WO2017115899A1 - Système de base de données en mémoire ayant une fonction de calcul de données d'objet mobile basée sur un traitement parallèle et procédé pour traiter les données - Google Patents

Système de base de données en mémoire ayant une fonction de calcul de données d'objet mobile basée sur un traitement parallèle et procédé pour traiter les données Download PDF

Info

Publication number
WO2017115899A1
WO2017115899A1 PCT/KR2015/014590 KR2015014590W WO2017115899A1 WO 2017115899 A1 WO2017115899 A1 WO 2017115899A1 KR 2015014590 W KR2015014590 W KR 2015014590W WO 2017115899 A1 WO2017115899 A1 WO 2017115899A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
moving object
object data
query
parallel processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/KR2015/014590
Other languages
English (en)
Korean (ko)
Inventor
정성훈
한혁
진성일
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
RealtimeTech Co Ltd
Original Assignee
RealtimeTech Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by RealtimeTech Co Ltd filed Critical RealtimeTech Co Ltd
Publication of WO2017115899A1 publication Critical patent/WO2017115899A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor

Definitions

  • the present invention accumulates and stores spatio-temporal data of a moving object every hour in an in-memory database system, and when storing the spatiotemporal data corresponding to a query using the stored large-capacity moving object data, the order of storing the moving object data is not important.
  • RDBMS relational database management systems
  • An in-memory database is a database management system that is installed and operated in main memory of data storage, and has an advantage in that processing speed is faster than that installed in disk.
  • information corresponding to a corresponding service can be generated using the stored data.
  • the size of spatiotemporal data corresponding to one moving object may increase from several M bytes to several tens of M bytes.
  • the operation processing time is very long, resulting in a problem of deterioration of service quality.
  • the present invention was created in view of the above circumstances, and when the space-time data operation is performed using a large amount of moving object data for a query in which the order of storing the moving object data is not important, the moving object data can be calculated. It is a technical object of the present invention to provide an in-memory database system and a data processing method having a parallel processing-based mobile object data processing function that can improve the continuous speed by dividing into parallel processing.
  • the database storing the moving object data including the spatiotemporal information on the moving object, and the large-capacity moving object data corresponding to the spatiotemporal query information provided from the query request terminal;
  • the moving object data management module and the large-capacity moving object data provided from the moving object data management module are divided into a plurality of data, and each of the divided data is generated according to a query using a multi-threaded number corresponding to the number of divided data.
  • An object having a parallel processing-based mobile object data calculation processing function comprising: a parallel processing module configured to generate spatiotemporal result information corresponding to a query based on the result of the calculation operation and to provide the moving object data management module.
  • a parallel processing module configured to generate spatiotemporal result information corresponding to a query based on the result of the calculation operation and to provide the moving object data management module.
  • the parallelizable query may include a query capable of performing maximum and minimum distances between different objects or a total length related to an object, a mutual location related query between different objects, or an object location related query at a specific time.
  • An in-memory database system having a parallel processing-based moving object data processing function is provided.
  • the parallel processing module is configured to generate a split data by setting a position at which the change of coordinate information in the mass moving object data differs by a predetermined level or more into a split position, or to split the bulk moving object data by a predetermined size unit.
  • the parallel processing module may be configured to generate the split data by adding a header including the total size of the corresponding data portion and the number of moving object position data included in the data, to the data obtained by dividing the large-capacity moving object data.
  • An in-memory database system having a parallel processing based mobile object data operation processing function is provided.
  • Step 3 for the query capable of parallel processing in the mobile object data management module, in step 4, providing the massive mobile object data called in the second step to the parallel processing module; To generate a plurality of divided data by dividing into arithmetic unit In step 5, the parallel processing module generates multi-threads corresponding to the number of split data, and performs sixth operation on the corresponding split data through each thread, and performs each thread in the parallel processing module. And a seventh step of generating the spatiotemporal result information corresponding to the query based on the result of each partitioned data operation processing processed through the split data operation processing, and providing the result to the moving object data management module.
  • a processing based moving object data operation processing method is provided.
  • the parallel processing query in the third step may include a maximum and minimum distance between different objects or a total length related query for an object, a mutual location related query between different objects, or an object location related query at a specific time.
  • a parallel processing based mobile object data processing method in an in-memory database system is provided.
  • the mobile object data management module determines whether the query is capable of parallel processing only when the size of the large-capacity mobile object data called in the second step is equal to or larger than a preset reference size.
  • the parallel processing module divides the data by dividing the data by setting a position where the change of the coordinate information differs by a predetermined level or more from the large-capacity moving object data or by dividing the data by a certain size unit.
  • the parallel processing module partitions the large amount of moving object data by a predetermined size unit, the parallel processing module sets the number of divided data so as not to exceed the maximum number of threads set corresponding to the number of CPU cores for performing the processing.
