WO2017022005A1

WO2017022005A1 - Database management device and method

Info

Publication number: WO2017022005A1
Application number: PCT/JP2015/071762
Authority: WO
Inventors: 圭弘野田; 雄森中; 隆一星野; 憲宏原
Original assignee: 株式会社日立製作所
Priority date: 2015-07-31
Filing date: 2015-07-31
Publication date: 2017-02-09

Abstract

Provided are a database management device and method for managing the data arrangement of a database held in a storage apparatus in which a hierarchical storage function is installed, wherein the device and method are configured so as to: determine, for each process, one data arrangement plan that satisfies a condition pertaining to a time prescribed in advance regarding the process from among a plurality of data arrangement plans for arranging some of the data of the database that is frequently accessed in the storage area of an upper hierarchy, the one data arrangement plan being determined as a final data arrangement plan; determine whether the capacity of the storage area of the upper hierarchy is sufficient when the data of the database is rearranged in accordance with the determined final data arrangement plan for each process; and cause the storage apparatus to rearrange the data of the database in accordance with the determined final data arrangement plan for each process when the capacity of the storage area of the upper hierarchy is sufficient.

Description

Database management apparatus and method

The present invention relates to a database management apparatus and method, and is particularly suitable when applied to a database management apparatus that manages a database held in a storage apparatus equipped with a hierarchical storage function.

Conventionally, a hierarchical storage function is widely known as one of the functions of a storage device (see, for example, Patent Document 1 below). The hierarchical storage function is a function that appropriately changes the data placement destination so that data with a higher access frequency or data with a specific data attribute is placed in a storage device with higher response performance based on the data access frequency and data attributes It is.

In practice, a storage apparatus equipped with such a hierarchical storage function is mounted with a plurality of types of storage devices with different response performance, and manages storage areas provided by storage devices with higher response performance as higher-level storage areas. . Then, the storage apparatus appropriately moves data so as to store data having a higher access frequency or data attributes having a higher importance in a storage area of a higher hierarchy.

According to such a hierarchical storage function, data can be arranged in an appropriate storage device according to access frequency or data attribute, so that the response performance of the entire system is improved while suppressing the cost of building the system. There are advantages that can be made.

Japanese Patent No. 5070315

By the way, when a database is stored in a storage apparatus equipped with the hierarchical storage function as described above, there is a problem that the determination and change of the arrangement hierarchy in the logical structure unit (that is, “table” unit) of the database is inefficient. . This is because there are data with high access and data with low access in the logical structure of the database. For example, if database data is moved to a higher hierarchy in the logical structure unit, data that is rarely accessed This is because the storage area of the uppermost hierarchy is unnecessarily used as a result.

In addition, according to the existing hierarchical storage function that changes the data location based only on the access frequency and data attributes, the quality assurance contract concluded with the customer (hereinafter referred to as SLA (Service Level Agreement)) There is a problem that optimal data placement is difficult.

The present invention has been made in view of the above points, and intends to propose a database management apparatus and method capable of appropriately managing a database held by a storage apparatus equipped with a hierarchical storage function.

In order to solve such a problem, in the present invention, data having a higher access frequency is moved to a higher-level storage area having a higher response performance, and data other than the data is lower-level hierarchy having a lower response performance than the first storage area In a database management apparatus that manages a database held in a storage apparatus having a hierarchical storage function to be moved to a storage area, a process execution unit that executes a process related to the database, and session information at the time of execution of each of the processes A session management unit that obtains and manages the data, and a data placement management unit that causes the storage device to rearrange the data of the database as needed based on the session information of each of the processes, The data placement management unit accesses the database data for each process. Of the plurality of data arrangement plans for allocating a part of the data having a large number in the storage area of the upper hierarchy, one data arrangement plan that satisfies a condition relating to the time defined in advance for the process is selected as the final data of the process. Each data storage plan is determined, and when the data of the database is rearranged according to the final data allocation plan for each of the determined processes, it is determined whether the storage area of the upper hierarchy is sufficient. When the storage area of the upper hierarchy is sufficient, the storage apparatus is made to relocate the data of the database according to the final data arrangement plan for each determined process.

In the present invention, data having a higher access frequency is moved to a higher-level storage area having a higher response performance, and data other than the data is moved to a lower-level storage area having a lower response performance than the first storage area. A database management method that is executed in a database management device that manages a database held in a storage device having a hierarchical storage function to be executed, wherein the database management device includes a processing execution unit that executes processing related to the database; A session management unit that acquires and manages session information at the time of execution of each process, and causes the storage apparatus to relocate data in the database as necessary based on the session information of each process A data placement management unit, and the data placement management unit One piece of the data that satisfies a condition relating to a time defined in advance for the processing from among a plurality of data placement plans in which a part of the data in the database having a large number of accesses is placed in the upper-layer storage area A first step of respectively determining an arrangement plan as a final data arrangement plan of the processing, and the data arrangement management unit rearranges the data in the database according to the final data arrangement plan for each of the determined processes The second step of determining whether or not the capacity of the upper-layer storage area is sufficient, and when the data placement management unit determines that the capacity of the upper-layer storage area is sufficient A third step of causing the storage device to rearrange the data in the database according to the final data allocation plan for each It was to be provided.

