JPH0577101B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a computer system, and in particular, constructs an index sequential file in an auxiliary storage device, and stores key blocks of an index group of this index sequential file in each hierarchy. This invention relates to an indexed sequential file system having a buffer area for separate storage.

[Conventional technology]

Figure 4 is a schematic configuration diagram of a computer system having this type of indexed sequential file system. ), 3 is a main storage device (hereinafter referred to as memory) that stores programs, data, etc. and also reserves a buffer area for index sequential files (hereinafter referred to as IS files), and 4 is used for exchanging data between computer systems. The interface 5 is a disk device (hereinafter referred to as disk) as an auxiliary storage device that stores a large amount of data as files.

Among the operations of this computer system, the case where IS files are used will be briefly explained.

First, it is assumed that an IS file has been constructed on the disk 5, and a user of this computer system 1 creates a user program to perform a desired process, loads it into the memory 3, and executes it.

Next, when the executed user program accesses the IS file, it issues a request to the indexed sequential file system (referred to as the IS system) and accesses the IS file.
The system accesses the IS file built on the disk 5 via the disk interface 4 and returns the result to the user program.

In this way, the user accesses the IS file on the disk 5 from the user program via the IS system and implements the desired processing.

FIG. 5 is a conceptual diagram of the IS file described here, and the IS file 11 arranges records 14, which are a combination of key items (hereinafter referred to as keys) 12 and data 13, in ascending order (descending order is also possible) by key. It is a comparison. Ascending order in this case means only logically ascending order, but not necessarily physically arranging in ascending order. IS systems usually handle a plurality of such IS files at the same time, and there are two types of access: sequential access and direct access. Among these, sequential access is accessing the records 14 in ascending or descending order with respect to the key 12, and direct access is accessing the record 14 having the key 12 equal to that value by specifying the value of the key.

FIG. 6 shows the physical configuration of the IS file, which is mainly divided into a data group 11a and an index group 15.

Of these, the data group 11a is a collection of the records 14 described above, and in some cases, the data group 11a
4 are collected in ascending order with respect to key 12 to form a record block, and this is the data group 11a.
There are also IS files that make up the .

On the other hand, the index group 15 has a three-level structure including a first index 16 at the highest level, a second index 17 at the middle level, and a third index 18 at the lowest level in order to improve access efficiency. Among these, the third index 18 at the lowest level is a large number of key blocks 21 in which keys 12 of records 14 are assembled in ascending order, and each key is linked to the corresponding record 14 by some means, and the key blocks 21 are also linked in ascending order. has been done. Also, the middle second index 17 is the key block 2.
It is made up of a plurality of key blocks 20 that collect the largest keys in one and are connected in ascending order. The first index 16 at the highest level is the key block 20.
One key block 1 that collects the largest keys in
It is familiar with 9.

Next, we will explain how to access this IS file.

In the case of sequential access, an arbitrary record 14 is first specified by the user program. Subsequently, when a sequential access request is received from the user program, the record linked to the key next to the third lowest index key linked to this record 14 is accessed. Then, each time there is a sequential access request from the user program,
The above operation is repeated and record 14 is changed to key 12.
are accessed in ascending order (or may be accessed in descending order).

Here, when referring to the key next to the key in the third index 18 at the lowest level, if there are no more keys in the key block 21, the first key in the next key block must be referred to. In this example, since the keyblocks 21 are connected, the next keyblock can be found without referring to the upper indexes 16 and 17, thereby increasing access efficiency.

On the other hand, in the case of direct access, the user program specifies the value of a certain key and requests direct access. At this time, a key that is equal to or greater than the first index 16 at the highest level with the specified value is found, and among the key blocks 20 of the second index 17 at the middle level to which that key is connected, Find keys that are also equal or greater. When the key block of the lowest index 18 is reached, a key having a value equal to the designated value is found and the record 14 to which that key is linked is accessed. Note that, in practice, these accesses are executed by the IS file, so the user program does not need to be concerned with them.

