JPH04311216A - External storage controller - Google Patents

Abstract

PURPOSE:To surely increase the accessing speed to an external storage device by always effectively functioning a look-ahead cache system. CONSTITUTION:A block management memory 9 stores unused area management information which manages the unused area of a buffer memory 8 by blocks and data management information which manages data by blocks. A CPU 6 discriminates whether or not read requesting data are stored in the memory 8 based on the data management information and, when the data are stored, accesses the read requesting data in the memory 8. When the data are not stored in the memory 8, the CPU 6 accesses the data in an HDD 3. When the data in the HDD 3 is accessed, the CPU 6 reads look-ahead data from the HDD 3 and constitutes a block of the look-ahead data and read requesting data, and then, stores the block in an unused area.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an external storage control device for controlling access to an external storage device such as a hard disk drive in a computer system and having a cache memory system.

0002

2. Description of the Related Art Conventionally, computer systems have used external storage devices such as hard disk drives (HDD). The HDD has a large capacity storage medium and is normally used as a file device.

[0003] An HDD is a disk controller (HDC).
) to read/write data from/to the host computer. When the HDC receives a read command (access command) from the host computer, the HDC executes control to access data in the sector of the read request from the HDD and transfer it to the host computer.

By the way, external storage devices such as HDDs have a data access speed lower than that of the main memory of the system. For this reason, the host computer
It takes a considerable amount of time to access and read the requested data.

[0005] In order to solve these problems, HDCs employing a cache memory system have been developed. This HDC stores some of the data stored on the HDD (
(in sector units) is stored in buffer memory. When receiving an access command from the host computer, the HDC determines whether the requested data is stored in the buffer memory. If it is stored (hit), the HDC accesses the requested data from the buffer memory and transfers it to the host computer.

On the other hand, if the requested data is not stored, the HDC controls the access of the requested data from the HDD, transfers it to the host computer, and stores it in the buffer memory. At this time, there is a pre-read cache method in which the HDC accesses a sector consecutive to the sector of the requested data as pre-read data and stores it in a buffer memory. This increases the probability that the data that the host computer requests access to will be stored in the buffer memory, and as a result, it becomes possible to increase the access speed of the HDD.

[0007]

SUMMARY OF THE INVENTION By using an HDC that employs a cache memory method, it is possible to increase the data access speed of an external storage device such as an HDD. By the way, the buffer memory that HDC has is usually FIF
It is controlled by an O (first-in-first-out) method. Therefore, when the above-mentioned read-ahead cache method is adopted, when read requests from the host computer are executed alternately to different locations (sectors of the recording medium) on the HDD, the data stored in the buffer memory A situation occurs in which the read-ahead data becomes invalid.

That is, when a read request from a host computer is executed sequentially to a plurality of positions on a recording medium, each time the read position moves, the pre-read data previously stored in the buffer memory is Disabled by lead. Therefore, especially when the host computer accesses the HDD under the control of a multitasking OS, it may not be possible to increase the access speed of the HDD.

An object of the present invention is to enable the read-ahead cache method to function effectively even when a plurality of access positions in the external storage device are accessed alternately in an external storage control device that employs the read-ahead cache method. The object of the present invention is to reliably increase the access speed of an external storage device.

0010

[Means for Solving the Problems] The present invention provides an external storage control device that controls external storage means in response to instructions from a host computer. It has buffer memory means for storing request data and pre-read data. The buffer memory means temporarily stores data in units of blocks, each consisting of a plurality of sectors, for example.

[0011]Furthermore, the present device includes block management memory means for managing data stored in the buffer memory means and unused areas, and a first control unit for controlling access to data in response to read requests from the host computer. and second control means for controlling when the read request data and pre-read data are stored in the buffer memory means.

0012

[Operation] In the present invention, the block management memory means stores unused management information for managing unused areas on a block-by-block basis in the storage area of the buffer memory means, and stores unused management information on a block-by-block basis stored in the buffer memory means. Stores data management information for managing data. The first control means is
In response to a read request from the host computer, it is determined whether the read request data is stored in the buffer memory means based on the data management information stored in the block management memory means, and if it is stored, the buffer The read request data is accessed from the memory means, and if it is not stored, the read request data is accessed from the external storage means. Furthermore, when the first control means accesses from the external storage means, the second control means reads data in the access unit related to the read request data from the external storage means as pre-read data, and reads the pre-read data and the read request data. It forms a block unit with data and stores it in an unused area of the buffer memory means indicated by the unused management information.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an external storage system according to the same embodiment. This system consists of a host computer 1, which is the main body of the computer system, and a hard disk drive (HDD) 3, which is an external storage device.
and H, which is an external storage control device that controls the HDD3.
Consists of DC2.