According to the database management apparatus and method, only data that satisfies the time conditions specified for each process can be relocated to the upper-layer storage area. In comparison with this, it is possible to more efficiently use the storage area of the upper hierarchy, and it is possible to perform hierarchical storage of data in the database so as to satisfy the time conditions defined for each process.

According to the present invention, it is possible to realize a database management apparatus and method capable of appropriately managing a database held by a storage apparatus equipped with a hierarchical storage function.

It is a block diagram which shows the whole structure of the database system by this Embodiment. It is a conceptual diagram with which it uses for description of the data arrangement management function by this Embodiment. It is a graph with which it uses for description of the data arrangement management function by this Embodiment. It is a conceptual diagram with which it uses for description of a session information management table group. It is a conceptual diagram which shows the structure of a session information management table. It is a conceptual diagram which shows the structure of an importance management table. It is a conceptual diagram which shows the structure of a SLA management table. It is a flowchart which shows the process sequence of a data arrangement management process. It is a flowchart which shows the process sequence of a data arrangement plan determination process. It is a flowchart which shows an example of the process sequence of an executable determination process. It is a flowchart which shows the other example of the process sequence of an execution availability determination process.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

(1) Configuration of Database System According to this Embodiment In FIG. 1, reference numeral 1 denotes a database system according to this embodiment as a whole. This database system includes a database server device 2 and a storage device 3.

The database server device 2 is a server device having a function of executing online processing in response to a request from an external device (not shown) and batch processing scheduled in advance. The database server device 2 includes a processor 10, a memory 11, a disk interface 12, and a communication interface 13.

The processor 10 is hardware that controls the operation of the entire database server device 2, and is composed of, for example, a CPU (Central Processing Unit). The memory 11 is composed of, for example, a semiconductor memory and is used to hold various programs and various data. When the CPU 10 executes the program stored in the memory 11, online processing, batch processing, and various types of processing described below are executed. A data management unit 20, a process execution unit 21, a session management unit 22, a data arrangement management unit 23, a session information management table group 24, an importance management table 25, and an SLA management table 26, which will be described later, are also stored and held in the memory 11. The

The disk interface 12 is composed of, for example, a fiber channel interface, and executes protocol control when transmitting / receiving database data to / from the storage apparatus 3.

The communication interface 13 is composed of, for example, a NIC (Network Interface Card), and is connected to the storage device 3 via the network 4. The communication interface 13 instructs a database operation command for instructing insertion, update or deletion of data issued from the database server apparatus 2 to the storage apparatus 3 at the time of the above online processing or batch processing, or a search on the database. Protocol control at the time of communication of commands transmitted / received to / from the storage apparatus 3 other than the search command is executed.

The storage device 3 is composed of a storage device such as RAID (Redundant Arrays of Inexpensive Disks) equipped with a plurality of types of storage devices each having a different response speed, and is used for storing and holding a database managed by the database server device 2 Is done. The storage device 3 includes a storage unit 30 and a controller 31.

The storage unit 30 is used to store and hold a database, and a high-speed storage device 40 such as an SSD (Solid State Drive) that can read and write data at high speed, and a hard disk that is slower in reading and writing data than the high-speed storage device 40 And a low-speed storage device 41 such as a device (for example, SATA (Serial Advanced Technology Attachment)).

A RAID group is configured by one or a plurality of storage devices of the same type (high-speed storage device 40 or low-speed storage device 41), and one or a plurality of logical volumes are respectively defined on storage areas provided by each RAID group. . Data of the database managed by the database server device 2 is distributed and stored in these logical volumes.

The controller 31 has a function of controlling reading and writing of data with respect to the storage unit 30, and includes a processor 42 and a memory 43. The processor 42 is hardware that controls the operation of the entire storage apparatus 3, and includes, for example, a CPU. The memory 43 is constituted by a semiconductor memory, for example, and is used for temporarily holding various programs and various data.

The memory 43 stores a hierarchy management unit 44, and when the CPU executes the hierarchy management unit 44, logical data formed on a storage area provided by the high-speed storage device 40 with data having a higher access frequency is stored. A logical volume (hereinafter referred to as the lower layer, as appropriate) formed on the storage area provided by the low-speed storage device 41 is moved to the volume (hereinafter referred to as the upper level storage area as appropriate) Data management is performed by a hierarchical storage system that is moved to a storage area.

