JP2001022715A - Data format conversion mechanism and conversion method under multi-platform environment - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To speed up conversion processing for converting data having a mainframe data format into data having an open system data format, and reduce the load on a host computer. An intermediate file of a disk subsystem is a file shared by host computers and has a mainframe data format. The open file 7 is a file having an open system data format. The data conversion mechanisms 12 and 13 of the channel adapter 11 are mechanisms for converting a data format by hardware. The data conversion mechanism 12 reads the intermediate file 6, converts it into the data format of the open file 7, and transfers it to the host computer 2. In addition, the data conversion mechanism 13 stores the open file 7
Is read, converted into the data format of the intermediate file 6, and transferred to the host computer 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data format between data having a mainframe data format and data having an open system data format in a multi-platform environment where a mainframe system and an open system coexist. More particularly, the present invention relates to a data format conversion mechanism that realizes high-speed conversion by hardware control.

[0002]

2. Description of the Related Art Conventionally, in a multi-platform environment, when a data format is converted between data having a mainframe data format and data having an open system data format, the data format is converted by software on the open system side. Has been converted.

FIG. 8 is a configuration diagram of a system in a multi-platform environment including host computers 1 and 2 and a disk subsystem 4. Reference numeral 1 denotes a host computer on the mainframe side. Reference numeral 2 denotes a host computer on the open system side, which has a code conversion function 3 of software. The storage device of the disk subsystem 4 stores a mainframe database 5, an intermediate file 6, an open file 7, and an open system database 8. Intermediate file 6 is
This file is shared by the host computer 1 and the host computer 2 and has a mainframe data format. The open file 7 is a file having an open system data format such as a UNIX file. (Note) UNIX is a registered trademark of X / Open in the United States and other countries.

When converting mainframe data into open system data, first, a utility of the host computer 1 reads data in the database 5 and writes it in the intermediate file 6. Next, the code conversion function 3 of the host computer 2 reads the intermediate file 6 and converts it into data in the open system data format.
Write to open file 7. Next, the utility of the host computer 2 reads the open file 7 and loads the data into the database 8. When the open system data is converted to the mainframe data, the conversion is performed in the reverse procedure. That is, the code conversion function 3 reads the open file 7, converts it into data in the data format of the mainframe, and then writes it into the intermediate file 6.

[0005]

According to the above-mentioned prior art, since the data conversion between the intermediate file 6 and the open file 7 is executed by the software of the host computer 2, the speed cannot be increased, and the host computer 2 is considerably slow. Load.

An object of the present invention is to speed up the conversion of the data format and reduce the load on the computer on the open system side.

[0007]

According to the present invention, a data format conversion function is realized by hardware separated from a host computer, thereby reducing the load on the host computer.

[0008]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a configuration diagram of the disk subsystem 9 of the present embodiment. The disk subsystem 9 is connected to a mainframe host computer 1 and an open system host computer 2. The mainframe channel adapter 10 has a channel interface of the host computer 1 and controls data transfer between the host computer 1 and a storage device. The open system channel adapter 11 has a channel interface of the host computer 2 and controls data transfer between the host computer 2 and a storage device.

[0010] The storage device is a magnetic disk drive,
It includes an optical disk drive and the like, and is recognized as a volume by the host computers 1 and 2. The mainframe database 5 is a database stored on a volume accessible only from the host computer 1. Database 8 for open systems
This is a database stored on a volume accessible only from the host computer 2. The open file 7 is a UNIX file or the like, and is a sequential file stored on a volume accessible only from the host computer 2. Intermediate file 6 is
Stored on a volume shared by the host computer 1 and the host computer 2,
2 is a sequential file accessible from both sides, and has a mainframe system record format.
The utility provided by the database management system of the host computer 1 reads the data of the database 5 and outputs the intermediate file 6,
And loads the data into the database 5. The utility provided by the database management system of the host computer 2 reads the data of the database 8 and outputs the open file 7, and reads the open file 7 and stores the data in the database 8
To load.

