JP3838195B2 - Information recording device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an information recording apparatus for recording information on a recording medium such as a magnetic tape.
[0002]
[Prior art]
A so-called tape drive is known as a drive device capable of recording / reproducing digital data on a magnetic tape. Such a tape drive can have a huge recording capacity of, for example, several tens to several hundreds gigabytes, depending on the tape length of a tape cassette as a medium. Widely used in applications such as backing up data recorded on media. It is also suitable for use in storing image data having a large data size.
[0003]
As a tape drive as described above, for example, a recording medium that records and reproduces data by adopting a helical scan method using a rotating head using an 8 mm tape as a recording medium is proposed and commercialized. ing. (Standard ECMA-292, AIT2 format).
[0004]
In the tape drive as described above, the data transferred from the host computer through the SCSI interface is packed in a predetermined unit (Basic Group) to form a logical block, and an error correction code is configured and recorded for each logical block.
[0005]
In the tape drive format, it is important to increase the reliability by reducing the redundancy of error correction parity. In other words, if the recording density per unit area of the magnetic tape is the same, the recording capacity of the data transferred from the host computer can be increased in the format in which the redundancy of the error correction parity is small. Even if the redundancy of the error correction parity is the same, the error correction capability for a burst error varies depending on how the data is distributed on the magnetic tape.
[0006]
A data flow of the AIT2 format will be described as a conventional tape drive device.
(1) Generate Basic-Group in a predetermined unit.
(2) Basic-Group is divided into 18 to generate G1 Sub-Group.
(3) G2 Sub-Group generation by dividing G1 Sub-Group into two bytes by byte
(4) G2 Sub-Group is equally divided to generate 288 blocks
(5) Divide 288 blocks into 6 G3 Sub-Groups by 48
(6) Generate a product code with 8 outer code parities and 12 inner code parities added for each G3 Sub-Group.
(7) Interleave the six product codes (described later) and record in one track
12A is a schematic diagram showing product codes in the AIT2 format, and FIG. 12B is a schematic diagram showing the relationship between the product codes and tracks.
[0007]
The product code in the AIT2 format is composed of 56 inner code blocks (in the horizontal axis direction in FIG. 12A), and 8 parity is added to the outer code parity. (Inner code block numbers 0 to 3 and 52 to 55)
Then, six product codes having the configuration of FIG. 12A are collected, interleaved for each product code, and recorded on one track.
[0008]
As shown in FIG. 12B, the inner code blocks of each product code are arranged on the track at a period of 6 blocks. Thus, for example, even when there is a burst error having a size corresponding to six inner code blocks in succession on a track, only one inner code block is erroneous in each product code unit. As described above, AIT2 has six product codes per track, and by interleaving and recording them, by distributing errors generated in a part of the track to six product codes, The track format is robust against burst errors.
[0009]
[Non-Patent Document 1]
Standard ECM-292 No. December 1999, pages 45-81 (Standard ECMA-292 1999.12 P45-P81)
[0010]
[Problems to be solved by the invention]
However, since the conventional configuration is composed of 288 (6 * 48) data blocks per track and 48 (6 * 8) outer code parity blocks, a burst error of 11% or more per track. If it occurs, there is a problem that error correction becomes impossible.
[0011]
In order to solve the above-described problems, an object of the present invention is to provide a format that enables error correction with a relatively small redundancy even when a large burst error occurs in the scanning direction of a magnetic head.
[0012]
[Means for Solving the Problems]
  To solve the above problems, Claim 1The present invention is an information recording apparatus for recording data on a track on a magnetic tape by a head, comprising a packet in which data is packed in a predetermined unit, and dividing the packet into N pieces (N Is an integer of 2 or more), each sector has L (L is an integer of 2 or more) inner code blocks, and L inner code blocks of each sector are M (M Are arranged substantially evenly on tracks of 2 or more integers). When L / M is not an integer, N, M, and L are set so that (N * L) / M is a natural number.
[0013]
With the above configuration, since (N * L) inner code blocks constituting N sectors having a product code configuration can be distributed substantially evenly to M tracks, a burst error of one track can be reduced to M. To provide a format having strong error correction capability against burst errors occurring in a track with a relatively small redundancy, which can use all the error correction capability of N product codes distributed to a plurality of tracks Is possible.