(2) Data Placement Management Function According to this Embodiment Next, the data placement management function implemented in the database server device 2 will be described. In this database server device 2, in order to satisfy the time condition defined by the SLA for online processing or batch processing regarding online processing that occurs 24 hours and batch processing that occurs periodically, any block of the database A data placement plan (hereinafter referred to as a data placement plan) for deciding in which storage area the data of which data is to be placed is determined for each online process and each batch process, and according to the determined data placement plan, A data arrangement management function for operating the storage apparatus 3 to relocate necessary data to a storage area of an upper hierarchy or a lower hierarchy is installed.

“Online processing” here refers to processing for inserting, updating, or deleting data in the database in response to an operation command for the database from the external device, or the search command in response to a search command from the external device This is a process for searching the data stored in the storage apparatus 3 for data satisfying the conditions specified in (1) and transmitting the search result to the transmission source of the search command. “Batch processing” refers to processing that occurs periodically, such as tabulation processing.

Usually, for online processing, the SLA for response time is preliminarily concluded with the user as a time condition, and for batch processing, time is set so that processing can be completed in the time zone such as nighttime that does not affect online processing. As a condition, the SLA for the processing time is concluded in advance with the user. The data arrangement management function according to the present embodiment performs online processing and batch processing of data arrangement proposals that can comply with the time conditions (response time and processing time) defined by the SLA of each online processing and each batch processing. This is a function for controlling the storage apparatus 3 so as to rearrange necessary data according to the decided data arrangement plan and rearrange the necessary data in the storage area of the upper hierarchy or the lower hierarchy.

Specifically, when executing online processing or batch processing, the database server device 2 executes a plurality of SQLs (Structured Query Query Language) corresponding to the online processing or batch processing, but each time the SQL is executed, Information indicating how many times each block of the database has been accessed and the average value of response times for these blocks (hereinafter referred to as average response time) is acquired and stored as session information.

Then, the database server apparatus 2 changes the configuration of the storage apparatus 3 such as when a new high-speed storage device 40 is added, or periodically thereafter, based on session information for each SQL stored at that time. For each online process and batch process, the storage area in which the high-speed storage device 40 provides the data of the upper several percent of the blocks in the database that are frequently accessed in the online process or batch process (that is, the storage area in the upper hierarchy) ), The estimated values of the processing time of the online processing or batch processing are respectively calculated.

At this time, the database server device 2 sets the ratio of the data to be arranged in the storage area of the upper hierarchy, as shown in FIG. 2, in the upper 10%, 20%, 30%,... As shown in FIG. 3, the processing time required for online processing or batch processing for each predetermined ratio as shown in FIG. 3 is sequentially changed to a plurality of predetermined ratios (hereinafter referred to as predetermined ratios). Are respectively calculated. In addition, in the following, each of the plans for placing the upper 10%, 20%, 30%,..., Block data with the highest number of accesses in the upper-tier storage area (the default ratio is 10%, 20%). %, 30%, and so on) are called data placement plans.

The database server device 2 is defined by the SLA concluded for the online processing or batch processing based on the estimated value of the processing time for each data arrangement plan calculated as described above for each online processing and batch processing. The data arrangement plan that satisfies the time condition and uses the least amount of storage space in the high-speed storage device 40 (that is, the data arrangement plan that has the smallest value of the default ratio) is the final data arrangement plan of the online processing or batch processing. Decide each.

Then, the database server device 2 determines whether or not the capacity of the high-speed storage device 40 is sufficient when data rearrangement is performed in accordance with the data allocation plan for each online process and batch process, and determines that the capacity is sufficient. If it is obtained, the storage apparatus 3 is caused to perform data rearrangement in accordance with the data arrangement plan of each online process and each batch process.

As described above with reference to FIG. 1, the memory 11 of the database server device 2 as a means for realizing the data arrangement management function according to the present embodiment as described above includes the data management unit 20, the process execution unit 21, and the session management. A unit 22, a data arrangement management unit 23, a session information management table group 24, an importance management table 25, and an SLA management table 26 are stored.

The data management unit 20 is a program having a function of managing the table structure of the database stored in the storage device 3 and the storage location (address) of each data. The process execution unit 21 is a program having a function of actually executing online processing and batch processing related to a database, such as executing SQL corresponding to online processing or batch processing.

Further, the session management unit 22 acquires session information for each SQL executed in online processing and batch processing, and stores and manages the acquired session information in a corresponding session information management table 50 (FIG. 5) described later. It is a program having Furthermore, the data arrangement management unit 23 is a program having a function of executing various processes related to the data arrangement management function according to the above-described embodiment.

The functions of the data management unit 20, the process execution unit 21, and the session management unit 22 are all functions of a general-purpose database management system (DBMS: Database Management System) program. Therefore, by adding the function of the data arrangement management unit 23 to the general-purpose database management system program, a program for realizing the data arrangement management function according to the present embodiment can be constructed.