Next, the buffer area for increasing the access efficiency of the IS system will be explained.

In the IS system, in order to improve the access efficiency of IS files, a buffer area corresponding to the size of the keyblock is secured for each IS file and for each hierarchy of the index group, and the keyblock once read is stored in each IS file.
There are things held for each layer of the file.
This will increase access efficiency, but
When using many IS files at the same time, the capacity of the buffer area becomes extremely large. FIG. 7 is an explanatory diagram showing the buffer area secured state when 30 IS files with the above-described three-layer structure are used, and a total of 90 buffers (B ₁₀₁ to B ₃₃₀ ) are required.

In addition, some IS systems reserve one buffer area for each layer and use this buffer for all IS files in order to prevent the capacity of the buffer area from increasing. In this way, no matter how many IS files are used at the same time, the number of buffers will be small.
If it is an IS file, as shown in Figure 8, a total of 3
buffer areas (B ₁₀₁ , B ₂₀₁ , B ₃₀₁ ) are sufficient.
However, although this configuration certainly increases buffer efficiency when only one IS file is used, if two or more IS files are used at the same time and must be accessed alternately, the contents of the buffer are The keyblock becomes the keyblock of another IS file, and the keyblock must be read out again, which reduces access efficiency.

[Problem that the invention seeks to solve]

Among the conventional IS systems mentioned above, access efficiency can be improved by securing a buffer area for the size of the key block for each IS file and for each hierarchy of the index group of each IS file, but access efficiency can be improved because many IS files are handled. In some cases, the capacity of the buffer area becomes too large, and in computer systems whose memory capacity is not so large, there is a problem in that it puts pressure on the user program.

On the other hand, if you allocate only one buffer area for each hierarchy that is common between the index groups of each IS file, there is no risk of overloading the user program, but because the IS files are handled alternately, However, there was a problem in that access efficiency was significantly reduced.

This invention was made in order to solve the above-mentioned problems of the conventional system, and it is possible to appropriately secure a buffer area according to the user's usage pattern, thereby reducing the pressure on user programs and reducing access efficiency. The purpose is to obtain an IS system that can suppress the degradation.

[Means for solving problems]

The IS system according to the present invention is provided with area setting means for changing the required capacity of a buffer area for holding key blocks of a common hierarchy among index groups of a plurality of index sequential files.

[Effect]

This invention is based on the IS files that are handled simultaneously in the IS file.
You can increase or decrease the capacity of the buffer area that holds the key blocks of the common hierarchy between the index groups of each IS file depending on the amount of files, so when the number of IS files handled at the same time increases, the capacity of the buffer area can be increased. This will improve access efficiency, and if there is a risk of overloading the user program, the capacity of the buffer area can be reduced before accessing.

Therefore, the user can freely set the usage pattern, thereby suppressing the burden on the user program and the decrease in access efficiency.

〔Example〕

FIG. 1 is a schematic configuration diagram of a computer system according to an embodiment of the present invention, and the same reference numerals as in FIG. 4 indicate the same elements. A switch 6 serves as an area setting means that can set the capacity of the buffer area that holds the key blocks of the common hierarchy among the index groups of each IS file by switching the setting method for each hierarchy in four stages. The difference from FIG. 4 is that an interface 7 for importing the state of this switch 6 into the computer system 1 is newly added.

Here, the operation of this embodiment will be described below assuming that the physical configuration of the IS file is the same as shown in FIG.

First, when the switch 6 is set to the first stage, the areas of the first index buffer B ₁₀₁ , the second index buffer B ₂₀₁ , and the third index buffer B ₃₀₁ are secured in the memory 3 as shown in FIG. be done. Therefore, this stage has the lowest buffer capacity and the lowest access efficiency.

