JP3180336B2 - Data access method using multilayer buffer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a secondary storage device used as a storage means of a database and a data access using a multi-layer buffer which is a method of storing and expressing data in a main storage device. About the method.

(Prior Art) Conventionally, buffering for accelerating access to a secondary storage device (a method of holding data once accessed from a secondary storage device on a main storage device and reusing the data) is as follows.
There is known a method of storing data in a format called a page (or a block) and having a page buffer on a main storage device. This method is called Wieder Hold (Gio Wiede
rhold) Database Design (Database Desig)
n, Second Edition, McGaw-Hill Book Company, 19
1983, pp. 56-59).

(Problems to be Solved by the Invention) Object-oriented databases, which have been actively researched recently, are intended for advanced application fields such as CAD and expert systems. In an application program in such a field, it is necessary to read a large number of data (objects) into a main storage device and execute an operation using them at high speed. However, in a conventional database system, by buffering data in page units, a large number of objects used in an application program are not sufficiently stored in the main storage device as compared with the size of the main storage to be secured. There is a problem that the performance of the operation is low. Further, in the conventional system, the processing overhead is increased by retrieving the object from the page each time the data value is accessed and then retrieving the data value from the object, so that the performance degradation due to the increase in the processing overhead is reduced. There was a problem of inviting.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a data access method using a multi-layer buffer for executing data access from a secondary storage device and a main storage device at high speed.

(Means for Solving the Problems) According to the present invention, in order to speed up data access from a secondary storage device of a computer to a main storage device, data once accessed from the secondary storage device is stored on the main storage device. In the method of holding and reusing, a page buffer holding a plurality of pages corresponding to a data block in the secondary storage device, and holding a plurality of used objects extracted from each page. An object buffer and a data buffer holding a plurality of in-use data values extracted from each object are prepared in the main storage device, and when accessing a specified data value, an object ID and an item are used. A data area corresponding to the hash value obtained from the ID is found from the data buffer, and the data area is connected to another data area. The data buffer, the object buffer, the page buffer, and the secondary storage device are sequentially searched while following the interface, and when the specified data value is found from the data buffer, the specified data value is returned. When an object corresponding to the specified data value is found from the buffer, the specified data value is fetched from the object, the specified data value is written to a new data area created in the data buffer, and the new data value is written. Performing a first data holding operation including setting a pointer of a data area immediately before having the hash value to the data area and returning the specified data value;
When a page corresponding to the specified data value is found from the page buffer, a second data holding operation including fetching an object corresponding to the specified data value from the page and writing the object to the object buffer. After the operation, the first holding operation is performed,
When a page corresponding to the specified data value is found in the secondary storage device, after the operation of writing the page from the secondary storage device to the page buffer, the second
The data holding operation and the first data holding operation are sequentially performed.

(Operation) In the present invention, frequently accessed data values are stored in a data buffer in the main memory, and when there is a request from an application program, the data value is immediately retrieved from the data buffer and returned. If the corresponding data value is not in the data buffer, the corresponding data value is fetched from the object buffer. If the corresponding data value is not in the object buffer, the corresponding data value is fetched from the page buffer. It is. Retrieving a data value from a data buffer is much simpler than retrieving a data value from a page buffer, and is therefore faster than conventional methods using only a page buffer. When using the same amount of main storage area, it is more applicable to divide the area into a page buffer, an object buffer, and a data buffer than to use the entire area as a page buffer as in the conventional method. Since more data values are used in the main memory by the program, the number of accesses to the secondary memory is reduced, and the speed of data access can be increased.

(Example) An example of the present invention will be described with reference to the drawings.

FIG. 1 is a flowchart showing one embodiment of a data access method using a multilayer buffer according to the present invention. FIG. 2 is a diagram showing the internal structure of an object and a page used in the data access method using the multi-layer buffer of the present invention. FIG. 3 is a data buffer and object buffer used in the data access method using the multi-layer buffer of the present invention. FIG. 4 is a diagram showing the internal configuration of a data buffer and a data area, and FIG. 5 is a diagram showing the device configuration of a computer to which a data access method using a multilayer buffer according to the present invention is applied. FIG.

In FIG. 2, a set of data to be processed by the application program is called an object. An object has a plurality of data values as its contents. A unit of reading / writing data (corresponding to a block) with the secondary storage is called a page. A page has a plurality of objects as contents.

In FIG. 3, all the buffers exist in the main storage device. In the figure, a data buffer (abbreviation DBF) holds a plurality of data values as contents. The object buffer (abbreviated as OBF) holds a plurality of objects as contents. The page buffer (abbreviation PBF) holds a plurality of pages as contents.