On the other hand, as shown in FIG. 4, the session information management table group 24 includes a plurality of tables for managing session information related to individual SQL executed in each online process and each batch process (hereinafter referred to as “session information management”). 50).

As shown in FIG. 5, the session information management table 50 includes columns associated with CPU, network, and I / O (hereinafter referred to as CPU column 50A, network column 50B, and I / O column 50C, respectively). And a row provided corresponding to each block on the database accessed in the corresponding SQL (hereinafter referred to as a block row) 50D and a row provided corresponding to the time (hereinafter referred to as this). 50E).

In the field intersecting with the time line 50E in the CPU column 50A (the lowest field in the CPU column 50A in FIG. 5), the processor 10 (FIG. 1) of the database server device 2 required when executing the corresponding SQL is performed. The database server device 2 and the storage device required for executing the SQL are stored in the field where the processing time is stored and intersects with the time row 50E in the network column 50B (the lowest field in the network column 50B in FIG. 5). The usage time of a network (not shown in FIG. 1) connecting the three is stored.

Further, each block row 50D in the I / O column 50C is divided into a location field 50CA, a frequency field 50CB, and an average response time field 50CC, respectively. The location field 50CA stores the block address of the block accessed when the SQL is executed. The number field 50CB stores the number of times the corresponding block is accessed when the SQL is executed. The time field 50CC stores an average value of response times when the corresponding block is accessed (hereinafter referred to as an average response time). Further, in the field intersecting with the time line 50E in the I / O column 50C (the lowest field in the I / O column 50C in FIG. 5), the total time of I / O for the database required for executing the SQL is shown. Stored.

Therefore, in the case of the example of FIG. 5, in the corresponding SQL of the online process or batch process “SQL1”, the blocks “Block1”, “Block6”,..., “Block10” and “Block40” on the database Each access is “5”, “4”,…, “190” or “70”, and the average response time of these accesses is “11 ms”, “1 ms”,…, “ It is shown that it was “9 ms” or “10 ms”. In FIG. 5, in this SQL, the total processing time of the processor 10 of the database server device 2 is “140 ms”, the total network usage time is “40 ms”, and the total I / O time is “5900 ms”. It has been shown.

The session information management table 50 updates the value of the stored information every time the corresponding SQL is executed, based on the session information acquired at that time.

The importance management table 25 is a table used for managing the importance for each online process and batch process registered in advance by the system administrator. As shown in FIG. A degree column 25B is provided.

The process name column 25A stores the names of all online processes and batch processes executed in the database server device 2, and the importance column 25B stores the corresponding process (process in the process name column 25A on the same line). The degree of importance registered by the system administrator is stored. In the example of FIG. 6, the smaller the numerical value stored in the importance column 25B, the higher the importance of the corresponding online process or batch process.

Furthermore, the SLA management table 26 is a table used for managing the time conditions defined in the SLA for each online process and batch process registered in advance by the system administrator. As shown in FIG. A column 26A and an SLA column 26B are provided.

The process name column 26A stores the process names of all the processes executed in the database server device 2, and the SLA column 26B concludes the corresponding process with the user of the database system 1. The time conditions defined in the SLA (maximum values of response time and processing time) are stored.

(3) Various Processes Related to Data Arrangement Management Function According to this Embodiment Next, specific processing contents of various processes executed in the database server device 2 in relation to the data arrangement management function according to this embodiment described above. explain. In the following, the processing subject of various processes will be described as the data arrangement management unit 23 (FIG. 1), but in practice, the processing is executed by the processor 10 of the database server device 2 based on the data arrangement management unit 23. It goes without saying that it is done.

(3-1) Data Arrangement Management Process FIG. 8 shows a specific processing procedure of the data arrangement management process executed by the data arrangement management unit 23. This data arrangement management process is executed when the configuration of the storage apparatus 3 is changed, for example, when a new high-speed storage device 40 (FIG. 1) is added to the storage apparatus 3, and thereafter periodically thereafter. To be executed.

When starting the data arrangement management process, the data arrangement management unit 23 first acquires the capacity of the high-speed storage device 40 registered in advance (SP1). The capacity acquired at this time becomes the capacity of the storage area of the upper tier in the data management method using the tiered storage method applied to the storage apparatus 3.

Subsequently, the data placement management unit 23 determines a final data placement plan that satisfies a time condition specified by the SLA for each online process and batch process executed by the database server device 2. A determination process is executed (SP2).

Next, the data arrangement management unit 23 has a capacity for the high-speed storage device 40 to have enough capacity to arrange all necessary data in the storage area of the upper hierarchy according to the final data arrangement plan for each online process and batch process determined in step SP2. Executability determination processing for determining whether or not it is present is executed (SP3).