Next, assume that the switch 6 is set to the second stage. As shown in FIG. 2, the buffer area reservation state at this stage is such that each area of the first index buffer B ₁₀₁ , the second index buffer B ₂₀₁ , and the third index buffer B ₃₀₁ to _{B 330} is secured. At this stage, only the lowest index hierarchy is secured for each IS file, but the other index hierarchies are secured as common to all IS files. In this way, the buffer capacity becomes smaller than when the switch 6 is set to the first stage.
In the case of sequential access, if you only refer to the key next to the key in the keyblock, efficiency is reduced because it is guaranteed that the keyblock at the lowest index level will always remain in the buffer once it is read. It gets expensive. However, in the case of direct access, it is the same as setting the switch 6 to the first stage.
Note that when using 30 IS files, 32 buffers are required.

Next, assume that the switch 6 is set to the third stage.
As shown in FIG. 3, the buffer area secured state at this stage is such that each area of the first index buffer B ₁₀₁ to B ₁₃₀ , the second index buffer B ₂₀₁ , and the third index buffer B ₃₀₁ to _{B 330} is secured. . That is, the highest and lowest index hierarchies are secured for each IS file, and the middle index hierarchies are secured as common to all IS files. In this way, the buffer capacity becomes the second largest after setting the switch 6 to the fourth stage. When considering access efficiency, in sequential access, it is as efficient as setting it to the second stage, and in direct access, once the key block in the top index hierarchy is read, it always remains in the buffer, so When searching for a key in an index group using a value specified by the user, efficiency is improved because it is not necessary to read out the key blocks of all index hierarchies for the highest index. Furthermore, when using 30 IS files, a total of 61 buffers are required.

Next, assume that the switch 6 is set to the fourth stage. This stage is exactly the same as that described using FIG. 7, and the buffer capacity is at its maximum and the access efficiency is at its highest.

The user selects the stage of use of the switch 6, taking into account memory capacity, access efficiency, and usage format (for example, sequential access only, direct access only, or not using more than one IS file at the same time). do.

Thus, according to this embodiment, the buffer area can be appropriately secured according to the usage pattern of the user.

In the above embodiment, the required capacity of the buffer area was set using a switch, but even if the required capacity is set by instructing the CRT display instead of the switch, the buffer area can be changed in the same way as described above. can.

Further, in the above embodiment, the required capacity of the buffer area is set in four stages using a switch, but the number of stages may be increased or decreased.

Furthermore, in the above embodiment, the case where the index group of the IS file has three hierarchies was explained.
The present invention is not limited to this, and the present invention can be applied even if the index group has a higher or lower hierarchy.

〔Effect of the invention〕

As explained above, according to the present invention, since the area setting means is provided which can set the required capacity of the buffer area of the main storage device in multiple stages, the user can select the state that suits his usage style and requirements. By adjusting the IS system, it is possible to utilize the computer system without wasting it, thereby reducing the burden on user programs and reducing access efficiency.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram of a computer system according to an embodiment of the present invention, FIGS. 2 and 3 are explanatory diagrams for explaining the state of buffer area reservation in the same example, and FIG. 4 is a conventional index A schematic configuration diagram of a computer system with a sequential file system. Figure 5 is a conceptual diagram of an index sequential file system. Figure 6 is a schematic diagram of an index sequential file system. Figures 7 and 8 are conventional index files. FIG. 2 is an explanatory diagram for explaining the state of buffer area reservation in the sequential file system. 1... Computer system, 2... Central processing unit,
3...Main storage device, 5...Disk device, 6...
Switch, 11... Index sequential organization file, 11a
...Data group, 15...Index group. In addition, in each figure,
The same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] 1 A plurality of sequential index files are constructed in a randomly accessible auxiliary storage device, and a buffer area for holding key blocks for each layer common among the index groups of the plurality of control sequential index files is provided in each layer. A computer system that secures key blocks in a main storage device according to the present invention, characterized by comprising an area setting means for changing the required capacity of a buffer area for holding key blocks of a common hierarchy among index groups of the plurality of index sequential files. indexed sequential file system.