[0015]
  Claim2The invention of claim 1 is characterized in that, in the invention of claim 1, each track has (L / M + 1) or less inner code blocks constituting each sector.
[0016]
As a result, even when the number L of inner code blocks per unit sector is not divisible by the number M of recording tracks, N * L inner code blocks can be distributed evenly to M tracks. Burst errors can be made using all the error correction capabilities of N product codes.
[0017]
  Claim3The information recording apparatus according to the invention is an information recording apparatus for recording data on a track on a magnetic tape by a head, comprising a packet in which data is packed in a predetermined unit, and the packet is divided into M tracks (M is an integer of 2 or more) N segments (N is an integer equal to or greater than 2) are generated by dividing the packet and product code is applied, and each sector is composed of L1 (the outer code data portion is a component). L1 is an integer equal to or greater than 2) and L2 (L2 is an integer equal to or greater than 2) second inner code blocks having an outer code parity part as a component. The first and second inner code blocks are1 Under the condition of L2 or less per track, the tracks are arranged substantially evenly on M tracks (M is an integer of 2 or more)..
[0018]
Thus, even when a burst error occurs in one entire track, data can be restored by error correction.
[0019]
  Claim4The invention described in claim3In the present invention, N, M, L1, and L2 are set so that (N * (L1 + L2)) / M is a natural number.
[0020]
As a result, even when the number of recording blocks L per unit sector is not divisible by the number of recording tracks M, N * L inner code blocks can be distributed evenly over M tracks, so that the entire track can be distributed. Even when a burst error occurs, data can be restored by error correction.
[0021]
  Claim5The invention described in claim 1 provides2In the described invention, N and M are set so as to satisfy N <M.
[0022]
As a result, the number of inner code blocks included in one track is reduced, and the L2 is also relatively reduced, so that the circuit scale required for outer code correction can be reduced.
[0023]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0024]
(Embodiment)
FIG. 1 shows a schematic configuration of the tape drive of the present invention. In FIG. 1, 1 is a SCSI interface means for transmitting and receiving data blocks using a host computer and a SCSI interface, 2 is a compression / decompression means for compressing data blocks transferred from the SCSI interface means 1, and 3 is compression / decompression. Packetizing means 4 for converting the data block transferred from the means 2 into a packet composed of predetermined bytes, 4 generates a subpacket by dividing the packet, and generates an error correction code having a product code structure in units of subpackets. Data converting means for generating and converting them into sync blocks, which are the minimum unit of recording, 5 is a recording means for receiving the sync block as input, performing modulation, etc., and transmitting a signal amplified by a recording amplifier to a magnetic head , 6 is a magnetic tape running control means for controlling the running of the magnetic tape, and 7 is a magnetic head. , 8 magnetic tape, 9 SCSI interface unit 1, the compression means 2, packetization means 3, the data converter 4, a tape interface means 5, a system controller for controlling the tape running means 6.
[0025]
An outline of the processing flow of the tape drive configured in FIG. 1 will be described with reference to FIG.
[0026]
FIG. 2A is a conceptual diagram of a data block transferred from the host, and the data block transferred from the host is denoted as LB (Logical Block). In the figure, LB1 is denoted by LB number 1. Similarly, LB2, LB3,...
[0027]
The plurality of LBs shown in FIG. 2A are input to the compression / decompression means 2. The compression / decompression means 2 combines the LBs into N blocks (N is a natural number) into an LBS (Logical Block Set) and compresses the data in units of LBS (FIG. 2B). As means for data compression, reversible compression such as ALDC is used. Since the operating principle of ALDC is described in Standard ECMA-222 (Adaptive Lossless Data Compression Algorithm), the description is omitted.
[0028]
The LBS compressed by the compression / decompression means 2 is input to the packetizing means 3 and packeted with a predetermined number of words (FIG. 2 (c)). Since the block size of the LBS is not constant due to the difference in the block size from the host and the compression efficiency of the data block, one packet is composed of a part of the LBS and a plurality of LBSs. In order to manage these, a packet information table (PIT) and a packet management table (PMT) are added to each packet.
