JPH01184763A - Data storage system in file system - Google Patents

Abstract

PURPOSE:To optimize the number of blocks to be used in each record and to attain rapid access by previously preparing plural kinds of block sizes and selecting a block size corresponding to a record size. CONSTITUTION:A volume control area 1 for storing volume control information for defining the positions, sizes, etc., of respective areas, an area 2 for storing file control information for defining the positions, sizes, etc., of blocks constituting each record and a record area 3 storing the content data of respective records are formed. A large record such as an image data record is dividedly stored in blocks with a large size and a small data record consisting of code data such as key words is dividedly stored in blocks with a small size. Consequently, the space of a file medium can be effectively utilized and access to respective records can be rapidly executed.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a data storage method in a file system, and in particular uses a large-capacity file medium such as a magneto-optical disk on which data can be erased and rewritten. Concerning data storage methods suitable for file systems.

[Conventional technology]

In fields such as office automation where large amounts of data are managed by computers, file systems that use optical disks as information distribution media are used. The optical discs that have been put into practical use to date are "write-once" types, where data cannot be erased.
However, in recent years, magneto-optical disks on which data can be erased and rewritten have been developed, replacing the conventional "
It is attracting attention as a large-capacity information storage medium that can replace write-once optical discs.

These file systems that use optical or magneto-optical disks are suitable as filing devices that store document information in the form of image data by taking advantage of their large capacity.6 Normally, image data is compressed. - Because it is encoded and stored, even if the input document size is the same, the amount of data differs for each image record on the storage medium. Furthermore, in these file systems, a code file that stores keywords, etc. for searching image files is usually created on the same medium as the image file; The amount of data is significantly smaller compared to .

Since data cannot be erased on conventional write-once optical discs, each data record to be registered as a file has only to be stored sequentially in sectors according to the physical order on the optical disc. However, in a file system using a magneto-optical disk on which data can be rewritten, new record data can be stored in a recording area obtained by erasing registered data. In this case, due to the compression encoding, the amount of data in the erased record does not necessarily match the amount of data in the newly registered record, so in order to efficiently store each record on the storage medium assuming rewriting, A data storage method is required.

For example, as a data storage method on a general rewritable storage medium such as a fixed magnetic disk or a floppy disk, see pages 91 to 102 of a document entitled rUNIXJ written by Haruhisa Ishida and published by Kyoritsu Shuppan. The data storage scheme described is known. According to the method described herein, a storage area forming a file is divided into a plurality of blocks of 512 bytes each, and the numbers of the plurality of blocks forming each record are managed as file management information. According to this method, one record is divided into multiple blocks based on the above-mentioned block size, and can be stored distributed in storage areas that are not necessarily physically contiguous on the storage medium. Each record can be stored space-efficiently under conditions where size changes occur.

[Problem to be solved by the invention]

However, when the conventional data storage method described above is applied to recording data with a large record size such as an image file, there is a problem that the efficiency of accessing each record decreases.6For example, in the case of black and white binary images used in facsimile , the amount of data is several 10 to 1 for an A4 document size.
00 kilobytes or more, and for color images or grayscale images such as photographs, the amount of data is several times to several tens of times that amount. Therefore, if the block size is set to the above-mentioned 512 bytes, 10 per roux code
Approximately 2 to 104 blocks are required. Therefore, access to a huge number of blocks is required for one record access, and the seek time of the read/write head increases due to block dispersion, resulting in a reduction in access speed.

In this case, in order to reduce the number of blocks per routine code, for example, in order to limit the number of accesses to a block within 10 times, each block size is set to about 5 to 500 kilobytes. There is a problem that data consumes a large storage area. Also, as mentioned above, in this type of image file, 1
Code files such as image keywords are stored in the same file.
These need to be managed on the system, but these are usually 10
Most files are between 0 bytes and 5 kilobytes, so with large blocks like the one above.

There was a problem that the space efficiency deteriorated to 11.

An object of the present invention is to provide a file management method that achieves high space efficiency and high access speed for each file on the premise of erasing and updating files of greatly different sizes.

[Means to solve the problem]

In order to achieve the above object, in the present invention, when one record is divided into multiple blocks and stored on a file medium, multiple types of block sizes are prepared in advance, and records are selected from among these. By selecting the block size according to the size, the number of blocks used for each record is optimized.

[Operation] According to the present invention, a large record such as an image data record is divided into large-sized blocks and stored, and small data consisting of code data such as a keyword is stored.
Since the records are divided into small blocks and stored, the space of the file medium can be used effectively and each record can be accessed at high speed.

〔Example〕

The present invention will be described below with reference to detailed drawings.7 Figure 1 is an overall configuration diagram of a file system for implementing the present invention, 11 is a processing device that controls the overall operation of this system, 12 13 is a keyboard used for inputting operation commands and record data, etc.; 14 is an image data input device for inputting image data; 15 is a file device for storing image data and code data, and 16 is a display device on which input images, search images, etc. are displayed.

FIG. 3 is a volume configuration diagram of the file device 15.
In the figure, 1 is a volume management area for storing volume management information that defines the position, size, etc. of each area, which will be described later, and 2 is a volume management area for storing volume management information that defines the position, size, etc. of each area, which will be described later. An area 3 is used to store defined file management information, and a record area 3 is used to store content data of each record (or file).