When the processing result of step SP3 is affirmative, the data arrangement management unit 23 stores the necessary data in accordance with the final data arrangement plan for each online process and batch process determined in step SP2 by the high-speed storage device 40. Whether or not the storage apparatus 3 is equipped with a function for moving to a storage area to be provided (higher-level storage hierarchy) and moving necessary data to a storage area (lower-level storage area) provided by the low-speed storage device 41 Is determined (SP4).

When the data placement management unit 23 obtains a positive result in this determination, it instructs the user to rearrange the necessary data according to the final data placement plan for each online process and batch process determined in step SP2 (hereinafter, This is referred to as a data relocation instruction) is transmitted to the storage apparatus 3 (SP5), and then this data allocation management process is terminated.

Thus, the storage apparatus 3 that has received the data relocation instruction moves necessary data to the storage area provided by the high-speed storage device 40 according to the data relocation instruction, and the low-speed storage device 41 provides the necessary data. Data relocation processing to be moved to the storage area is executed.

On the other hand, if the data allocation management unit 23 obtains a negative result in the determination at step SP4, the data allocation management unit 23 creates the processing at step SP2 by the hierarchical storage function based on the hierarchical management unit 44 (FIG. 1) installed in the storage device 3. According to the final data arrangement plan for each online process and batch process, necessary data is moved to the storage area provided by the high-speed storage device 40 and necessary data is moved to the storage area provided by the low-speed storage device 41. A dummy I / O command (hereinafter referred to as a dummy I / O command) is issued to the storage device (SP6).

Specifically, the data arrangement management unit 23 reads data that is currently stored in the storage area provided by the low-speed storage device 41 and that should be moved to the storage area provided by the high-speed storage device 40. The dummy I / O command is continuously and repeatedly transmitted to the storage apparatus 3 until the data is moved to the storage area provided by the high-speed storage device 40. The data arrangement management unit 23 similarly performs such processing for all data to be moved to the storage area provided by the high-speed storage device 40.

Then, the data placement management unit 23 moves the necessary data to the storage area provided by the high-speed storage device 40 according to the final data placement plan for each of the on-run processing and batch processing determined in step SP2, and also transfers the necessary data. When the data has been moved to the storage area provided by the low-speed storage device 41, the data arrangement management process ends.

(3-2) Data Placement Plan Determination Process FIG. 9 shows an example of specific processing contents of the data placement plan determination process executed by the data placement management unit 23 in step SP2 of the data placement management process described above with reference to FIG. Show.

When the data placement management unit 23 proceeds to step SP2 of the data placement management process, the data placement management unit 23 starts the data placement plan determination process shown in FIG. 9, and first, the database server device 2 according to a request from an external device or a preset schedule. One online process or batch process is selected from all the online processes and batch processes executed by (SP10). Hereinafter, the online process or batch process selected at this time is referred to as a selection process.

Subsequently, the data arrangement management unit 23 obtains the time condition defined by the SLA for the selection process from the SLA management table 26 (FIG. 7) (SP11), and the past session of each SQL executed in the selection process. Information is acquired from the corresponding session information management table 50 (FIG. 5) (SP12).

Next, the data arrangement management unit 23 extracts the processing time required for I / O and the processing time other than I / O for each SQL from the past session information acquired in step SP12 (SP13). Specifically, the data arrangement management unit 23 sets the I / O of the session information management table 50 (FIG. 5) associated with each SQL executed in the selection process as “processing time required for I / O”. Extract the time stored in the field where column 50C and time row 50E intersect. Further, the data arrangement management unit 23 sets the time stored in the field where the CPU column 50A and the time row 50E of the session information management table 50 intersect, the network column 50B and the time row as “processing time other than I / O”. Times obtained by adding the times stored in the fields intersected by 50E are calculated.

Further, the data arrangement management unit 23, for each SQL executed in the selection process, from the past session information acquired in step SP12, the block address of each block accessed by the SQL, the number of accesses and the average response time for each block Are extracted (SP14).

Thereafter, the data placement management unit 23 estimates the I / O processing time when the data of a certain ratio from the upper block to the upper access block is placed in the upper-layer storage area for each SQL of the selection process. Are calculated for each of a plurality of predetermined ratios (for example, 10%, 20%, 30%,...) (SP15).

Specifically, the data placement management unit 23 sets, for example, a certain default ratio of a certain SQL to a total value of the number of accesses to each block corresponding to a higher default ratio with a large number of accesses among the blocks accessed by the SQL. Multiply the response time that would be required for one I / O to the high-speed storage device, the multiplication result, and the measured value of the I / O time for each other block (the number of accesses and the average response time for that block) And the total value of the I / O processing time at the predetermined ratio of the SQL. The “response time that would be required for one I / O to the high-speed storage device” is obtained from the measurement information made in advance or from the specification information of the high-speed storage device 40 and is given to the database server device 2 in advance. Take advantage of your time.