  • a parallel processing based moving object data processing method in an in-memory database system is provided.
  • the parallel processing module generates split data by adding a header including the total size of the corresponding data portion and the number of moving object position data included in the data, to the data obtained by dividing the large-capacity moving object data.
  • the large amount of moving object data is divided into a plurality of small data to perform parallel operation processing through multi-threading.
  • FIG. 1 is a schematic diagram of an in-memory database system having a parallel processing based moving object data arithmetic function according to a first embodiment of the present invention.
  • FIG. 2 is a diagram illustrating a process of dividing a large amount of moving object data into a plurality of divided data in the moving object data partitioning block 141 shown in FIG.
  • FIG. 3 is a diagram illustrating a multi-thread operation process for trajectory length calculation of a moving object in the multi-thread operation module 142 shown in FIG. 1.
  • FIG. 4 is a flowchart illustrating a data processing method of an in-memory database system having a parallel processing-based space-time data processing function shown in FIG.
  • FIG. 1 is a diagram illustrating a schematic configuration of an in-memory database system having a parallel processing based mobile object data arithmetic function according to a first embodiment of the present invention.
  • an in-memory database system 100 having a parallel processing-based mobile object data operation processing function includes a database 110, a mobile object data management module 120, and a query processing module ( 130, and a parallel processing module 140.
  • the database 110 is an in-memory database and stores the moving object data provided from the moving object data management module 120.
  • the database 110 includes a moving object table so as to correspond to the moving object, and each moving object table stores moving object data including position information (coordinates) and time information, that is, space-time information.
  • the moving object data management module 120 stores the moving object data provided from the outside in a corresponding moving object table of the database 110.
  • the mobile object data management module 120 calls the mobile object data corresponding to the query information provided from the query processing module 130 in the database 110, and analyzes the query information to perform parallel processing. For, it is configured to provide the called moving object data to the parallel processing unit 140 to provide the spatiotemporal result information provided from the parallel processing unit 140 to the query processing module 130.
  • the query capable of parallel processing is a spatiotemporal data related query in which the order of data is not important, for example, a query relating to the maximum and minimum distances between different objects or the total length (outermost length or trajectory distance) for a specific object, A mutual location related query between different objects or an object location related query of a specific time may be included.
  • the mobile object data management module 120 may include a lookup table including information on whether parallel processing is possible for each query function, and may determine whether a query is capable of parallel processing based on the lookup table.
  • Table 1 below illustrates a lookup table for parallel query function determination.
  • ST_MEETS (A, B) X Evaluate whether the moving space object A meets at the boundary of the moving space object B
  • ST_INSIDES (A, B) X Evaluate whether the moving space object A stays inside the moving space object B.
  • ST_ENVELOPE O Extract the outermost bounding box (MBB) surrounding the moving object
  • ST_SNAPSHOT O Extract location information at specific time
  • the mobile object data management module 120 may perform parallel processing query determination only when the length of the called large object data is greater than or equal to a preset reference length, for example, 128 KB. This is because data below the reference length may have a large parallel processing overhead.
  • a preset reference length for example, 128 KB.
  • the reference length can be changed to be optimized for the system.
  • the query processing module 130 provides the mobile object data management module 120 with the spatiotemporal query information related to the moving object requested from the outside, for example, the query request terminal, and from the moving object data management module 120. It is configured to provide the spatiotemporal result information to the corresponding query request terminal.
  • the parallel processing module 140 divides the large-capacity mobile object data provided from the mobile object data management module 120 into a plurality of data, and each divided data performs parallel arithmetic processing using multi-threads, and each thread. It is configured to generate the spatiotemporal result information corresponding to the query information based on the operation processing result calculated through.
  • the parallel processing module 140 includes a moving object data partitioning block 141, a multithreaded operation block 142, and a parallel processing result calculation block 143.
  • the moving object data partitioning block 141 performs a preprocessing operation for parallel processing of a large amount of moving object data.
  • the moving object data dividing block 141 divides the large-capacity moving object data into units capable of space-time calculation.
  • the size and number of data to be divided may be set in consideration of available resource information of the operating system, for example, CPU performance of operating the system. That is, the mobile object data partitioning block 141 may set the number of cores of the CPU that operates the system to the maximum number of partitioned data in partitioning a large amount of mobile object data.
  • the multi-threaded operation may be reset so as to reset the maximum size of the divided data or to correspond to the number of divided data divided into the maximum size.
  • the maximum number of threads set in block 142 may be reset.
  • the criterion for dividing the large-capacity moving object data may be set to a position where a change in coordinate information differs by a predetermined level or more, or may be set in units of a predetermined size.
  • a maximum size for example, 1MB, may be preset for the divided data, which may be changed and set appropriately by the user.
  • the moving object data partitioning block 141 divides the data portion of the large-capacity moving object data into three first to third data, and a header corresponding to the data corresponding to the first to third data.
  • the header of the divided data includes the total size of the divided data portion, the number information of the position data of moving object included in the data portion, and the total time and coordinate range of the position data included in the data portion.
