JP2006350633A - Data management method and data management system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To maintain the cache hit rate of frequently accessed data by preventing the data to be accessed frequently from being removed as much as possible. <P>SOLUTION: An application 4 specifies a cache control system according to the type of an access request at each access request, and a data managing device 5 controls a cache 8 according to the specified cache control system. Data to be held in the cache 8 are managed in the order of the time of use. The data managing device 5 holds the used data at the latest time position of the cache, when the cache control system controls data to be accessed frequently to prevent the data from being removed from the cache, and holds the used data at an older predetermined time position in the cache, when the cache control system controls data that are not accessed frequently. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a data management method and a data management system, and in particular, a data management method suitable for use in holding file data and database data held in an external storage device in a cache of a main storage device, and The present invention relates to a data management system.

  As a conventional technique related to management for holding file data and database data held in an external storage device in a cache of a main storage device, for example, a technique described in Patent Document 1 is known.

In the prior art described in Patent Document 1, file data or database data held in an external storage device is read into the cache of the main storage device, and thus any of the data already read into the cache is driven out. . At this time, a criterion for determining which data is left in the cache is necessary, and the above-described conventional technique caches frequently used data as the determination criterion. Therefore, data frequently used for each logged-in user is used. The cache management is performed so that a high priority is set to the data and such data is left in the cache with a high probability. Specifically, it has an access frequency database that records a user name and data frequently accessed by the user in association with each other, and sets a priority for data copy in the cache.
JP-A-7-93205

  The above-mentioned conventional technology can expedite low-priority data from the cache and prevent frequently accessed data from being expelled when referring to or updating frequently used data in small quantities. On the other hand, it can be accessed from the cache with a high probability. However, there are a wide variety of operations on data, and when performing different operations to access different data, the above-mentioned conventional technology sets a high priority for data that has been frequently used before. Because it is not possible to cope with the change in the setting of the data and unused data is not evicted from the cache, the cache capacity for the data to be accessed is reduced, and the cache is frequently evicted for data access. And the process of fetching data of the external storage device into the cache is performed, which causes a problem that the performance of the system deteriorates.

  In addition, with the above-described conventional technology, if a large amount of search or update is performed on data with low usage frequency, the capacity of the cache is limited, so even high priority data is evicted from the cache. The problem is that the external storage device will be read directly until it is evicted from the cache area and read again into the cache area, degrading the performance of the entire system. have.

  The object of the present invention is to solve the above-mentioned problems of the prior art, and can cope with each work content of data to be accessed. An object of the present invention is to provide a data management method and a data management system that can be realized.

  According to the present invention, in the data management method for controlling to hold the data in the external storage device in the cache of the main storage device by executing the application of the computer, the access is performed each time the application makes an access request. This is achieved by a data management apparatus that specifies a cache control method according to the type of request and controls the cache, and controls the cache according to the specified cache control method.

  In addition, the object is to provide a data management system including a data management device that performs control to hold data in an external storage device in a cache of a main storage device by executing an application of the computer, each time the application makes an access request. This is achieved by designating a cache control method according to the type of access request, and controlling the cache by the data management device according to the designated cache control method.

  According to the present invention, it is possible to maintain the cache hit rate of frequently accessed data by preventing frequently accessed data from being expelled from the cache as much as possible.

  Embodiments of a data management method and a data management system according to the present invention will be described below in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a data management system according to an embodiment of the present invention. In FIG. 1, 1 is a computer, 2 is an external storage device, 3 is a main storage device, 4 is an application, 5 is a data management device, 6 is an access request control unit, 7 is a cache management control unit, 8 is a cache, It is a data area.

  As shown in FIG. 1, the data management system according to the embodiment of the present invention includes a computer 1 that executes an application 4 and an external storage device 2 that is accessed when the application 4 is executed. The computer 1 stores on the main storage device 3 an application 4, a data management device 5 that accepts a request from the application 4 and makes a data access request, and a cache 8 as a shared memory. The data management device 5 includes an access request control unit 6 and a cache management control unit 7, and accesses the data area 9 on the external storage device 2 through the cache 8.