[0029]
The configuration of one packet will be described in detail with reference to FIG.
[0030]
One packet is composed of 374400 bytes, and as shown in the figure, 36 bytes from the first byte of the packet are allocated to the PIT. The LBS is packed backward from the 37th byte from the top. Each time the LBS is packed, the PMT is packed forward by 4 bytes from the last byte of the packet. When the data size of the LBS is small and many LBSs are included in one packet, the number of bytes allocated to the PMT increases in proportion to the number of LBSs.
[0031]
A data structure of a 4-byte PMT corresponding to the unit LBS is shown in FIG.
The PMT is composed of three elements: PMT-Type, Comp, and Byte-Count. PMT-Type indicates the type of the corresponding LBS, and is identified as follows.
(1) All bytes of the LBS are included in the current packet.
(2) The first byte of the LBS is included in the current packet, but not all bytes.
(3) The first byte and the last byte of the LBS are not included in the current packet.
(4) The last byte of the LBS is included in the current packet, but the first byte is not included.
(5) The LBS is a file mark.
(6) The LBS is a set mark.
(7) Indicates the end of the PMT.
[0032]
Comp is a bit for identifying whether or not the LBS is an LBS subjected to data compression by the compression / decompression means 2.
[0033]
Also, the byte length of each LBS belonging to the current packet is registered as Byte-Count.
[0034]
The PMT-Type indicating the end of the PMT is the last data block transferred from the host computer or the like, and when the data ends in the middle of one packet, this PMT-Type is registered and the packet is terminated.
[0035]
FIG. 5 shows the data structure of PIT. The PIT is composed of the following seven elements.
(1) Packet-No,
(2) Max-File-Mark-Count,
(3) Max-Set-Mark-Count,
(4) Max-Record-Count,
(5) Packet-No Of Previous File-Mark,
(6) Packet-No Of Previous Set-Mark,
(7) Packet-No Of Previous Record
(1) indicates the serial number of the packet from the beginning of the tape. (2) is the number of file marks from the beginning of the tape, (3) is the number of set marks from the beginning of the tape, and (4) is the number of data blocks from the beginning of the tape. (5) is the packet number including the previous file mark, (6) is the packet number including the previous set mark, and (7) is the packet number including the previous data block.
[0036]
In this embodiment, an EOD packet is provided as identification data indicating the end of a data block. The EOD packet and the normal data packet described above are identified by the PIT of the packet. The data structure of the EOD packet is shown in FIG. As shown in the figure, the EOD packet stores 24′hFFFFFF in the Packet-No part of (1). In the present embodiment, the number of packets that can be recorded on the magnetic tape cannot reach 24'hFFFFFF, and the Packet-No 24'hFFFFFF can be used as an identification code for determining an EOD packet.
[0037]
The packetized packet by the packetizing means 3 is input to the data converting means 4 and divided into 15 pieces.
[0038]
FIG. 7 shows a conceptual diagram for dividing a packet. As shown in the figure, a packet having 374400 bytes of data is divided every 24960 bytes, resulting in 15 subpackets. In this embodiment, 15 subpackets are created in order from the beginning of the packet (FIG. 2 (d)).
[0039]
Each of the 15 subpackets constitutes an error correction code. In the error correction code, a Reed-Solomon code that allows a large distance between codes is generally used. FIG. 8 is a block diagram of the error correction code. In each subpacket, 12 bytes of error correction parity are added to 160 bytes in the vertical axis direction of FIG.
[0040]
In general, the 160-byte data and 12-byte error correction parity in the vertical axis direction in the figure are called outer codes, and the 12-byte error correction parity is called outer code parity. Outer codes are created with 156 codes per unit subpacket.
[0041]
After the outer packet is generated as described above, the subpacket is divided into 172 blocks having 156 words in the horizontal direction in FIG.
[0042]
Further, the block is obtained by adding 16 bytes of error correction parity to 156 bytes of data. An error correction parity generation method is described in detail in ECMA292, and will not be described.