Further, the data arrangement management unit 23 calculates the estimated value of the I / O processing time for each predetermined ratio for the SQL by repeatedly executing the same processing while sequentially changing the predetermined ratio for the SQL.

Further, the data arrangement management unit 23 calculates the estimated value of the I / O processing time for each predetermined ratio for each SQL of the selection process by executing the above process for all the SQLs of the selection process.

Subsequently, the data arrangement management unit 23 calculates an estimated value of the processing time for each data arrangement plan based on the calculation result of step SP15 (SP16). Note that the data placement plan is a plan for placing each block of upper 10%, 20%, 30%,. Each plan is 10%, 20%, 30%, etc.).

Specifically, the data arrangement management unit 23 sets the I / O processing time of one predetermined ratio with the SQL for one SQL of the selection process to an estimated value other than the SQL I / O obtained at step SP13. In this case, the estimated value of the processing time of the SQL when the block data corresponding to the upper predetermined ratio with the higher number of accesses is arranged in the storage area of the upper hierarchy is calculated. In addition, the data arrangement management unit 23 similarly calculates other SQL processing time estimates for the same predetermined ratio, and adds all of these calculation results to obtain the upper limit of the predetermined ratio with the highest number of accesses. An estimated value of the processing time of the selection process when the block data is arranged in the storage area of the upper hierarchy is calculated. The estimated value calculated at this time is the estimated value of the processing time of the data arrangement plan in which the data of the block corresponding to the predetermined ratio of the upper number with the highest number of accesses is arranged in the storage area of the upper hierarchy.

Therefore, the data arrangement management unit 23 calculates the estimated value of the processing time for each data arrangement plan by executing the above process in the same manner for all the predetermined ratios.

Then, the data arrangement management unit 23 determines whether there is a data arrangement plan that satisfies the time condition defined by the SLA of the selection process based on the calculation result of step SP16 (SP17). Specifically, the data placement management unit 23 compares the estimated processing time for each data placement plan calculated in step SP16 with the SLA time condition for the selection processing in order, and the estimated processing time is Such a determination is made depending on whether or not a data arrangement plan smaller than the condition exists.

If the data allocation management unit 23 obtains a negative result in this determination, it displays a warning that the capacity of the high-speed storage device 40 is insufficient on the database server device 2 (SP18), and thereafter, this data allocation plan The determination process is terminated, and the data arrangement management process described above with reference to FIG. 8 is also terminated.

On the other hand, if the data arrangement management unit 23 obtains a positive result in the determination at step SP17, the data arrangement that uses the least amount of the high-speed storage device 40 among the data arrangement proposals that satisfy the SLA time condition for the selection process. The plan is determined as the final data arrangement plan of the selection process (SP19).

Thereafter, the data placement management unit 23 determines whether or not the final data placement has been determined for all online processing and batch processing executed by the database server device 2 (SP20). If the data arrangement management unit 23 obtains a negative result in this determination, it returns to step SP10, and thereafter, the processing selected in step SP10 is sequentially switched to other unprocessed online processing or batch processing, and step SP10 to step SP10 are performed. Repeat the process of SP20.

When the data placement management unit 23 finally determines the final data placement plan for all online processes and batch processes executed by the database server device 2 and obtains a positive result in step SP20, the data placement plan The determination process ends and the process returns to the data arrangement management process of FIG.

(3-3) Executability Determination Process FIG. 10 shows an example of specific processing contents of the execution permission determination process executed by the data arrangement management unit 23 in step SP3 of the data arrangement management process described above with reference to FIG.

When the data placement management unit 23 proceeds to step SP3 of the data placement management process, the data placement management unit 23 starts the execution determination process shown in FIG. 10, and first, for each online process and batch process in the data placement plan determination process described above with reference to FIG. When the data rearrangement is changed in accordance with the final data arrangement plan determined respectively, it is determined whether or not the capacity of the high-speed storage device 40 (the capacity of the storage area of the upper hierarchy) is sufficient (SP30).

Specifically, for each online process and batch process, the data placement management unit 23 calculates the sum of blocks to be placed in the storage area provided by the high-speed storage device 40 for the online process or batch process, and displays the block in the calculation result. By multiplying the capacity per one, the sum total of the data amount of data to be arranged in the upper tier storage area is calculated. Then, the data arrangement management unit 23 determines whether or not the total amount of data to be arranged in the higher-level storage area calculated for each online process and batch process is larger than the capacity of the high-speed storage device 40. Thus, the determination at step SP30 is performed.

If the data arrangement management unit 23 obtains an affirmative result in this determination, the data arrangement management unit 23 ends the execution determination process and returns to the data arrangement management process (FIG. 8).