  • the multi-threaded operation block 142 creates a thread to correspond to the number of divided data generated by the moving object data partitioning block 141, and generates a plurality of pieces of divided data provided from the moving object data partitioning block 141. It is configured to perform spatiotemporal processing using different threads. At this time, the maximum number of threads generated may be set to the number of CPU cores of the calculation module, and may be appropriately changed and set as necessary. In addition, an operation processing program executed in each thread may be provided in response to a query from the mobile object data management module 120.
  • each thread performs the same operation on different partitioned data to generate an operation result for each partitioned data.
  • each thread may calculate the trajectory length of the moving object in the corresponding partitioned data.
  • the split data includes position coordinate information on the corresponding moving object.
  • 3 illustrates a multi-threaded operation process for trajectory length calculation of a moving object. That is, as shown in FIG. 3, the trajectory calculation processing for the corresponding region is performed through the first to third threads on the first to third partitioned data divided by the moving object data partitioning block 141. Each operation result corresponding to the number of threads is generated.
  • each thread may calculate whether coordinates of the moving object are included in a specific area in the corresponding partitioned data.
  • the split data includes position coordinate information on the corresponding moving object.
  • the parallel processing result calculation block 143 uses a result of the calculation information provided from each thread of the multithreaded operation block 142 to obtain a space-time result corresponding to the query requested from the mobile object data management module 120. Generates information and provides it to the moving object data management module 120. For example, the parallel processing result calculation block 143 may calculate a total trajectory length for the corresponding moving object by summing the trajectory length calculation results provided from each thread of the multithreaded operation block 142 or from at least one or more threads. When the moving object coordinates are included in a specific area, the spatiotemporal information of "True" is provided to the moving object data management module 120.
  • the moving object data management module 120 stores the moving object data provided from the outside in the corresponding moving object table in the database 110 (ST10).
  • the moving object data includes position coordinates and time information, that is, space-time information.
  • real-time spatiotemporal information about the moving object is accumulated and stored in each moving object table of the database 110.
  • the query processing module 130 moves the space-time object.
  • the query information is provided to the moving object data management module 120.
  • the moving object data management module 120 calls a large amount of moving object data for the corresponding moving object in the database 110 based on the moving object space-time query information provided from the query processing module 130 (ST30).
  • the moving object data management module 120 analyzes the moving object space-time query information provided from the query processing module 130 and determines whether the query is capable of parallel processing (ST40).
  • the query capable of parallel processing is a query relating to the maximum and minimum distances between different objects or the total length (outermost length or trajectory distance) for a specific object, a query relating to mutual position between different objects, or an object position at a specific time. Include the query.
  • the mobile object data management module 120 provides the parallel processing module 140 with the large amount of mobile object data called in step ST30, when it is determined that the query is a parallel processable query in step ST40. do.
  • the parallel processing module 140 generates a plurality of pieces of divided data by dividing the large-capacity moving object data provided from the moving object data management module 120 into units that can be calculated to correspond to a query (ST50). For example, the parallel processing module 140 may generate the split data by setting a position at which the coordinate change is changed by a predetermined level or more as the data split position with respect to the trajectory length calculation query. In addition, the parallel processing module 40 may divide the data by a unit of a predetermined data size for the query via the specific location. In this case, in the case of dividing data by data size unit, an appropriate number of pieces of divided data may be generated in consideration of the number of cores of a CPU that operates the system.
  • the parallel processing module 140 generates multi-threads corresponding to the number of split data (ST60).
  • the parallel processing module 140 performs arithmetic processing on each partitioned data through each thread (ST70).
  • the parallel processing module 140 generates the spatiotemporal result information corresponding to the query based on the result of each partition data operation processing processed by each thread in step ST70 and provides the same to the moving object data management module 120. (ST80).
  • the mobile object data management module 120 provides the spatiotemporal result information provided from the parallel processing module 140 to the query processing module 130, thereby providing spatiotemporal information corresponding to the query to the query requesting terminal (ST90). .
  • the mobile object data management module 120 determines that the query is a non-parallel query, that is, a query whose order is important, as a result of judging the parallel processing possible query in step ST40, the mass mobile object data called in step ST30. Space-time arithmetic processing is performed directly (ST100).
  • step ST30 when the length of the called large-capacity mobile object data is less than or equal to a preset reference length, the mobile object data management module 120 does not perform the parallelizable query determination and performs the corresponding large-capacity mobile object data. Performs spatiotemporal processing directly on.
  • the data when performing a spatiotemporal operation on a large amount of moving object data, for a query request in which the order of moving object position data is not important, the data is divided into several smaller data to perform parallel processing. It is possible to quickly perform spatiotemporal information processing related to moving objects.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