[0043]
A 172-byte error correction code obtained by combining the 156-byte data and the 16-byte error correction parity is referred to as an inner code, and the error correction parity is referred to as an inner code parity. As with the outer code, the inner code uses a Reed-Solomon code.
[0044]
In the configuration of the error correction code shown in FIG. 8, an outer code is generated by adding an outer code parity to the subpacket data in the vertical axis direction of FIG. 8, and 156 bytes of data in the horizontal axis direction of FIG. An inner code is generated by adding an inner code parity. Since these outer code and inner code are orthogonal to each other, they are generally called product codes.
[0045]
As described above, the subpacket is a sector that has been subjected to an error correction product code using an outer code and an inner code. One inner code (unit inner code) of 172 bytes on the horizontal axis in FIG. 8 is called an inner code block. Each sector has 172 inner code blocks. Of the 172 inner code blocks, 160 are inner code blocks composed of data, and 12 are inner code blocks composed of an outer code parity part.
[0046]
The 160 inner code blocks per sector composed of data are further divided into 20 blocks, and 8 blocks are mapped to one recording track (FIG. 2 (e)). Since eight inner code blocks per sector are mapped to one recording track, the number of inner code blocks composed of data included in one track is 8 * 15 = 120 blocks.
[0047]
Further, the inner code block is added with a 3-byte identification code (SYNCID) and a 2-byte synchronization code (SYNC) to become a sync block. A sync block is a basic unit constituting a recording track.
FIG. 9 shows a schematic diagram of the sync block.
[0048]
The SYNC ID is a sync block identification code.
(1) Serial number from the track start end of each recording block in each track (sync number)
(2) Recording track serial number (track number) in a unit packet
(3) Identification number of recording contents (data / EOD / Amble) of each recording block
Is stored.
[0049]
These inner code blocks are interleaved in sector units and mapped to tracks. That is, if the data of the first sync block of the track is the inner code block belonging to sector 1, the data of the next sync block is the inner code block belonging to sector 2, and the data of the next sync block is the inner code block belonging to sector 3. Blocks are interleaved like this.
In formula,
The track number is A (0 ≦ A ≦ 19; A is an integer),
The data of the sync block having the sync block number B (0 ≦ B ≦ 159; B is an integer)) is
・ Internal code block number of sector: 20 * INT (B / 15) + A
-Sector number: (2 * B + 4 * A)% 15
It matches the data of the inner code block.
[0050]
INT (): integer operator, ()% 15: remainder of dividing () by 15
FIG. 10 shows the correspondence between each sector and track. The vertical axis indicates the sync block number of the track, and the horizontal axis indicates the track number in the unit packet. Also, the internal codes enclosed by rectangles are shown in the order of the inner code block number and the sector number in the sector to which the inner code block constituting each sync block belongs.
[0051]
For example, the second sync block from the top of the leftmost track in the figure has a code 0-2, which indicates that it is the data of the 0th inner code block of the second sector. .
[0052]
When sync block number ≧ 160, an inner code block including outer code parity is mapped to each sync block. A method for mapping the inner code block composed of the outer code parity will be described below.
[0053]
The inner code block composed of the outer code parity is 12 blocks per sector, and 12/20 is not an integer. Therefore, the inner code block composed of the outer code parity per sector cannot be equally divided into 20 tracks.
[0054]
Therefore, in this embodiment, 9 blocks of inner code blocks each consisting of 12 * 15 = 180 outer code parities per packet are distributed over 20 tracks. Further, in order to evenly arrange the recording blocks of the unit sectors on the tracks as much as possible, the number of inner code blocks made up of outer code parity per sector included in one track is set to one or two.
[0055]
FIG. 10 shows the mapping of the inner code block of the outer code parity part to each sync block.
[0056]
As shown in the figure, the inner code blocks including the outer code parity arranged in the sync blocks having the sync block numbers of 120 or more are also arranged substantially equally.
[0057]
As described above, 129 sync blocks are generated per track, and 172 * 15 inner code blocks per packet are recorded by the recording means 5 on 20 tracks.