  In the foregoing, as is well known, the computer 1 includes a CPU in addition to the main storage device 3, and includes an access request control unit 6 constituting the data management device 5 in the embodiment of the present invention. The cache management control unit 7 can be realized by loading an application program stored in the external storage device 2 into the main storage device 3 and causing the CPU to execute it under the OS.

  In the embodiment of the present invention configured as described above, the application 4 requests the data management device 5 for the data area 9 allocated on the external storage device 2 for each processing access. At that time, the application 4 sets a cache management code indicating a cache control method for each processing access, and also passes this cache management code to the data management apparatus 5. The access request control unit 6 of the data management apparatus 5 analyzes the data access request from the application 4 and requests data from the cache management control unit 7 based on the analysis result. If there is requested data on the cache 8, the cache management control unit 7 reads the data and transmits it to the request source. If the requested data is not on the cache 8, the cache management control unit 7 reads the data area 9 from the external storage device 2. Are read out and written into the cache 8, and the data is transmitted to the request source. At that time, the cache management control unit 7 determines the data storage position in the cache 8 based on the cache control method indicated by the cache management code designated for each processing access from the application 4.

  FIG. 2 is a flowchart for explaining processing operations until data requested in response to an access request from an application is returned in the embodiment of the present invention. Next, this will be described.

(1) When the application 4 in the computer 1 makes a data access request to the data management device 5, the access request control unit 6 in the data management device 5 analyzes the data access request, and based on the result, To the cache management control unit 7. The cache management control unit 7 that has received the access request performs data access control processing, and determines whether or not the requested data has been read into the cache 8 (step S1).

(2) If the requested data has already been read into the cache 8 in the determination in step S1, the cache management control unit 7 performs processing in the designated cache control method, and rewrites the position on the cache. (Restore location) is determined, the target data is read from the cache 8, and the result is returned to the application 4 (steps S2 and S3).

(3) If it is determined in step S1 that the requested data has not been read into the cache 8, the cache management control unit 7 reads the data from the data area 9 of the external storage device 2 and designates the specified cache control method. Is performed to determine the position of data to be purged on the cache 8 and the write position (storage position) of the data read from the data area 9 of the external storage device 2 in the cache 8 (steps S4 and S5).

(4) Thereafter, the cache management control unit 7 updates the cache 8 by writing the data read from the data area 9 of the external storage device 2 to the position determined in the process of step S5 of the cache 8, and the held data. The result is sent back to the application 4 (steps S6 and S7).

  According to the processing in the embodiment of the present invention described above, an access request from an application can be processed by a cache management method designated for each request, and the request temporarily refers to data by this processing. Even if this is the case, it is possible to efficiently allocate a finite cache by setting an appropriate cache management method, and control the cache so that frequently used data is left in the cache area with a high probability. It can be carried out. As a result, the cache area can be used efficiently, so that frequently accessed data can be constantly left on the cache, and the speed of data access can be increased.

  FIG. 3 is a diagram illustrating an example in which a cache is used according to a different cache control method for each processing access. In the example shown in FIG. 3, in the case where the application 4 is a business that refers to a small amount of data, the cache control method that expects the corresponding data to remain on the cache and speed up the access processing is LRU. Specify a cache control method that is not required to be reused even if the data is read into the cache in the case of business such as a case where all data is processed, etc., as a non-LRU method It is supposed to be possible.

  Then, the LRU method data access by the access request a10 of the application 4 and the non-LRU method data access by the access request b11 of the application 4 are executed on the computer 1, and the cache 8 stores these access requests a10 and b11. It is assumed that it is shared with.

  FIG. 4 is a diagram for explaining the arrangement of data in the cache and the outline of processing in the cache management control unit 7 when the cache is shared by different cache control methods. Next, referring to FIG. It will be described that a finite cache area can be effectively used in different cache control methods by the arrangement of data in the memory and the control of the cache management control unit. In the following description, an LRU method and a non-LRU method will be described as examples of the cache control method.