The channel adapter 11 has a data conversion mechanism 12 and a hardware mechanism of the data conversion mechanism 13. The data conversion mechanism 12 reads the intermediate file 6 having the data format of the mainframe, converts the data into data having the data format of the open system, and transfers the data to the host computer 2. The host computer 2 writes this data in the open file 7. Data conversion mechanism 1
3 receives the data of the open file 7 from the host computer 2 which has read the data, converts the data into data having a mainframe data format, and converts the data into an intermediate file 6.
To be stored.

In the disk subsystem 9, the channel adapter 10 and the channel adapter 11 are mounted on a disk controller (DKC), and the database 5, the intermediate file 6, the open file 7, and the database 8 are stored in the disk unit (DKU). Information.

The mainframe system and the open system use different record data formats and use different character codes. FIG. 2A is a diagram showing a record format of a mainframe system, and FIG. 2B is a diagram showing a record format of an open system. In the mainframe system, a record length is added to the beginning of a record, and EBCDIC is used for character data. On the other hand, in the open system, Carriage Retur
Record boundaries are indicated by inserting n (CR) or Line Feed (LF) codes, and character data uses ASCII codes. Other data indicates data other than character data such as binary data and hexadecimal data. In order to exchange data between the mainframe and the open system, it is necessary to perform both record format conversion and character code conversion. In order to convert the character code, the head position (byte position) and the data length of the character data must be specified in advance.

FIG. 3 is a diagram showing a hardware configuration of the data conversion mechanism 12 for converting mainframe data into open system data. The input buffer 18 stores the data input from the intermediate file 6, and extracts the record length stored at the head of the record. The byte number calculation circuit 16 counts the number of bytes of the data output from the input buffer 18 and compares it with the extracted record length. The CR / LF generation circuit 17 outputs the generated CR / LF when the record boundary is reached. The character detection circuit 34 detects that the data input to the input buffer 18 is character data. EBCDIC-A
The SCII character conversion circuit 19 converts the character data into ASCII codes on a byte-by-byte basis and outputs them.
The output buffer 20 outputs data after adding CR / LF to the end of the record.

The procedure for converting mainframe data into open system data is as follows. The channel adapter 11 transmits the intermediate file 6 from the host computer 2.
Receive a read command for In this command, the head position where the character data in the record exists and the data length are set. The channel adapter 11 inputs the data read from the intermediate file 6 to the input buffer 18 according to this command. Input buffer 18
Extracts the record length added to the head of the record and passes it to the byte number calculation circuit 16. Byte number calculation circuit 1
No. 6 detects the data input following the record length, counts the number of bytes, and compares it with the record length. The character detection circuit 34 detects a data byte input following the record length, and outputs an ON signal for the data length of the character data. As a result, data other than the character data in the input buffer 18 is directly transferred to the output buffer 20, and the output of the character data is directly suppressed and input to the EBCDIC-ASCII character conversion circuit 19 byte by byte. EBCDIC-ASCII
The character conversion circuit 19 converts the input character code into an ASCII code and outputs it. When the record boundary is reached, the byte count calculation circuit 16 outputs an ON signal, so that the CR / LF generation circuit 17 outputs CR / LF, and the output buffer 20 adds CR / LF to the record and sends it to the host computer 2. Forward. The host computer 2 writes the converted record in the open file 7.