[0058]
Next, an error correction operation in the case where all of the 20 tracks constituting one packet are wrong in the reproduction of the data recorded on the magnetic tape 8 according to the present embodiment will be described. . Since the entire configuration and data flow are the reverse of those shown in FIGS. 1 and 2, the description thereof will be omitted, and the configuration and operation during reproduction of the data conversion means 4 shown in FIG.
[0059]
Data recorded on the magnetic tape 8 is reproduced by the reproducing head 10 and input to the reproducing means 11. The reproduction means detects reproduction data reproduced from the reproduction head 10 by a reproduction amplifier, an equalizer, and the like. The detection data output from the reproduction means 11 is input to the inner code correction means 12 to detect the sync code of the sync block shown in FIG. 9 of the sync block and the sync block identification code. Thereby, the track number of the detected sync block and the sync block number in the track can be identified. Further, error correction is performed on the inner code block belonging to the sync block. In this embodiment, 16 bytes of inner code parity are added, so that up to 8 bytes of data can be corrected for each inner code block. It is known that errors up to (generally error correction parity number / 2) are corrected by Reed-Solomon codes. When there are 9 or more errors, an error flag is output for each inner code block.
[0060]
The inner code block subjected to inner code correction and error detection by the inner code correction means 12 is input to the packet storage means 13. The packet storage means 13 has an internal code block for 15 sectors per packet and a memory capable of storing the error correction possibility of the internal code block, and the inverse mapping of FIG. 11 is performed for each internal code block. For example, the inner code block belonging to the sync block of track number 0 and sync block number 1 is stored in the memory location of sector number 1 and inner code block number 0.
[0061]
If all the sync blocks with track number 0 are erroneous, all the error corrections of the inner code block corresponding to the sync block with track number 0 are all denied and recorded in the memory.
[0062]
The inner code blocks that are arranged and recorded in the sync blocks of track numbers 1 to 19 are determined to be error-correctable by the inner code correction means 12, and are mapped in the opposite directions to those in FIG. Stored.
[0063]
The inner code block and error flag stored in the packet storage means 13 are subjected to outer code correction by the outer code correction means 14 for each sector. Outer code correction is performed using the error flag in the direction of the vertical axis in FIG.
[0064]
The error flag indicates the byte position of the error in the outer code, and it is known that if the error position can be determined, an error corresponding to the number of outer code parities can be corrected.
[0065]
Since the outer code of the present embodiment has a 12-byte outer code parity, errors up to 12 bytes can be corrected.
[0066]
Further, in this embodiment, even if all the tracks 0 are wrong, the error flag of the inner code block of each sector is 9 blocks at most, so that error correction is possible by outer code correction, and one track is damaged. Even in this case, the data can be restored.
[0067]
As described above, according to the present embodiment, data transferred from the host computer is packed in units of 374400 bytes to form a packet, and the packet is divided into 15 pieces in order to form a 24960-byte sub-packet. A product code is applied to the subpacket to generate a sector composed of 172 inner code blocks, and the inner code blocks constituting each sector are interleaved approximately evenly over 20 tracks.
[0068]
With the above configuration, 15 * 172 inner code blocks constituting 15 sectors having a product code configuration can be distributed almost evenly over 20 tracks. The error of the code block) is an error of 8 to 9 inner code blocks per product code, and the error correction capability of 15 product codes can be used for one track. An information recording device that is resistant to burst errors can be provided.
[0069]
Further, in this embodiment, there are 172 inner code blocks constituting one sector, and 172/20 is not an integer. Therefore, these cannot be equally distributed over 20 tracks, but sectors constituting one packet. The number is set to 15 so that (15 * 172/20) is an integer, and the number of inner code blocks constituting each sector included in one track is 9 or less (172/20 + 1 = 9.6) is interleaved.
[0070]
By setting each parameter as described above, even when the number of recording tracks is larger than the number of sectors and the number of recording tracks is not divisible by the number of sectors, the inner code blocks constituting each sector are changed. It is possible to provide an information recording apparatus that can interleave almost 20 tracks and is resistant to burst errors.