La présente invention concerne un système de base de données en mémoire ayant une fonction de calcul de données d'objet mobile basée sur un traitement parallèle et un procédé pour traiter les données, pour accumuler et sauvegarder, dans le système de base de données en mémoire, des données spatio-temporelles suivant le déplacement d'un objet mobile toutes les heures, et diviser et traiter au moyen d'un traitement parallèle des données d'objet mobile massives correspondant à une interrogation dont l'ordre n'est pas important, lors du calcul des données spatio-temporelles correspondant à l'interrogation à l'aide des données d'objet mobile massives qui sont sauvegardées, permettant ainsi d'améliorer la vitesse de traitement de calcul spatio-temporelle.
PCT/KR2015/014590 2015-12-30 2015-12-31 Système de base de données en mémoire ayant une fonction de calcul de données d'objet mobile basée sur un traitement parallèle et procédé pour traiter les données Ceased WO2017115899A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR10-2015-0189450 2015-12-30
KR20150189450 2015-12-30

Publications (1)

Publication Number Publication Date
WO2017115899A1 true WO2017115899A1 (fr) 2017-07-06

Family

ID=59224931

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2015/014590 Ceased WO2017115899A1 (fr) 2015-12-30 2015-12-31 Système de base de données en mémoire ayant une fonction de calcul de données d'objet mobile basée sur un traitement parallèle et procédé pour traiter les données

Country Status (1)

Country Link
WO (1) WO2017115899A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112732852A (zh) * 2020-12-31 2021-04-30 武汉大学 一种跨平台的时空大数据分布式处理方法及软件
CN116701446A (zh) * 2023-05-23 2023-09-05 浪潮云信息技术股份公司 一种数据处理的多线程优化方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120131584A1 (en) * 2009-02-13 2012-05-24 Alexey Raevsky Devices and Methods for Optimizing Data-Parallel Processing in Multi-Core Computing Systems
US20140019984A1 (en) * 2012-07-11 2014-01-16 Sap Ag Feedback-driven tuning for efficient parallel execution
US20150197248A1 (en) * 1997-10-22 2015-07-16 American Vehicular Sciences Llc Vehicle speed control method and arrangement