  First, the arrangement of data in the cache 8 will be described. The blocks of data stored in the cache 8 are managed in the order of time used in the order of latest 1, latest 2,..., Oldest 3, oldest 2, and oldest 1. For this reason, each block of data is managed by connecting information of its storage position (data storage address in the cache) in order by a chain as indicated by an arrow in FIG. In the example shown in FIG. 4, the storage position of the oldest 3 to oldest 1 data block is set as a cache area used in the non-LRU method, and all the data blocks from the latest 1 to the oldest 1 are set. Is set so that it can be used in the LRU method.

  Note that only the address indicating the storage location of the data block may be managed by being connected by the chain, and the actual data may be stored in any location in the cache 8. In addition, each data block in the cache is assigned a non-LRU flag to data used in the non-LRU method in order to indicate whether the cache control method was difficult when the data was most recently used. I shall keep it.

  The cache management control unit 7 discharges the oldest cache data stored in the cache, in the example shown in FIG. Whether reading or writing data or reading data from the cache area, the data is always stored in the latest cache position, in the example shown in FIG. As a result, the corresponding data remains on the cache 8 constantly.

  In addition, the cache management control unit 7 discharges the oldest cache data stored in the cache, in the example shown in FIG. 4, the oldest 1 data when performing control by the non-LRU method. Whether the data is read and written from the device 2 or the data is read from the cache area, the latest cache position of the cache area in which the data is used in the non-LRU method, the oldest 3 position in the example shown in FIG. Store. However, if the data read from the cache area does not have the above-mentioned non-LRU flag and the data is written in the cache by the LRU method, the data is the latest cache location, in the example shown in FIG. Is stored at the position of. Thereby, data on the cache used in the LRU method can be prevented from being discharged from the cache even when used in the non-LRU method.

  FIG. 5 is a flowchart for explaining the processing operation in the cache management control unit. Next, this will be explained.

(1) Upon receiving the access request, the cache management control unit determines whether the cache management code specified by the application specifies the LRU method or the non-LRU method (step S8).

(2) If the LRU method specified by the access request a10 is specified in step S8, it is determined whether the request data has been read into the cache. If the request data has been read into the cache, the corresponding data The cache storage location information is moved to the latest location, that is, the latest location 1 in FIG. 4 (steps S9 and S10).

(3) If the request data has not been read into the cache in step S9, the data is read from the external storage device, and the data at the cache location with the oldest data and the oldest data in FIG. The read data is written in the latest position of the cache, the latest position of FIG. 4 (steps S11 to S13).

(4) After the processing of steps S10 and S13 and step S18 described later, if the target cache information has a non-LRU flag, the non-LRU flag is deleted, and the requested data on the cache is returned to the application (step) S14, S15).

  The processing in the case of the LRU method is performed as described above, and the cache of the data to be read is always moved to the latest position. Therefore, the cache is not expelled from frequently accessed data, and the cache can be maintained. it can.

(5) If the non-LRU method specified by the access request b11 is specified in the determination in step S8, it is determined whether the request data has been read into the cache. If the request data has been read into the cache, the target It is determined whether or not the non-LRU flag is set in the data cache information (steps S16 and S17).

(6) If it is determined in step S17 that the non-LRU flag is not set in the cache information of the target data, the data is frequently accessed data that was previously used in the LRU method. Is moved to the latest position, that is, the latest position described with reference to FIG. 4, and thereafter, the process proceeds to the process from step S14 described above (step S18).

(7) If the non-LRU flag is set in the cache information of the target data in the determination in step S17, the data is data that has been read in the non-LRU method before and is less likely to be read again. Since it is data, the cache position of the data is moved to the newest position of the cache area used in the non-LRU method, the oldest 3 position shown in FIG. 4 (step S19).