[0071]
Further, according to the present embodiment, the data transferred from the host computer is packed in units of 374400 bytes to form a packet, the packet is divided into 15 pieces in order to form a 24960 byte sub-packet, A sector in which a product code is applied to the subpacket is generated, and the outer code configuration of each sector is a configuration in which a 12-byte outer code parity is added to 160-byte data. Therefore, each sector is composed of a first inner code block having a data part of 160 outer codes as a constituent element and a second inner code block having twelve outer code parity parts as a constituent element. In this example, at most 9 (<12) inner code blocks are arranged substantially evenly per unit sector.
[0072]
With the above configuration, 15 * 172 inner code blocks constituting 15 sectors having a product code configuration can be distributed almost evenly over 20 tracks. (Code block error) results in 8 to 9 inner code block errors per product code. On the other hand, since the product code configuration of each sector can correct up to 12 errors in the inner code block, error correction is possible even when one track is completely erroneous, and information that is more resistant to burst errors. A recording device can be provided.
[0073]
【The invention's effect】
As described above, according to the present invention, (N * L) inner code blocks of N sectors having a product code configuration can be distributed substantially evenly to M tracks. The error correction capability of N product codes distributed over M tracks can be used, and a format having strong error correction capability is provided for burst errors that occur in tracks with relatively small redundancy. Is possible.
[0074]
Further, by setting N and M to N <M, the number of sync blocks constituting one track can be set relatively small, and one track correction can be performed with a relatively small number of outer code parities. The scale can also be reduced.
[Brief description of the drawings]
FIG. 1 is a block diagram of a tape drive device according to an embodiment of the present invention.
FIG. 2 is a data flow diagram of the tape drive device according to the embodiment of the present invention.
FIG. 3 is a structural diagram of a packet according to an embodiment of the present invention.
FIG. 4 is a structural diagram of a PMT described in an embodiment of the present invention.
FIG. 5 is a structural diagram of the PIT described in the embodiment of the present invention.
FIG. 6 is a structural diagram of a PIT of an EOD packet according to an embodiment of the present invention.
FIG. 7 is a conceptual diagram of a subpacket described in the embodiment of the present invention.
FIG. 8 is a configuration diagram of a sector described in the embodiment of the present invention;
FIG. 9 is a configuration diagram of a sync block according to the embodiment of the present invention.
FIG. 10 is a layout diagram of inner code blocks according to an embodiment of the present invention on a track.
FIG. 11 is an internal configuration diagram of data conversion means at the time of reproduction described in the embodiment of the present invention;
FIG. 12 is a layout diagram of inner code blocks on a track in a conventional information recording apparatus.
[Explanation of symbols]
1 SCSI interface means
2 Compression / decompression means
3 Packetization means
4 Data conversion means
12 Inner code correction means
13 Packet storage means
14 Outer code correction means

Claims (5)

An information recording apparatus for recording data on a track on a magnetic tape by a head, comprising a packet in which data is packed in predetermined units, and dividing the packet into N pieces (N is 2 or more) generating a sector integer), wherein each sector have a inner code blocks of L (L is an integer of 2 or more), the M pieces (M L number of inner code blocks of each sector is an integer of 2 or more ) substantially and uniformly distributing constituting the track, when L / M is not a natural number also, information and sets the (N * L) / M so that a natural number, N, M, L Recording device.   2. An information recording apparatus according to claim 1, wherein each track has (L / M + 1) or less inner code blocks constituting each sector. An information recording apparatus for recording data on a track on a magnetic tape by a head, comprising a packet in which data is packed in a predetermined unit, and recording the packet by dividing it into M tracks (M is an integer of 2 or more). , N packets (N is an integer of 2 or more) are generated by dividing the packet and product code is applied, and each sector is L1 (L1 is an integer of 2 or more) having the data part of the outer code as a component. ) First inner code block and L2 second inner code blocks (L2 is an integer of 2 or more) having the outer code parity part as components, and the first and second of each sector An information recording apparatus characterized in that the inner code blocks are arranged substantially evenly on M tracks (M is an integer of 2 or more) under the condition of L2 or less per track. 4. The information recording apparatus according to claim 3 , wherein N, M, L1, and L2 are set so that (N * (L1 + L2)) / M is a natural number. 3. The information recording apparatus according to claim 1, wherein N and M are set so as to satisfy N <M.

Images