[0015] The HDC 2 has a host interface 4,
Disk interface 5, CPU 6, control memory 7
, a buffer memory 8 and a block management memory 9. The host interface 4 is connected to the host computer 1
This is an interface for transferring data, commands, etc. between The disk interface 5 is an interface for transferring data between the HDD 3 and the HDD 3 .

The CPU 6 constitutes first and second control means, and is a microprocessor that performs overall control of the HDC 2. The control memory 7 is a ROM that stores programs for the CPU 6 to perform control operations. The buffer memory 8 is a memory for executing a pre-read cache function, and is controlled by the CPU 6. The block management memory 9 is a memory that stores management information for the buffer memory 8 related to the gist of the present invention. Next, the operation of this embodiment will be explained.

First, as shown in FIG. 4(B), the HDD 3 has a plurality of disks (recording media) 10 and 11.
Under the control of HDC2, read/read data in sector units.
Do a light. As shown in FIG. 4A, a sector is a unit recording area obtained by partitioning a track 12 in a plurality of tracks provided on each surface of the disks 10 and 11. This sector unit becomes the access unit.

As shown in step S1 in FIG. 2, when a read command (read request) is output from the host computer 1, the HDC 2 receives the read command through the host interface 4.

The CPU 6 determines whether the read request data (target sector) is stored in the buffer memory 8 or not. That is, the CPU 6 searches the block management memory 9 and determines whether management information corresponding to the read request data exists (step S2).

As shown in FIG. 3B, the block management memory 9 stores a management list of prefetch blocks (data management information) and a management list of unused blocks (unused management information). The prefetch block management list is an information table for managing prefetch data (cache data) in block units. Here, one block consists of, for example, 4 sectors, and 4 sectors consecutive from the first sector.
Consists of sectors. The prefetch block management list has data management information for each block. This data management information includes the first sector number (for example, sector A), the total number of blocks read ahead (for example, 4), the block number of the first four sectors (for example, 2), the block number (N) of the next four sectors, and the third block number (for example, 4). It consists of the block number of the sector and the block numbers of the last four sectors.

On the other hand, the buffer memory 8 is shown in FIG.
As shown in , it has a storage area in block units (block numbers 1 to N), and stores pre-read data of blocks according to a management list of pre-read blocks. That is, for example, in the area of block number 2, the first sector numbers A to A
Four sectors of data up to +3 are stored. In addition,
Block numbers 1 and 4 are empty areas, respectively, and are unused areas that are not currently used as storage areas for pre-read data.

If the target sector of the read request from the host computer 1 is, for example, sector A (YES in step S3), the CPU 6 accesses the area of block number 2 from the buffer memory 8 and reads the data of sector number A. and transfers it to the host computer 1 (step S7).

On the other hand, if it is determined that the target sector does not exist in the buffer memory 8 based on the data management information (NO in step S3), the CPU 6 accesses the HDD 3 through the disk interface 5 (step S4). . That is, the CPU 6 accesses the requested sector from the disks 10 and 11 of the HDD 3. The CPU 6 transfers the requested sector data read from the HDD 3 to the host computer 1.

At the time of this access, the CPU 6 searches the unused block management list (unused management information) in the block management memory 9 to find an unused block (empty area) in the buffer memory 8. The CPU 6 stores the data of the requested sector read from the HDD 3 in the searched empty area of the buffer memory 8 (step S5). The management list of unused blocks is as shown in (B) of FIG.
Management that indicates the total number of unused blocks in the buffer memory 8 and the area in which new data can be stored from the first entry (for example, block number 1), second entry (for example, block number 4) to the Nth entry of unused blocks. Consists of information.

Here, if the read command from the host computer 1 is continuing (YES in step S6)
, the CPU 6 executes a prefetch process of prefetching data in a plurality of consecutive sectors (four sectors), using the requested sector as the first sector (step S8). At this time, CPU6
searches for an unused block based on the unused management information in the block management memory 9, and stores the pre-read data in the empty area of the unused block number in the buffer memory 8 (steps S9 and S10). In this case, the CPU 6 stores four sectors worth of pre-read data (including the requested sector) in the empty area of block number 1 of the first entry that stores the requested sector. If the read command is continuing, CP
U6 further reads four sectors worth of pre-read data from the HDD 3 and stores it in the empty area of block number 4 of the second entry of the buffer memory 8, for example.

By the way, when the read request data (request sector) from the host computer 1 is stored in the buffer memory 8 (hit), the CPU 6 uses the data from the buffer memory 8 as read data from the HDD 3 to read from the host computer 1. Transfer to. After this transfer, C
The PU6 stores the read block in the block management memory 9.
, and register it as a new unused block in the unused block management list.

In this way, in response to a read request from the host computer 1, if the requested data (requested sector) is stored in the buffer memory 8, it is accessed from the buffer memory 8 and transferred to the host computer 1. . Therefore, by accessing from the buffer memory 8 without accessing the HDD 3, data can be accessed at high speed.

