one ARIMA model. A data set of estimated I/O statistics is determined. No. US10078569 I/O statistics mentioned in the United States patent document are instantaneous period, current data set and at least one ARIMA model A predicted I/O statistics for the next time period using The data group is determined.

Bulusun Kisa Açiklamasi Bu bulusun amaci veri tabani seviyesinde çalistirilan islerde çesitli makine ögrenme algoritmalari ile islenerek gerçeklestirilen islerin sistem üzerinde olusturdugu yükün tahminlenmesi ile yükün zaman bazinda dagitilarak sistem performansinin artirilmasi ve tikanmanin önlenmesini saglayan bir sistem gerçeklestirmektir. Brief Description of the Invention The purpose of this invention is to use various machine tools in jobs run at the database level. The work done by learning algorithms is processed on the system. By estimating the load it creates, the system is distributed by distributing the load on a time basis. A system that improves performance and prevents clogging is to realize.

Bulusun Ayrintili Açiklamasi Bu bulusun amacina ulasmak için gerçeklestirilen “Bir Yük Dagitim Sistemi” ekli sekilde gösterilmis olup, bu sekil; Sekil-1 Bulus konusu sistemin sematik blok diyagramidir. Detailed Description of the Invention In order to achieve the purpose of this invention, "A Load Distribution System" is attached. It is shown in the figure and this figure; Figure-1 is the schematic block diagram of the system subject to the invention.

Sekilde yer alan parçalar tek tek numaralandirilmis olup, bu numaralarin karsiliklari asagida verilmistir. 1. Sistem 2. Isleri çalistiran (birincil) sistem 3. Islerin çalistigi (ikincil) sistem 4. Sunucu Bulus konusu, veri tabani seviyesinde çalistirilan islerde çesitli makine ögrenme algoritmalari ile islenerek gerçeklestirilen islerin sistem üzerinde olusturdugu yükün tahininlenmesi ile yükün zaman bazinda dagitilarak sistem performansinin artirilmasi ve tikanmanin önlenmesini saglayan saglamak üzere yapilandirilan bir sistem (1); - isleri çalistiran (birincil) sistem (2), - islerin çalistigi (ikincil) sistem (3), - isleri çalistiran (birincil) sistemin (2) gerçeklestirdigi veri toplama isleminin islerin çalistigi (ikincil) sistem (3) üzerinde olusturacagi yükün tahminlenmesi ve yükün zaman dilimlerine dagitilmasini saglamak üzere yapilandirilan en az bir sunucu (4) içermektedir. The parts in the figure are numbered one by one, and these numbers The equivalents are given below. 1. System 2. The (primary) system that runs things 3. The (secondary) system in which jobs run 4. Server The subject of the invention is the use of various machines in jobs run at the database level. The work done by learning algorithms is processed on the system. By estimating the load it creates and distributing the load on a time basis, the system providing increased performance and prevention of clogging A system configured to (1); - the (primary) system that runs the jobs (2), - the (secondary) system in which the jobs run (3), - data collection performed by the (primary) system (2) that runs the jobs The load that the process will create on the (secondary) system (3) where the jobs run to ensure estimation and distribution of the load over time periods. It contains at least one configured server (4).

Bulus konusu sistemde (1) yer alan isleri çalistiran (birincil) sistem (2) çesitli dis kaynaklardan veri toplama Ve depolama islemleri gerçeklestirmek üzere yapilandirilan bir sistemdir. Isleri çalistiran (birincil) sistem (2) sunucu (4) ile temas halindedir. Isleri çalistiran (birincil) sistem (2) islerin baslama ve bitis tarihleri gibi yük tahminlenmesinde kullanilacak zaman boyutunu bir girdi olarak sunucuya (4) saglamak üzere yapilandirilmaktadir. The (primary) system that runs the jobs in the system (1) that is the subject of the invention (2) includes various external To perform data collection and storage operations from sources It is a structured system. The (primary) system that runs the jobs (2) with the server (4) is in contact. The (primary) system that runs the jobs (2) starts and ends the jobs The time dimension to be used in load estimation, such as dates, as an input It is configured to provide the server (4).

Bulus konusu sistemde (1) yer alan islerin çalistigi (ikincil) sistem (3) isleri çalistiran (birincil) sistem (2) tarafindan çalistirilmak üzere yapilandirilmaktadir. The (secondary) system (3) jobs in which the jobs in the system (1) that are the subject of the invention run It is configured to be run by the operating (primary) system (2).

Islerin çalistigi (ikincil) sistem (3) sunucu (4) ile temas halindedir. Islerin çalistigi (ikincil) sistem (3) sürekli olarak sunucuya (4) veri saglamaktadir. The (secondary) system (3) where the jobs run is in contact with the server (4). things are working The (secondary) system (3) constantly provides data to the server (4).