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150197248A1 (en) * 1997-10-22 2015-07-16 American Vehicular Sciences Llc Vehicle speed control method and arrangement
US20120131584A1 (en) * 2009-02-13 2012-05-24 Alexey Raevsky Devices and Methods for Optimizing Data-Parallel Processing in Multi-Core Computing Systems
US20140019984A1 (en) * 2012-07-11 2014-01-16 Sap Ag Feedback-driven tuning for efficient parallel execution

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"Kairos User Manual Release 5.0", 2009, Retrieved from the Internet <URL:http://www.realtimetech.co.kr/download/?pageid=3&uid=157&mod=document> *
SHUAI LIAO: "A SPATIOTEMPORAL AGGREGATION QUERY METHOD USING MULTI- THREAD PARALLEL TECHNIQUE BASED ON REGIONAL DIVISION", ISPRS INTERNATIONAL WORKSHOP ON SPATIOTEMPORAL COMPUTING, vol. II-4/W2, 13 July 2015 (2015-07-13), Fairfax, Virginia, USA, pages 1 - 5, XP055599591, ISSN: 2194-9050, Retrieved from the Internet <URL:http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-4-W2/1/2015/isprsannals-II-4-W2-1-2015.pdf> DOI: 10.5194/isprsannals-II-4-W2-1-2015 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112732852A (zh) * 2020-12-31 2021-04-30 武汉大学 一种跨平台的时空大数据分布式处理方法及软件
CN116701446A (zh) * 2023-05-23 2023-09-05 浪潮云信息技术股份公司 一种数据处理的多线程优化方法

Similar Documents

Publication Publication Date Title
US20140331235A1 (en) Resource allocation apparatus and method
CN113568940A (zh) 数据查询的方法、装置、设备以及存储介质
CN114416352A (zh) 算力资源分配方法、装置、电子设备及储存介质
JP6421247B2 (ja) マルチコアシステムのメモリ同期のためのシステム及び方法
CN108776934A (zh) 分布式数据计算方法、装置、计算机设备及可读存储介质
US20170228190A1 (en) Method and system providing file system for an electronic device comprising a composite memory device
CN112416960A (zh) 多场景下的数据处理方法、装置、设备及存储介质
US12079472B2 (en) Data reduction method, apparatus, computing device, and storage medium for forming index information based on fingerprints
US20180232889A1 (en) Generating and reviewing motion metadata
CN102915344B (zh) 一种sql语句处理方法及装置
CN106294815B (zh) 一种url的聚类方法及装置
WO2023005771A1 (fr) Procédé et dispositif d&#39;interrogation de piste, support de stockage et produit-programme informatique
CN102968491A (zh) 一种数据分配方法和装置
CN116860662A (zh) 冷热内存页的识别方法、内存回收方法及其相关设备
US20220375185A1 (en) Method of recognizing image, electronic device, and storage medium
WO2017115899A1 (fr) Système de base de données en mémoire ayant une fonction de calcul de données d&#39;objet mobile basée sur un traitement parallèle et procédé pour traiter les données
CN110851249B (zh) 一种数据导出的方法及设备
JPWO2012066620A1 (ja) 情報処理装置、情報処理プログラム、および情報処理方法
CN116244380A (zh) 数据库的同步事务的处理方法、存储介质与计算机设备
CN109976809B (zh) 调度方法及相关装置
CN108595367B (zh) 一种基于局域网内计算机集群的服务器系统
CN117992857A (zh) 一种热页和冷页的识别方法、装置、设备及存储介质
CN117437579A (zh) 视频处理方法、装置、设备及介质
CN111160283A (zh) 一种数据接入方法、装置、设备和介质
KR20250050311A (ko) 강화학습 기반 이기종 멀티코어 프로세서 스케줄링 장치 및 방법

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15912135

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15912135

Country of ref document: EP

Kind code of ref document: A1