Next, management is performed to store pre-read data in blocks in the buffer memory 8. This has the following effects. That is, as shown in FIG. 4(B), when read requests are made alternately to sectors A and sectors B at different positions in the same cylinder N of disks 10 and 11, continuous pre-read data in each sector A and B is can be used effectively. Specifically, as shown in (A) of FIG. 3, when sector A is accessed from HDD 3, pre-read data (A+1 to A+3) is stored in block number 2, and when sector B is accessed from HDD 3. , the pre-read data (B+1 to B+3) is transferred to block number 3.
Store in. Therefore, when next requests are made to read sectors A and B alternately, each data can be accessed from the buffer memory 8. In other words, when read requests are made alternately for sectors A and Sector B, a situation in which the pre-read data of one is deleted from the buffer memory 8 and becomes invalid does not occur.

In the above-mentioned embodiment, when a non-sequential read request is output while processing of pre-read data is continued, there is the following method for controlling the end of the pre-read processing. The first method is to set a predetermined threshold (number of sectors) for a block in advance, and then continue read-ahead processing until the end of the block based on this threshold, or invalidate the block and return it to an unused block. control. For example, when the threshold value is m/2 sectors, if half or more of the block has been read, the prefetch process is continued until the end, and if it is less than half, the prefetch process is interrupted. As a second method, a minimum number of prefetch blocks is set in advance, and if the prefetch process is progressing beyond this minimum number of prefetch blocks, the first method described above is used.
Apply the method. Otherwise, the prefetch process continues until the minimum number of prefetch blocks is reached.

Next, there are two methods for determining the maximum number of read-ahead blocks in the read-ahead process: a method in which the maximum number of read-ahead blocks is fixedly set, and a method in which the maximum number of read-ahead blocks is set using a function that takes the read data length as input. be.

Furthermore, as a control method when there are no free blocks (unused blocks) during prefetch processing,
There is a method that immediately stops the read-ahead process when there are no more free blocks.

[0032]

As described in detail above, according to the present invention, in an external storage control device that employs a read-ahead cache method, the read-ahead data stored in the buffer memory is managed block by block, thereby making it possible for the host computer to Even when read requests are executed alternately to different storage locations in the external storage device, the pre-read data can always be used effectively. Therefore, as a result, it is possible to reliably achieve faster access speeds to the external storage device.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an external storage system according to an embodiment of the present invention.

FIG. 2 is a flowchart for explaining operations related to the embodiment.

FIG. 3 is a conceptual diagram for explaining the contents of a memory according to the same embodiment.

FIG. 4 is a conceptual diagram for explaining the structure of a hard disk device according to the same embodiment.

[Explanation of symbols]

1...Host computer, 2...HDC, 3...HDD, 6
...CPU, 8...Buffer memory, 9...Block management memory.

Claims (2)

[Claims] 1. An external storage control device that controls access to external storage means in response to instructions from a host computer, wherein a block unit consisting of a plurality of access units is selected from among a data group stored in the external storage means. buffer memory means for temporarily storing data; and unused management information for managing unused areas in the storage area of the buffer memory means in units of blocks;
and block management memory means storing data management information for managing the data in units of blocks stored in the buffer memory means; It is determined whether the read request data is stored in the buffer memory means based on data management information, and if it is stored, the read request data is accessed from the buffer memory means, and if it is not stored, the read request data is accessed from the buffer memory means. a first control means for accessing the read request data from the external storage means; and a first control means for accessing the read request data from the external storage means; read from the external storage means as pre-read data,
A second control means for configuring the block unit by the pre-read data and the read request data and storing it in the unused area of the buffer memory means indicated by the unused management information. external storage controller. 2. An external storage control device that controls access to external storage means in response to instructions from a host computer, wherein a block unit consisting of a plurality of access units is selected from a data group stored in the external storage means. buffer memory means for temporarily storing data; and unused management information for managing unused areas in the storage area of the buffer memory means in units of blocks;
and block management memory means storing data management information for managing the data in units of blocks stored in the buffer memory means; It is determined whether the read request data is stored in the buffer memory means based on data management information, and if it is stored, the read request data is accessed from the buffer memory means, and if it is not stored, the read request data is accessed from the buffer memory means. a first control means for accessing the read request data from the external storage means; and a first control means for accessing the read request data from the external storage means; read from the external storage means as pre-read data,
a second control means for configuring the block unit with the pre-read data and the read request data and storing it in the unused area of the buffer memory means indicated by the unused management information; The means continues the read operation of the pre-read data up to the block unit based on a predetermined rule when a read request unrelated to the read request data occurs while the read operation of the pre-read data is continuing. or third control means for executing end control of a pre-read data read operation so as to interrupt the read operation and invalidate the read pre-read data.