Bulus konusu sistemde (1) yer alan sunucu (4) islerin çalistigi (ikincil) sistem (3) üzerinden sürekli olarak veri okumak üzere yapilandirilmaktadir. Sunucu (4) isleri çalistiran (birincil) sistem (2) üzerinden islerin baslama ve bitis tarihleri gibi bir Zaman boyutu bildiren veri edinmek üzere yapilandirilmaktadir. Sunucu (4) islerin çalistigi (ikincil) sistem (3) ve isleri çalistiran (birincil) sistem (2) üzerinden edinilen verilerin islenmesi ile isleri çalistiran (birincil) sistemin (2) islerin çalistigi (ikincil) sistem (3) üzerinde olusturacagi yükün tahminlenmesini saglamak üzere yapilandirilmaktadir. Sunucu (4) yapay zekâ teknolojileri kullanilarak islerin gerçeklestirilecegi (ikincil) sistem (3) üzerinde eylemin gerçeklestirilmesi için en iyi zamansal tahminin veya en iyi harekete geçme zamaninin belirlenmesini saglamak üzere yapilandirilmaktadir. Sunucu (4) islerin çalistigi (ikincil) sistem (3) üzerinden olusan yük durum bilgisine dair bir geri dönüs bilgisi almak ve islemek üzere yapilandirilmaktadir. Bulusun tercih edilen uygulamasinda sunucu (4) makine ögrenme algoritmalari kullanilarak Ögrenen bir yapi olusturmak üzere yapilandirilmaktadir. Sunucu (4) islerin çalistigi (ikincil) sistemdeki (3) yükün optimize edilmesini saglamak üzere islerin çalistigi (ikincil) Sistem (3) ve isleri çalistiran (birincil) sistem (2) üzerinden veri toplamak ve topladigi verileri islemek üzere yapilandirilmaktadir. The server (4) in the system (1) that is the subject of the invention is the (secondary) system (3) where the jobs run. It is configured to continuously read data from it. Server (4) names A system such as the start and end dates of the jobs is sent through the operating (primary) system (2). It is configured to obtain data that reports the time dimension. Server (4) jobs through the (secondary) system on which it runs (3) and the (primary) system (2) that runs the jobs. The (primary) system (2) that runs the jobs by processing the acquired data Estimating the load it will create on the (secondary) system (3) on which it operates It is configured to provide Server (4) artificial intelligence technologies action on the (secondary) system (3) where the work will be carried out using the best time estimate or best action to take It is configured to enable time determination. Server (4) jobs It provides feedback regarding the load status information via the (secondary) system (3) on which it operates. It is configured to receive and process return information. Find the preferred In the application, the server (4) is a learning device using machine learning algorithms. It is structured to create a structure. Server (4) where jobs are running (secondary) (secondary) where jobs run to ensure that the load on the system (3) is optimized Collecting data through the system (3) and the (primary) system (2) that runs the jobs and It is configured to process the data it collects.

Bulus konusu sistemin (1) çok çesitli uygulamalarinin gelistirilmesi mümkün olup, bulus burada açiklanan örneklerle sinirlandirilamaz, esas olarak istemlerde belirtildigi gibidir. It is possible to develop various applications of the system (1) subject to the invention. and the invention cannot be limited to the examples described herein, it is essentially stated in the claims. as stated.

Claims

CLAIMS: This invention is related to a system (1) that increases system performance and prevents clogging by estimating the load created by the jobs executed on the system by processing the jobs run at the database level with various machine learning algorithms and distributing the load on a time basis. CLAIMS . It is configured to increase system performance and prevent clogging by distributing the load on a time basis by estimating the load that the jobs created on the system by processing the jobs run at the database level with various machine learning algorithms; - estimating the load that the data collection process performed by the (primary) system (2) that runs the jobs (including the (primary) system (2), - the (secondary) system (3) where the jobs run)) will create on the (secondary) system (3) where the jobs run (3). and a system (1) characterized by at least one server (4) configured to ensure the distribution of the load across time periods. . A system (1) as in Claim 1, characterized by the (primary) system (2) that runs the jobs, which is a system configured to perform data collection and storage operations from various external sources. . A system (1) as in Claim 1 or Claim 2, characterized by the (primary) system (2) that runs the jobs in contact with the server (4). . A system (1) as in any of the above claims, characterized by the (primary) system (2) that runs the jobs configured to provide the time dimension to be used in load estimation, such as the start and end dates of the jobs, to the server (4) as an input. . A system (1) as in any of the above claims, characterized by the (secondary) system (3) in which the jobs configured to be run by the (primary) system (2) run the jobs. A system as in any of the above claims, characterized by the (secondary) system (3) in which the jobs that are in contact with the server (4) run. a system like any other (1). A system (1) as in any of the above claims, characterized by the server (4) configured to continuously read data from the (secondary) system (3) on which the jobs run. A system as in any of the above claims, characterized by the server (4) configured to obtain data indicating a time dimension such as the start and end dates of the jobs, through the (primary) system (2) that runs the jobs, and the (secondary) system (3) where the jobs run and the server (4) that runs the jobs. A system as in any of the above claims, characterized by the server (4) configured to process the data obtained through the (primary) system (2) and to estimate the load that the (primary) system (2) that runs the jobs will create on the (secondary) system (3) where the jobs run. system (1). A system (1) as in any of the above claims, characterized by the server (4) configured to ensure the determination of the best temporal estimate or the best action time for performing the action on the (secondary) system (3) where the work will be carried out using artificial intelligence technologies. A system as in any of the above claims, characterized by the server (4) configured to receive and process a return information regarding the load status information obtained through the (secondary) system (3) where the jobs are running. The server configured to create a learning structure using machine learning algorithms ( A system (1) as in any of the above claims, characterized by 4). The above is characterized by the server (4) configured to collect data from the (secondary) system (3) where the jobs run and the (primary) system (2) that runs the jobs and to process the data it collects, in order to optimize the load on the (secondary) system (3) where the jobs run. A system as in any of the claims (1).