(8) If the request data has not been read into the cache in the determination at step S16, the data is read from the external storage device, and the cache position where the oldest data is located and the data at the position of the oldest 1 in FIG. Then, the read data is written into the newest position of the cache area used in the non-LRU method of the cache, that is, the oldest 3 position shown in FIG. 4 (steps S20 to S22).

(9) After the processing in steps S19 and S22, the non-LRU flag is reset in the target cache information, and the requested data on the cache is returned to the application (steps S23 and S24).

  As described above, the processing in the case of control by the non-LRU method in the embodiment of the present invention is frequently accessed in order to always move the data written in the cache by the control of the LRU method to the latest position. Data is not evicted from the cache and the cache can be maintained. Also, data written in the cache under non-LRU method control is moved to the newest location in the cache area used in the non-LRU method, so that access using the non-LRU method can be performed using the minimum cache capacity. The cache can be maintained so that frequently accessed data is not evicted from the cache.

  The processing in the above-described embodiment of the present invention is configured by a program and can be executed by a CPU provided in a computer, and these programs are provided by being stored in a recording medium such as an FD, CDROM, or DVD. And can be provided by digital information over a network.

  According to the above-described embodiment of the present invention, the cache control method that operates every time an access request is made can be changed, so that cache hits of frequently accessed data are avoided so as not to expel data from the cache area as much as possible. The rate can be maintained. In addition, when referring to all data temporarily, the data is read into the cache so that it is not frequently accessed, thus preventing frequently accessed data from being expelled from the cache. Can do.

  In the above-described embodiment of the present invention, the block of data stored in the cache is described as being chain-managed by the address. However, the present invention manages the time of the block of data stored in the cache, and the old write data. The same management can be performed when the LRU method of discarding is performed.

  Further, the present invention can also be applied to the case of control in which priority is set as described in Patent Document 1 described in the related art. In other words, in this case, in the case of control in which a non-priority level is set, the data read into the cache is maintained in the priority level and remains in the cache, and the data that has not been read into the cache is forced to have a priority level. What is necessary is just to set to 0 which shows non-priority and to write the data to a cache. Also, in the case of control with a high priority set, if the priority is 0 for data read into the cache, the priority is reset, and the priority for data that has not been read into the cache is set as before. You can set it and write it to the cache.

  As described above, the present invention can be applied to various cache management controls, and the cache can be maintained so that frequently accessed data is not purged from the cache.

It is a block diagram which shows the structure of the data management system by one Embodiment of this invention. 6 is a flowchart illustrating processing operations until data requested in response to an access request from an application is returned in the embodiment of the present invention. It is a figure explaining the example which uses a cache by the different cache control system for every process access. It is a figure explaining the arrangement | sequence of the data in a cache in the case of sharing a cache by a different cache control system, and the outline | summary of the process in the cache management control part 7. FIG. It is a flowchart explaining the processing operation in a cache management control part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Computer 2 External storage device 3 Main storage device 4 Application 5 Data management device 6 Access request control part 7 Cache management control part 8 Cache 9 Data area 10 Access request a
11 Access request b

Claims (4)

  In a data management method for performing control to hold data in an external storage device in a cache of a main storage device by executing an application on a computer, the application performs a cache control method according to the type of access request each time an access request is made A data management method characterized in that a data management device that designates the cache and controls the cache controls the cache according to the designated cache control method.   The data held in the cache is managed in the order of time used, and the data management device controls the cache control method so that it is not frequently evicted from the cache because it is for frequently accessed data. The used data is kept at the most recent time position in the cache, and if the cache control method is for data that is not frequently accessed, the used data is 2. The data management method according to claim 1, wherein the cache is controlled so as to be held at an old predetermined time position.   3. The data management method according to claim 1, wherein an access control method for the frequently accessed data is an LRU method, and an access control method for the data that is not frequently accessed is a non-LRU method.   In a data management system including a data management device that performs control to hold data in an external storage device in a cache of a main storage device by executing a computer application, each time an access request is made, the application sets the type of access request. A data management system according to claim 1, wherein the cache control method is designated, and the data management device controls the cache according to the designated cache control method.

Images