On the other hand, if the data arrangement management unit 23 obtains a negative result in the determination at step SP30, it executes the final data arrangement plan for each online process and each batch process determined at step SP2 of the data arrangement management process. Displays a warning to the effect that the capacity of the high-speed storage device 40 is insufficient (SP31). Thereafter, the execution determination process is terminated, and the data arrangement management process described above with reference to FIG. 8 is also terminated.

On the other hand, FIG. 11 shows another process example of the execution feasibility determination process executed by the data arrangement management unit 23 in step SP3 of the data arrangement management process described above with reference to FIG.

When the data placement management unit 23 proceeds to step SP3 of the data placement management process, the data placement management unit 23 starts the execution determination process shown in FIG. 11. First, in the data placement plan determination process described above with reference to FIG. When the data rearrangement is changed in accordance with the final data arrangement plan determined respectively, it is determined whether or not the capacity of the high-speed storage device 40 (the capacity of the upper-layer storage area) is sufficient (SP40). Since the specific processing content of this step SP40 is the same as that of step SP30 of the execution determination processing described above with reference to FIG. 10, the description thereof is omitted here.

On the other hand, when the data arrangement management unit 23 obtains a negative result in the determination at step SP40, the data arrangement management unit 23 performs a display similar to that at step SP31 of the execution determination process described above with reference to FIG. 10 (SP41). With reference to the management table 25 (FIG. 6), the final data arrangement plan of each online process and each batch process determined in step SP2 of the data arrangement management process (FIG. 8) When adopting in order from the processing, it is calculated up to which data processing plan of what importance online processing or batch processing can be executed with the capacity of the current high-speed storage device 40, and the calculation result is stored in the database server device 2. (SP42).

Then, the data placement management unit 23 determines whether or not an operation input indicating consent from the user has been given to the execution of the online processing of the importance level displayed in step SP42 or the data placement plan of the batch processing (SP43). . If the data arrangement management unit 23 obtains a positive result in this determination, the data arrangement management unit 23 ends the execution determination process and returns to the data arrangement management process (FIG. 8). Thus, in this case, in step SP5 or step SP6 of the data arrangement management process, the data arrangement management unit 23 stores the data so as to perform data rearrangement up to the data arrangement plan of the importance online process or batch process agreed by the user. The apparatus 3 is controlled.

On the other hand, if the data placement management unit 23 obtains a negative result in the determination at step SP43, the data placement management unit 23 performs data rearrangement according to the data placement plan of each online process and each batch process determined at step SP2 of the data placement management process. After the database server device 2 displays a warning to the effect that it cannot be performed (SP44), the execution determination process is terminated and the data arrangement management process (FIG. 8) is also terminated.

(4) Effects of the present embodiment As described above, in the database system 1 of the present embodiment, a part of data having a high access count among the database data is stored in the upper layer for each online process and batch process. Among the plurality of data arrangement plans to be arranged in the area, one data arrangement plan satisfying the SLA time conditions defined in advance for the online process or batch process is set as the final data arrangement plan of the online process or batch process. Each is determined, and the storage apparatus 3 is caused to rearrange the data in the database according to the final data allocation plan for each determined online process and batch process.

Therefore, according to this database system 1, only the minimum data satisfying the respective time conditions defined by the SLA concluded for each online process and each batch process among the data in the database is stored in the storage area of the upper hierarchy. Since data can be rearranged, it is possible to use the storage area of the upper hierarchy more efficiently than when data is rearranged in units of the logical structure of the database, and each online process and each batch process are concluded. Optimal data management can be performed with the SLA line of sight. In this way, it is possible to appropriately manage the database held by the storage apparatus 3 equipped with the hierarchical storage function.

(5) Other Embodiments In the above-described embodiment, the case where the present invention is applied to the database server device 2 that executes both online processing and batch processing has been described. The present invention is not limited to this, and the present invention can also be applied to a database server device that executes only one of online processing and batch processing.

Further, in the above-described embodiment, the number of accesses is large as a plurality of data arrangement plans for arranging some data in the database data in the storage area of the upper hierarchy for each online process and batch process. Although a case has been described in which the proposal of allocating data for different upper-level predetermined ratios (upper 10%, 20%, 30%,...) In the upper-tier storage area has been described, the present invention applies to this. Not limited to, for example, a default size may be applied instead of a default ratio, and a plurality of data arrangement plans are created in which a part of data in the database data is arranged in a storage area of an upper hierarchy. Various other methods can be widely applied as the method.

The present invention can be widely applied to database management apparatuses having various configurations for managing a database held by a storage apparatus equipped with a hierarchical storage function.

DESCRIPTION OF SYMBOLS 1 ... Database system, 2 ... Database server apparatus, 3 ... Storage apparatus, 10, 42 ... Processor, 11, 43 ... Memory, 20 ... Data management part, 21 ... Process execution part, 22 ... Session management unit 23... Data arrangement management unit 24 24 Session information management table group 25. Importance management table 26... SLA management table 44 44 Hierarchy management unit 40 40 High-speed storage device 41... Low-speed storage device, 50... Session information management table.