In FIG. 4, the hash table is located at the head of the data buffer and has a pointer (address) to the data area.
Including the set of The data area corresponds to one data value,
A pointer to the next data area with the same hash value,
Object ID corresponding to its own size and data value
And the item ID and data value. Note that empty data areas are joined by a pointer and pointed to from the first entry in the hash table.

The method of the present invention will be described with reference to FIGS. Here, an example is shown in which the application program requests a value (data value) of a specific item in a specific object. The data buffer, the object buffer, and the page buffer are stored in the main storage device 52.

First, in step 11, the object ID (OID)
And a hash value from the item ID (IID).

In step 12, a hash table is drawn based on the hash value to obtain the entry.

In step 13, it is checked whether the pointer in the hash table or the data area is empty, and if not empty, the process jumps to step 14, and if empty, the process jumps to step 17.

In step 14, follow the pointer to get the first or next data area.

In step 15, the object ID in the data area
And the item ID are compared with the value requested by the application program.
Return to

In step 16, the data value is returned to the application program, and the process ends.

In step 17, it is checked whether or not there is an object containing the data value in the object buffer.

In step 18, it is checked whether or not there is a page including the object in the page buffer.

In step 19, the page is read from the secondary storage device 53 and held in the page buffer.

In step 20, the object is copied from the page buffer and put into the object buffer.

In step 21, the data value is extracted from the object in the object buffer.

In step 22, a new data area is created in the data buffer, and the data value is copied.

In step 23, the pointer of the immediately preceding data area in the data buffer is set to point to the data area,
Jump to step 16.

As described above, when the application program requests a specific data value, the number of execution steps of the central processing unit 51 and the number of accesses to the secondary storage device 53 are reduced as compared with the conventional method, so that data access is performed at high speed. Can be

(Effects of the Invention) According to the present invention, when an application program requests a specific data value, the number of execution steps of the central processing unit is reduced and the number of accesses to the secondary storage is reduced as compared with the conventional method. Therefore, the access to the secondary storage data can be executed at higher speed.

[Brief description of the drawings]

FIG. 1 is a flowchart showing one embodiment of a data access method using a multi-layer buffer of the present invention. FIG. 2 is a diagram showing the internal structure of an object and a page used in the data access method using a multi-layer buffer of the present invention. FIG. 3 is a diagram showing the configuration of a data buffer, an object buffer, and a page buffer used in the data access method using the multi-layer buffer of the present invention. FIG. 4 is a diagram showing the internal configuration of the data buffer and the data area. FIG. 1 is a diagram showing a device configuration of a computer to which a data access method using a multilayer buffer according to the present invention is applied. 51 central processing unit, 52 main storage unit, 53 secondary storage unit

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-63-313252 (JP, A) JP-A-3-142463 (JP, A) JP-A-1-206431 (JP, A) et al. "Per sistent Caching: An Implimentation Te chnique for Comple x Objects with Obj ect Indentity" Tech nical Report 89-021, Department of Info rmation Science, Un iversity of Tokyo (1989) (58) investigated the field (Int.Cl. ⁷ G06F 12/00 514 G06F 12/00 547 JICST file (JOIS)

Claims (1)

(57) [Claims] 1. A method for retaining and reusing data once accessed from a secondary storage device on the main storage device in order to speed up data access from a secondary storage device of the computer to the main storage device. A page buffer holding a plurality of pages corresponding to the data blocks in the secondary storage device; an object buffer holding a plurality of in-use objects extracted from each page; A data buffer holding a plurality of used data values used in the main storage device, and when accessing the specified data value, a hash value obtained from the object ID and the item ID. A data area corresponding to the data area is found from the data buffer, and the data area is traced while following a pointer connecting the data area to another data area. Data buffer, the object buffer, the page buffer, and the secondary storage device. When the specified data value is found from the data buffer, the specified data value is returned, and the specified data value is returned from the object buffer. When an object corresponding to is found, the specified data value is taken out from the object, the specified data value is written to a new data area created in the data buffer, and the specified data value is written to the new data area. Performing a first data holding operation consisting of setting a pointer to a data area immediately before having a hash value and returning the specified data value;
When a page corresponding to the specified data value is found from the page buffer, a second data holding operation including fetching an object corresponding to the specified data value from the page and writing the object to the object buffer. After the operation, the first holding operation is performed,
When a page corresponding to the specified data value is found in the secondary storage device, after the operation of writing the page from the secondary storage device to the page buffer, the second
A data access method using a multi-layer buffer, wherein the data holding operation and the first data holding operation are sequentially performed.