FIG. 4 is a diagram showing a hardware configuration of the data conversion mechanism 13 for converting open system data into mainframe data. The input buffer 22 stores data input from the host computer 2. The character detection circuit 34 is the same as the character detection circuit 34 of the data conversion mechanism 12. CR / LF detection circuit 3
5 compares the data input to the input buffer 22 byte by byte with CR / LF, and detects CR / LF when C / LF is detected.
It outputs an R / LF detection signal and outputs a byte count up signal for data bytes other than CR / LF. A
The SCII-EBCDIC character conversion circuit 23 converts the character data into EBCDIC one byte at a time and outputs it. The record length calculation circuit 21 counts the number of bytes of data output from the input buffer 22 in accordance with the byte count up signal output from the CR / LF detection circuit 35, and counts when a record boundary is reached by CR / LF detection. Output the number of bytes as the record length. The output buffer 24 is the input buffer 22 or the ASCII-
The data output from the EBCDIC character conversion circuit 23 is received, and when a record boundary is reached, the data is output after adding a record length to the beginning of the record.

The procedure for converting open system data into mainframe data is as follows. The channel adapter 11 transmits the intermediate file 6 from the host computer 2.
Receive a write command to In this command, the head position where the character data in the record exists and the data length are set. The channel adapter 11 inputs the data received from the host computer 2 to the input buffer 22 according to this command. The CR / LF detection circuit 35 searches the data input to the input buffer 22 for CR / LF one byte at a time,
If it is not R / LF, the byte count up signal is turned on. The record length calculation circuit 21 counts the byte count up signal. The character detection circuit 34 detects a data byte input to the input buffer 22, and outputs an ON signal for the data length of the character data. As a result, data other than the character data in the input buffer 22 is directly transferred to the output buffer 24, the output of the character data is directly suppressed, and the character data is input to the ASCII-EBCDIC character conversion circuit 23 byte by byte. The ASCII-EBCDIC character conversion circuit 23 converts the code of the input character into an EBCDI
Convert to C and output. CR / LF detection circuit 35
When / LF is detected, the CR / LF detection signal is turned on, and the record length calculation circuit 21 outputs the counted number of bytes as the record length. Thereby, the data conversion mechanism 1
Reference numeral 3 adds a record length to the beginning of the record, and outputs one record of data stored in the output buffer 24 to the intermediate file 6.

FIG. 5 is a diagram showing a hardware configuration of the CR / LF detection circuit 35. 1-byte input data 25 read from input buffer 22 and CR / LF code 2
6 is compared by a comparison circuit 27. If they match, a CR / LF detection signal is output to indicate that the record boundary has been reached. If they do not match, a byte count up signal is output.

FIG. 6 is a diagram showing a hardware configuration of the EBCDIC-ASCII character conversion circuit 19.
The EBCDIC-ASCII character conversion circuit 19
When 1-byte EBCDIC is input, the input EBCDIC is converted by an address conversion circuit 33 into an address of a memory in which the code conversion table 30 is mounted.
A one-byte ASCII code accessed by the address is read from the code conversion table 30 and output.

FIG. 7 is a diagram showing a hardware configuration of the ASCII-EBCDIC character conversion circuit 23.
The ASCII-EBCDIC character conversion circuit 23
When a one-byte ASCII code is input, the input ASCII code is converted by an address conversion circuit 33 into an address of a memory in which the code conversion table 31 is mounted. Reads and outputs EBCDIC.

In the above embodiment, the channel adapter 11
Provided with a data conversion mechanism 12 and a data conversion mechanism 13,
The intermediate file 6 has data in the mainframe record format, and the open file 7 has data in the open system record format.
3, a file for storing mainframe record format data is provided in place of the open file 7,
The present invention can be implemented even when the intermediate file 6 has data in the open system record format. Further, there is no problem even if the data conversion mechanism 12 and the data conversion mechanism 13 are configured as components of the disk subsystem 9 independent of the channel adapter 10 or the channel adapter 11.

Further, a device incorporating a data conversion mechanism 12 and a data conversion mechanism 13 is interposed between the host computer 2 and a channel adapter for the open system of the conventional disk subsystem 4, and the data format conversion is performed by this device. May go. This converter and the host computer 2 are connected by a channel interface of the host computer 2, and the channel adapter of the disk subsystem 4 is connected by the same channel interface of the host computer 2. When the host computer 2 issues a data conversion command by designating the intermediate file 6 and the open file 7 to the converter, in the case of a data conversion command from the mainframe to the open system, the converter starts the data conversion mechanism 12. The format of the record read from the intermediate file 6 is converted and written into the open file 7. In the case of a data conversion command from the open system to the mainframe, the conversion device activates the data conversion mechanism 13 and the open file 7
The format of the record read from is converted and written to the intermediate file 6.