Claims

Equipped with a hierarchical storage function that moves data with higher access frequency to a higher-level storage area with higher response performance and moves data other than the data to a lower-level storage area with lower response performance than the first storage area In the database management device that manages the database held in the storage device
A process execution unit that executes processes related to the database;
A session management unit that acquires and manages session information at the time of execution of each of the processes;
A data placement management unit that causes the storage device to rearrange the data of the database as needed based on the session information of each of the processes,
The data arrangement management unit
For each of the processes, from among a plurality of data arrangement plans for arranging a part of the data of the database having a large number of accesses in the storage area of the upper hierarchy, a condition regarding a time defined in advance for the process is set. Each of the data arrangement plans to be satisfied is determined as a final data arrangement plan for the processing,
When the data of the database is rearranged according to the final data arrangement plan for each of the determined processes, it is determined whether the storage area of the upper hierarchy is sufficient,
When the storage area of the upper tier is sufficient, the database management apparatus causes the storage apparatus to relocate data of the database according to the final data arrangement plan for each of the determined processes. .
A plurality of data arrangement plans for allocating a part of the data of the database having a high access frequency in the storage area of the upper tier are respectively different upper ratios of the database data having a high access frequency. The database management apparatus according to claim 1, wherein the data is arranged in a storage area of the upper hierarchy.
The data placement management unit, for each process,
Calculating a processing time for each of the plurality of data arrangement plans for arranging a part of the data of the database with a large number of accesses in the upper tier storage area;
Among the data arrangement plans for which the processing time has been calculated, the data arrangement plan that satisfies the conditions specified in advance for the process and uses the least amount of storage area in the upper tier is selected as the final of the process. The database management apparatus according to claim 2, wherein the database management apparatus determines the data arrangement plan.
The database management apparatus according to claim 3, wherein the condition is a condition relating to a response time or a processing time defined in a quality assurance contract concluded with a customer.
The data arrangement management unit
When a predetermined function is installed in the storage device, a predetermined command for causing the storage device to perform data relocation of the database according to the data allocation plan for each of the determined processes is stored in the storage device. To the device,
If the function is not installed in the storage device, a dummy I / O command is transmitted to the storage device so that the data in the database is rearranged according to the determined data placement plan for each process. The database management apparatus according to claim 2, wherein:
Equipped with a hierarchical storage function that moves data with higher access frequency to a higher-level storage area with higher response performance and moves data other than the data to a lower-level storage area with lower response performance than the first storage area A database management method executed in a database management device that manages a database held in a storage device,
The database management device includes:
A process execution unit that executes processes related to the database;
A session management unit that acquires and manages session information at the time of execution of each of the processes;
A data placement management unit that causes the storage device to rearrange the data of the database as necessary based on the session information of each of the processes;
For each process, the data arrangement management unit pre-determines the process from a plurality of data arrangement plans for arranging a part of the data in the database having a high access frequency in the storage area of the upper hierarchy. A first step of respectively determining one data arrangement plan that satisfies a condition relating to a specified time as a final data arrangement plan of the processing;
The data arrangement management unit determines whether or not the storage area of the upper hierarchy is sufficient when the data in the database is rearranged according to the final data arrangement plan for each of the determined processes. And the steps
The data placement management unit causes the storage device to rearrange the data of the database according to the final data placement plan for each of the determined processes when the storage area of the upper hierarchy is sufficient. 3. A database management method comprising: 3 steps.
A plurality of data arrangement plans for allocating a part of the data of the database having a high access frequency in the storage area of the upper tier are respectively different upper ratios of the database data having a high access frequency. The database management method according to claim 6, wherein the data is arranged in a storage area of the upper hierarchy.
In the first step, the data placement management unit, for each process,
Calculating a processing time for each of the plurality of data arrangement plans for arranging a part of the data of the database with a large number of accesses in the upper tier storage area;
Among the data arrangement plans for which the processing time has been calculated, the data arrangement plan that satisfies the conditions specified in advance for the process and uses the least amount of storage area in the upper tier is selected as the final of the process. The database management method according to claim 7, wherein the database management method is determined as the data arrangement plan.
The database management method according to claim 8, wherein the condition is a condition related to a response time or a processing time defined in a quality assurance contract concluded with a customer.
In the third step, the data arrangement management unit
When a predetermined function is installed in the storage device, a predetermined command for causing the storage device to perform data relocation of the database according to the data allocation plan for each of the determined processes is stored in the storage device. To the device,
If the function is not installed in the storage device, a dummy I / O command is transmitted to the storage device so that the data in the database is rearranged according to the determined data placement plan for each process. The database management method according to claim 7.