21 to 24 are individual file management information, 31
-34 represent record areas corresponding to 21-24, respectively. In this example, the basic size of the image data block is "5" and the basic size of the code/data block is "5".
For example, the record 31 consisting of one image data has three blocks (31-1 to 3l-3), and the record 32 consisting of code data has two blocks (32-1 to 32-2).
, record 33 consisting of code data consists of 2 blocks (
33-1 to 33-2), record 3 consisting of image data
4 indicates that it is composed of two blocks (34-1 to 34-2). Note that 35 indicates an empty area.

FIG. 3 is a table configuration diagram showing details of each file management information 2. In the figure, 51 is a column for storing information regarding file protection or users.

52 is a column that stores the record size, 53 is a column that stores the size of the blocks that make up the record, 54 is a column that stores the address of each block that makes up the record, and 55 is a column that stores the blocks other than 54 that make up the record. Column for storing pointers to each block address management area, 56
represents a column for storing additional information such as record creation, update date, etc. In this embodiment, the block basic size is set to be a multiple of the size for one image data record and the size for code data record, so the record size 1lI52 includes the block size for code data record. An integer value of size 1 can be stored. Second
In the illustrated example, the record sizes 52 in the file management information areas 21 to 24 are r15J and r2J, respectively.
, r2J, rlOJ, and the block sizes 53 are r5,1, rl, respectively.

For the same reason as above, the block address 54 can also be a number that is the size unit for the cop data record, and the record management information 21 can be set to "0" or "5". 5".

"20", rtoJ for record management information 22.

r14J, rHJ for record management information 23.

rl3j, record management information ff124 n5.

In this example, if the number of blocks constituting one record is within 10 blocks, only block address 54 is used, and for files exceeding 10 blocks, address 55 is used to duplicate the file. An indexing method is used.The double indexing method is detailed in the above-mentioned literature.

The process of adding a new record to the file system with the above configuration will be explained below.

FIG. 4 shows a flowchart of a process for determining an empty area in the record area 3 (for example, an erased area on a magneto-optical disk). In step 101, the record area is determined by referring to the volume management information 1. Find the address of 3. In step 103, the record area 3
In step 105, a variable i indicating the address used for the block address 54 and a variable K for counting consecutive empty areas (areas as one) in the record area 3 are initialized. It is determined whether the value currently pointed to by i is the end of the file, for example, by referring to volume management information 1. In step 107, the value currently pointed to by variable i is determined to be the end of the file. It is determined whether the portion equal to the block size is the above-mentioned empty area.If it is an empty area, the following steps 109 to 115 and 1 are performed.
Execute step 19. In step 109, the record area 3
Initializes table A rKJ for storing free space information within. Table A rKJ specifies the start address and area size of each continuous empty area as attributes for the blocks for code/data records.
It shall also be a unit of size. In step 111, the variable i is updated to be the next address, and in step 113, it is checked again whether the contents of the address are empty or not as in step 107.
Judgment is made in the same manner as . In the case of a region, in step 115,
Update the size attribute of table A rKJ and step 1
Repeat the process from step 13 onwards. If the area is not empty in step 107 or step 113, step 117 respectively.
Or 119 is executed, and the processing from step 105 onward is repeated. If it is determined in step 105 that the file area is the last, this process ends.

Note that the process for determining the empty area described above may be executed each time a new record is added, but it may be executed in advance.

It can also be stored in the volume management information 1 or in another file device. If the latter is adopted for a file system using a magneto-optical disk, the process of the above flowchart can be executed when the medium is installed in the system, and the data can be stored on another magnetic disk or semiconductor memory.

Next, a process for storing a new record in the empty area obtained through the above process will be described.

Figure 5 shows that after dividing records to be stored in a file medium into basic block sizes according to the record size,
This is a process of storing one block in one of the above empty areas. Steps 201 to 207 are processes for finding an empty area with a size equal to the block size, and Step 2
If the empty area cannot be found in step 03, then in steps 209 to 217, the smallest block among the empty areas whose size exceeds the block size is found. If this cannot be determined, in step 221 error processing is executed when the record area is insufficient. Through the above steps, the empty area in which the block should be stored is determined, and step 22
In step 5, data is actually stored in this location. At this time, the file management information 2 and free space table A rKJ are updated at the same time.

In the embodiment described above, multiple types of records with different block sizes are mixed in the record area 3, but the record area 3 is designed so that only records using blocks of the same size are stored in advance. If the areas are divided into multiple areas, the process of calculating the sky area described above can be
Steps 209-219 can be omitted.

〔Effect of the invention〕

As described above, according to the present invention, records or files of greatly different sizes, such as one image data record and one code data record, can be divided into multiple blocks using basic blocks of a size corresponding to the record size. Since it can be divided and managed, it has the advantage of being able to access large-capacity, rewritable storage media, such as magneto-optical disks, with good space efficiency and at high speed.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the overall configuration of an image file system, FIG. 2 is a diagram showing one embodiment of the file configuration according to the present invention, and FIG.
The figure shows a configuration diagram of a table that stores file management information, Figure 4 is a flowchart of the process of finding an empty area, and Figure 5 is a flowchart of the process of storing blocks that make up a file in the empty area found in Figure 4. be. 1... Volume management area, 2... File management area, 3... Record area.

Claims (1)

[Claims] 1. In a file system that stores multiple records with different amounts of data on the same storage medium, multiple types of basic block sizes with different sizes are determined in advance,
A basic block size is selected according to the data amount of the record to be stored, and the record is divided into a plurality of blocks determined by the selected basic block size and stored on the storage medium, and the basic block size is A data storage method in a file system characterized in that data is stored in a management information area for record access.