[0023]

As described above, according to the present invention, data format conversion between data having a mainframe data format and data having an open system data format is executed by hardware. The data input to the mechanism is converted at a speed substantially close to the transfer speed of the data stream, so that the conversion of the data format can be speeded up and the load on the host computer can be reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a disk subsystem according to an embodiment.

FIG. 2 is a diagram showing data formats of a mainframe and an open system.

FIG. 3 is a diagram illustrating a configuration of a mechanism for converting mainframe data into open system data according to the embodiment;

FIG. 4 is a diagram illustrating a configuration of a mechanism that converts open system data into mainframe data according to the embodiment;

FIG. 5 is a diagram showing an internal configuration of a CR / LF detection mechanism 35 of the embodiment.

FIG. 6 is a configuration diagram of an EBCDIC-ASCII character conversion circuit 19;

FIG. 7 is a configuration diagram of an ASCII-EBCDIC character conversion circuit 23;

FIG. 8 is a diagram illustrating a conventional data format conversion.

[Explanation of symbols]

1: host computer (mainframe), 2: host computer (open), 6: intermediate file, 7:
Open file, 10: channel adapter (mainframe), 11: channel adapter (open), 1
2: Data conversion mechanism, 13: Data conversion mechanism

Claims (5)

[Claims] 1. A conversion mechanism for converting a data format between data having a data format of a mainframe and data having a data format of an open system, comprising: a first mechanism for converting mainframe data into open system data; And a second conversion mechanism for converting the open system data into mainframe data. The first conversion mechanism receives the mainframe data, detects the head record length of the record, A means for counting the number of input bytes following the length to detect a record boundary, and input data other than character data is transferred and output as it is, and when character data is detected, the output is suppressed as it is and code conversion is performed byte by byte. Means for outputting later, and adding a CR / LF code to the end of the record when a record boundary is detected Means for inputting open system data, counting input record length until a CR / LF code at the end of the record is detected, and detecting a record boundary; Means for directly storing the input data in the output buffer and, when character data is detected, suppressing the output as it is, converting the code one byte at a time and storing it in the output buffer, and means for storing the output buffer when a record boundary is detected. Means for adding the record length to the beginning of the record and outputting the record. 2. A disk subsystem incorporating the data format conversion mechanism according to claim 1. 3. A method of reading data on one of a file having a data format of a mainframe and a file having a data format of an open system, converting the data format, and writing the data on the other file. 2. The data format conversion mechanism according to claim 1, wherein: 4. A method of converting a data format between data having a mainframe data format and data having an open system data format, the method comprising: converting data having a mainframe data format. Detects the record length at the beginning of a record by inputting mainframe data, counts the number of input bytes following the record length, detects record boundaries, transfers and outputs input data other than character data as is, Is detected, the output is suppressed as it is, the code is converted one byte at a time and output. When a record boundary is detected, a CR / LF code is added to the end of the record, and the data having the open system data format is added. When converting, input the open system data and detect the CR / LF code at the end of the record. Counts the input record length detecting a record boundary to the input data other than character data is stored in it an output buffer, upon detection of the character data, to suppress the raw output,
A data format conversion method comprising: converting a code one byte at a time, accumulating the result in an output buffer, and adding a record length to the beginning of the record in the output buffer when a record boundary is detected, and outputting the record. 5. A method of reading data on one of a file having a data format of a mainframe and a file having a data format of an open system, converting the data format, and writing the data on the other file. 5. The data format conversion method according to claim 4, wherein: