JP2010152603A - Device and method for migrating data recorded in recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve performance under the consideration of access frequency in data migration. <P>SOLUTION: In a controller 16 of a tape drive, if a command processing part 61 receives a data read instruction, a channel input/output part 63 reads data from a tape, Ram-count showing the number of times that the tape is mounted for reading is stored in each data on the tape, and Dm-readfreq showing the history of the number of data reads is stored in the read data. In data migration, a CM input/output part 64 reads the Ram-count and Dm-readfreq, a frequency calculation part 65 calculates the ratio of the number of times shown by Dm-readfreq to the Ram-count as access frequency, and a migration control part 66 controls a buffer management part 62 or the like so that the data read from the migration source tape can be recorded in the order of access frequency into the migration destination tape. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an apparatus and method for transferring data recorded on a recording medium. In particular, the present invention relates to an apparatus and method for supporting migration of data recorded on a recording medium to the recording medium or another recording medium.

The data recorded on the medium may be transferred to another new medium depending on the life of the medium or the medium drive device. Further, the transition to another medium may be performed for operational reasons, for example, switching to a medium with higher cost performance.
At the time of such data migration, it is desirable to record data having more accesses at a position advantageous in performance. For example, when accessing multiple data as one data group like a data object, the data recorded at the most disadvantageous position in performance may affect the performance of the entire data group. is there.

Conventionally, when data recorded on a medium is migrated to itself or another medium, a technique for improving performance by considering the frequency of access to the data has been known (see, for example, Patent Documents 1 to 3). .
In Patent Document 1, a logical volume with a high access frequency is assigned to a RAID group in a long-time energization mode, a logical volume with a medium access frequency is assigned to a RAID group in a first short-time energization mode, and a logical volume with a low access frequency is set to a second. A rearrangement plan is created to be placed in each RAID group in the short-time energization mode, and the execution time of the rearrangement plan is determined based on the access frequency to the RAID group.

In patent document 2, the access frequency of the file unit comprised by a some block is memorize | stored, this access frequency is more than a predetermined value, and several files with correlation are extracted as one group, and each of these extracted each The data of each block of the file is read from the optical disc and sequentially written into successive blocks on the optical disc or neighboring blocks.
In Patent Document 3, a memory is provided in a cartridge of a tape cassette, information on the access frequency for each partition is recorded in the memory, and when copying a tape, the access frequency information is read, Data is recorded from the top of the tape in descending order.

JP 2007-164650 A JP-A-8-263335 JP-A-11-31376

  However, the access frequency in the cited documents 1 to 3 is recorded for a data group (logical volume, file, partition) including a plurality of data. The data group may contain data that has existed for a long time and data that has been created recently. If the access frequency is obtained with reference to the time axis for such data with different creation times, the true access frequency cannot be obtained. Therefore, the techniques of the cited documents 1 to 3 have a problem that the performance cannot be improved in consideration of the true access frequency when data is migrated.

  An object of the present invention is to improve performance in consideration of a true access frequency without using a time axis when migrating data.

  For this purpose, the present invention is an apparatus for supporting the transfer of data recorded on the first recording medium to the second recording medium, and for each data recorded on the first recording medium, First information indicating the number of times any data is read from the first recording medium after each data is written to the first recording medium, and each data is read from the first recording medium The acquisition unit that acquires the second information indicating the number of times, the grasping unit that grasps the access frequency of each data using the first information and the second information obtained by the obtaining unit, and the grasping unit There is provided an apparatus including a determination unit that determines a recording position of each data in a second recording medium based on the grasped access frequency.

Here, when specific data is read from the first recording medium, the apparatus updates the first information for each data and updates the second information for the specific data. May further be included. In this case, the specific data includes main data and metadata that is additional information of the main data, and the update unit reads at least one of the main data and the metadata from the first recording medium. In this case, it may be determined that specific data has been read.
The first information is stored in the first recording medium after each data is written on the first recording medium according to the number of times the first recording medium is mounted on the storage device after each data is written on the first recording medium. It may indicate the number of times any data is read from one recording medium.
Further, the grasping unit may grasp a ratio of the number of times indicated by the second information to the number of times indicated by the first information as an access frequency, or the number of times indicated by the first information is within a predetermined range. The second information may be used to grasp the access frequency for the data in the data.
The second information is information representing a history of reading each data with a plurality of bit values, and the number of times each data is read out of M times of data reading from the first recording medium. , It may be represented by N (N <M) bit values among a plurality of bit values.
Further, the determining unit records data having a high access frequency at the first recording position on the second recording medium, and data having a low access frequency is stored on the second recording medium than the first recording position on the data. The recording position of each data on the second recording medium may be determined so as to be recorded at the second recording position that takes time to read. In this case, the second recording medium is a tape medium from which data reading is started from the first end toward the second end, and the second recording position is higher than the first recording position. The second recording medium may be a position close to the second end, and the second recording medium starts reading data from the first end toward the second end, and is long. A tape medium having a plurality of areas formed by dividing at least one position in the direction, wherein the first recording position is a position in the first area of the plurality of areas, The recording position may be a position in a second area that is closer to the second end than the first area among the plurality of areas.

  In addition, the present invention is an apparatus that supports rearrangement of data on a recording medium, and for each data recorded on the recording medium, any data is read from the recording medium after each data is written on the recording medium. An acquisition unit that acquires first information indicating the number of times of output and second information indicating the number of times each piece of data is read from the recording medium; first information acquired by the acquisition unit; and second information And a determination unit that determines the recording position of each data on the recording medium based on the access frequency that is grasped by the grasping unit. provide.

  The present invention is also an apparatus for transferring data recorded on a first tape medium to a second tape medium, recorded on the first tape medium, and showing main data and accompanying information of the main data. For each data including metadata, first information indicating the number of times the first tape medium is mounted on the tape drive for reading any data after each data is written to the first tape medium. And an acquisition unit that acquires second information indicating the number of times each data has been read from the first tape medium, and second information for the number of times indicated by the first information acquired by the acquisition unit An apparatus is also provided that includes a grasping unit that grasps the ratio of the number of times as an access frequency of each data, and a writing unit that writes data to the second tape medium in order from the data with the highest access frequency grasped by the grasping unit.

  Furthermore, the present invention is a method for supporting the transfer of data recorded on a first recording medium to a second recording medium, wherein each data recorded on the first recording medium is recorded in the first recording medium. First information indicating the number of times any data is read from the first recording medium after each data is written on the medium, and a first information indicating the number of times each data is read from the first recording medium. 2, the step of determining the access frequency of each data using the acquired first information and the second information, and the second based on the acquired access frequency Determining the recording position of each data on the recording medium.

  Furthermore, the present invention is a program for causing a computer to function as an apparatus for supporting the transfer of data recorded on the first recording medium to the second recording medium, and recording the computer on the first recording medium. For each of the recorded data, first information indicating the number of times any data is read from the first recording medium after each data is written to the first recording medium, and each of the data from the first recording medium. Grasping the access frequency of each data using the acquisition unit that acquires the second information indicating the number of times the data has been read, and the first information and the second information acquired by the acquisition unit And a program that functions as a determination unit that determines the recording position of each data on the second recording medium based on the access frequency grasped by the grasping unit.

  According to the present invention, when data is migrated, performance can be improved in consideration of the true access frequency without using a time axis.

  The best mode for carrying out the present invention (hereinafter referred to as “embodiment”) will be described in detail below with reference to the accompanying drawings. Although the present invention can be applied to any recording medium, a tape medium (hereinafter simply referred to as “tape”) will be described as an example here. In the current situation where it is difficult to store all of the ever-increasing amounts of data in high-speed storage devices, it is important to place data in consideration of access frequency for data stored in low-speed storage devices such as tape drives. Because it is.

First, a digital data archive device (hereinafter simply referred to as “archive device”) to which this embodiment is applied will be described.
FIG. 1 is a diagram illustrating a configuration example of an archive device 100 to which the exemplary embodiment is applied.
As illustrated, the archive device 100 includes a tape drive 10, a control mechanism 30, an accessor 40, and a cartridge slot 50.
Among these, the tape drive 10 includes a host interface (hereinafter referred to as “host I / F”) 11, a buffer 12, a channel 13, a write head 14 a, a read head 14 b, and a motor 15. Further, the controller 16, a head position control system 17, and a motor driver 18 are included. Furthermore, since a tape cartridge (hereinafter simply referred to as “cartridge”) 20 can be loaded into the tape drive 10, the cartridge 20 is also illustrated here. The cartridge 20 includes a tape 23 wound around reels 21 and 22. As the reels 21 and 22 rotate, the tape 23 moves in the longitudinal direction from the reel 21 to the reel 22 or from the reel 22 to the reel 21. In addition, although the magnetic tape is illustrated as the tape 23, tape media other than a magnetic tape may be sufficient.

  The cartridge 20 also includes a cartridge memory 24. The cartridge memory 24 records, for example, information on how data is written on the tape 23. Then, for example, by using an RF interface to check the index of data written on the tape 23 in a non-contact manner, the usage status of the tape 23, and the like, high-speed access to the data is possible. In the figure, an interface for accessing the cartridge memory 24 such as this RF interface is shown as a cartridge memory interface (hereinafter referred to as “CMI / F”) 19.

  Here, the host I / F 11 communicates with the host 200 via the control mechanism 30. For example, the host 200 receives a command for instructing data writing to the tape 23, a command for moving the tape 23 to a target position, and a command for instructing reading of data from the tape 23. Note that SCSI is exemplified as a communication standard used in the host I / F 11. In the case of SCSI, the first command corresponds to the Write command, the second command corresponds to the Locate command and the Space command, and the third command corresponds to the Read command. Further, the host I / F 11 returns a response to the host 200 as to whether the processing according to these commands has succeeded or failed.

The buffer 12 is a memory that stores data to be written to the tape 23 and data read from the tape 23. For example, it is configured by a DRAM (Dynamic Random Access Memory). The buffer 12 is composed of a plurality of buffer segments, and each buffer segment stores a data set that is a unit of reading and writing with respect to the tape 23.
The channel 13 is a communication path used for sending data to be written on the tape 23 to the write head 14a and receiving data read from the tape 23 from the read head 14b.
The write head 14a writes information to the tape 23 when the tape 23 moves in the longitudinal direction, and the read head 14b reads information from the tape 23 when the tape 23 moves in the longitudinal direction.
The motor 15 rotates the reels 21 and 22. In the figure, the motor 15 is represented by one rectangle, but it is preferable to provide two motors 15, one for each of the reels 21 and 22.

The controller 16 controls the entire tape drive 10. For example, data writing to the tape 23 and reading from the tape 23 are controlled in accordance with a command received by the host I / F 11. The head position control system 17 and the motor driver 18 are also controlled.
The head position control system 17 is a system that controls the write head 14a and the read head 14b to track one or more desired wraps. Here, the lap is a group of a plurality of tracks on the tape 23. When it is necessary to switch the wrap, it is also necessary to electrically switch the write head 14a and the read head 14b. Therefore, the head position control system 17 performs such switching control.
The motor driver 18 drives the motor 15. As described above, if two motors 15 are used, two motor drivers 18 are also provided.

The control mechanism 30 is a mechanism that controls the accessor 40 and the tape drive 10 in accordance with an instruction from the host 200. That is, the accessor 40 is instructed to load the cartridge 20 into the tape drive 10 so that the data instructed by the host 200 can be read and written, and the host 200 is instructed to the tape drive 10. The instructed data is read from the cartridge 20 loaded by the accessor 40 or written to the cartridge 20.
The accessor 40 takes out the cartridge 20 from the cartridge slot 50 and loads it into the tape drive 10 under the control of the control mechanism 30.
The cartridge slot 50 is a place for storing a cartridge 20 that has not been read or written. Here, the cartridge slot 50 is shown as a single rectangle, but in reality it is a plurality of slots for storing a plurality of cartridges.

  Although only one tape drive 10 is shown in the figure, a plurality of tape drives 10 may be provided. In that case, the control mechanism 30 instructs the loading of the cartridge 20 into the tape drive 10 by transmitting the identification information of the tape drive 10 to which the read command or the write command is sent to the accessor 40. Further, in the figure, the tape drive 10 capable of loading only one cartridge 20 is shown, but the tape drive 10 capable of loading a plurality of cartridges 20 may be used.

  By the way, the access performance to the data recorded on the tape 23 by the tape drive 10 varies greatly depending on the recording position on the tape 23. Accordingly, when data is transferred to the tape 23, if the access frequency to the data is not taken into consideration, the access speed to the data with high access frequency is slowed, and the access performance to the data is deteriorated.

  Further, in the current situation where the tape drive 10 is frequently used as an archive of photo data, video data, etc., not only the data (main data) itself but also metadata as its additional information is stored for a long time. It is becoming important as data to be used. Here, as the metadata, for example, for photo data, the photo shooting location, date / time, subject, photographer, photographing device, recording format, file name, file size, and the like can be considered. As described above, since the metadata is often strongly related to the data and is often handled together, if the data is stored on the tape 23 as an archive, if the metadata is also stored on the same tape 23, the data and If metadata is not properly placed, there is a problem that access performance is degraded.

Therefore, in the present embodiment, the access frequency of mutually related data recorded on the tape 23 is recorded, and the data is transferred using the access frequency. As a result, an appropriate arrangement of data and metadata groups in data migration is realized.
Specifically, the following configuration is adopted.
That is, data and metadata are associated with each other by dm-id, and each time access to data or metadata occurs, an access history is stored in an access counter variable group (hereinafter simply referred to as “counter”) corresponding to the dm-id. Record. In this case, the recording of the access history to the counter group is performed not on the time axis but on the mount count of the tape 23 for data reading, in consideration of the characteristics of the sequential access recording medium called tape. . The counter group records in the semiconductor memory attached to the tape 23 in consideration of the portability of the tape 23.

  After the access history is recorded in this way, when data is transferred from one recording medium A to another recording medium B, first, the access frequency to the data and metadata associated with the dm-id in the recording medium A is determined. The calculation is performed according to a policy based on the counter group. Then, from the viewpoint of access performance, they are rearranged on the recording medium B in the order of high access frequency. Here, as a rearrangement method, a method of sequentially writing the associated data and metadata as one unit from the top of the tape 23 is adopted. This improves the access performance when data and metadata are read sequentially from the beginning.

Hereinafter, a configuration for realizing such an operation will be described in detail.
First, the association between data and metadata will be described.
FIG. 2 is a diagram schematically showing how data and metadata are associated with each other.
As shown in (a), first, the host 200 creates an association between data and metadata by assigning an arbitrary identification number dm-id. The data is passed to the control mechanism 30 via the data port, and the metadata is passed to the control mechanism 30 via the metadata port.

  Next, in the control mechanism 30, information passed through the data port is handled as data, and the ID generation unit 31 generates d-id and assigns it to the data. Here, d-id is a data identification number and is a unique value at least in an environment in which the archive apparatus 100 operates. Also, information passed through the metadata port is handled as metadata, and the ID generation unit 31 generates m-id and assigns it to the metadata. Here, m-id is an identification number of metadata, and is a unique value at least in the environment in which the archive apparatus 100 operates. Even if the data and the metadata are passed apart in time, the relationship between them can be established in the archive device 100 by dm-id.

As a result, the control mechanism 30 records the relationship between the data based on the dm-id and the metadata in the database (DB) 32 as a dm-a table (hereinafter simply referred to as “dm-a”). (B) shows the format of dm-a. Here, one piece of metadata is associated with one piece of data, but the data and metadata can also have a relationship of M to N (M and N are natural numbers).
In addition, the control mechanism 30 records the recording location of the data and metadata in the archive device 100 as a dm-l table (hereinafter simply referred to as “dm-l”) in the DB 32 for each piece of data and metadata. . (C) shows the format of dm-l. Here, when a plurality of storage spaces are mounted in the archive device 100 and there are a plurality of copies of data or metadata in those spaces, entries corresponding to the number of copies are registered in dm-l. . In the figure, such entries are represented by Location-1, Location-2,.

Here, the association between data and metadata will be described using a specific example.
FIG. 3 is a diagram showing a specific example of association between data and metadata.
(A) is recorded on the recording medium with the medium name “A001”, with the data identified by d-id and the metadata identified by m-id having a relation identified by dm-id. It represents the state. Specifically, data with d-id d3 and metadata with m-id m4, m6, and m8 are recorded with an association with dm-id dm03, and data with d-id d1 and m-id The metadata with id m1 and m7 is recorded with the relationship dm-id dm01, the data with d-id d2 and the metadata with m-id m5 have the relationship dm-id dm02 The d-id d5 data and the m-id m9 metadata are recorded with the relationship dm-id dm04. In this specification, a combination of data and metadata associated with each other by dm-id is called “DM”.
Further, (b) shows a state in which DM identifiers (d-id, m-id) whose dm-id is dm03 are listed in dm-a.

However, since the data and metadata are written to the tape 23 when the preparation is completed, the recording order of the data and metadata on the tape 23 is not uniquely determined.
Therefore, even if the relationship between data and metadata is logically constructed by dm-id, the physical recording positions on the tape 23 can be dispersed when they are recorded on the tape 23. There is sex.

FIG. 4 is a diagram showing an example of data and metadata on the tape 23 distributed and recorded in this way. As illustrated, a plurality of wraps exist on the tape 23. Further, in the figure, the white arrow indicates the data writing direction, and the solid arrow indicates that the running direction of the tape 23 is reversed.
Here, first, in the forward direction with respect to wrap # 0, metadata with m-id m6, data with d-id d3, metadata with m-id m1, data with d-id d1, m -Metadata with id of m4 is written. Next, in the reverse direction to wrap # 1, metadata with m-id of m5 and data with d-id of d2 are written. Then, in the forward direction again for wrap # 2, metadata with m-id m7, metadata with m-id m9, data with d-id d5, and metadata with m-id m8 are written. ing.

  That is, even if data and metadata groups as shown in FIG. 3A are stored in a certain recording medium, if they are transferred to the tape 23, they are arranged regardless of the association between the data and metadata. there is a possibility. In this case, since the DM of a certain dm-id cannot be read unless the tape 23 is aligned a plurality of times, the access performance is lowered by the amount corresponding to the alignment of the tape 23.

Therefore, in the present embodiment, when data and metadata recorded in this way are migrated, they are rearranged based on the access frequency of the data and metadata.
Here, the following three counters are defined for each DM.

(1) Read Access Mount counter (Ram-count)
This Ram-count is a counter defined for each dm-id. The number of mounts of the tape 23 for reading after the time when the data and metadata associated with the dm-id are written on the tape 23 is held. In the present embodiment, it is assumed that the tape 23 on which the data is recorded is mounted on the tape drive 10 every time it is necessary to read the data. Hereinafter, the mounting of the tape 23 for reading will be referred to as “reading mount”.

  When the Ram-count is defined in this way, the dm-id having the largest Ram-count becomes the dm-id associated with the data and metadata written on the tape 23 the oldest. In addition, by referring to the Ram-count assigned to each dm-id, it is possible to know which data and metadata is relatively old or new and the probability of read access to that data and metadata. it can. This counter should be defined so that it can hold a large value so as not to wrap. If it wraps, it is reset to zero.

(2) Read Frequency counter (Dm-readfreq)
This Dm-readfreq is also a counter defined for each dm-id. A history of how much data and metadata associated with this dm-id have been read while the number of read mounts indicated by Ram-count has occurred is maintained. This counter is updated when either data or metadata is read.

Here, for example, it is assumed that the maximum value of Ram-count is 16. That is, only the past 16 read mount histories are counted. In this case, assuming that each dm-id data and metadata is read by each bit of Dm-readfreq, Dm-readfreq is represented by 16 bits.
However, if 1 bit is assigned to one read mount, for example, when it is desired to measure the frequency from the history of the 256 previous read mounts, a memory area of 32 bytes (= 256 bits) is required. End up.

Therefore, a method of expressing a plurality of read mounts with 1 bit is considered.
For example, if 8 read mounts are expressed by 1 bit, a memory area of 4 bytes (= 256 bits / 8 = 32 bits) is sufficient. When such a method is adopted, a Dm access counter (described later) that holds the number of times of reading data of dm-id among the past eight read mounts is required in addition to Dm-readfreq.
However, with this method, the counter becomes 1 if the dm-id data is read even once in the past eight read mounts. Therefore, it is not clear whether the number of times of reading dm-id data in the eight read mounts is eight or one.

Therefore, a method of expressing a plurality of read mounts with a plurality of bits is considered.
For example, if 8 read mounts are expressed by 2 bits, a memory area of 8 bytes (= 256 bits / 8 × 2 = 32 bits × 2) is required. In this case as well, a Dm access counter (described later) that holds the number of times of reading data of dm-id out of the past eight read mounts is required in addition to Dm-readfreq.
In this method, for example, the following frequencies can be expressed using 2 bits.
“00”: No access “01”: 1 to 2 times “10”: 3 to 5 times “11”: 6 to 8 times This definition can be changed according to the frequency calculation policy.

(3) Dm access counter (Dm-count)
In the above Dm-readfreq, this Dm-count is represented by N (<M) bits with M read mounts as a unit, and the DM of this dm-id is read out of the M read mounts. It is a counter that holds the number of times that has been performed.
For example, if the past eight read mounts are taken as one unit, Dm-count becomes 8 if the data of this dm-id is read by all eight read mounts. If read twice, Dm-count becomes 2.
Note that if there is a read mount in one unit of the read mount, for example, eight times on the tape 23, Dm-count is reset to zero. Here, in order to count one unit of 8 times, the above Ram-count is used. That is, if there are read mounts whose number is a multiple of 8, Dm-count is reset.

By using these three counters, the following information can be obtained.
Of the read mounts to the tape 23 of the past X times, the frequency with which the DM of the target dm-id is read. The number of times the DM of the target dm-id has been read. Access probability calculated based on the number of times the DM of the target dm-id is read out of the read mounts. Access tendency in the X read mounts, that is, whether there are many recent accesses or many accesses in the past. Or

  In general, when talking about the frequency of access to data, a time axis is considered. However, in the case of a removable recording medium such as a tape, if it is ejected from the storage device, or if there is no access to the tape in the storage device, it will be in a state where “time has stopped”. Therefore, in this embodiment, instead of taking the time axis into consideration, the example is that the tape is mounted for the purpose of reading.

The elements that determine the size of these three counters are as follows.
-How many past read mounts are targeted-How many read mounts are represented by how many bits For example, as shown above, when the past 256 read mounts are targeted for measuring access frequency, Ram-count is 1 byte.
Also, when 8 read mounts are taken as one unit and expressed by 2 bits, Dm-readfreq is 8 bytes and Dm-count is 4 bits.

FIG. 5 shows three counters in this case.
Here, since three counters are provided for each dm-id, dm-id is also shown.
However, since the number of bits assigned to dm-id is not defined above, it is simply represented by one rectangle.
On the other hand, since the number of bits to be allocated to the three counters is defined, one bit is represented in association with one elongated rectangle. Specifically, since Ram-count is 8 bits, it is represented by 8 rectangles. Since Dm-count is 4 bits, 8 rectangles are shown, and 4 of them are shaded to indicate that they are unused. Since Dm-readfreq is 64 bits, it is represented by 64 rectangles. In particular, in this Dm-readfreq, since eight read mounts are represented by 2 bits as a unit, the meaning of the rightmost 2 bits is also shown as an example.

The values of these three counters are recorded in the dm access counter table for each dm-id.
FIG. 6 shows an example of the dm access counter table.
In the figure, in the Ram-count column, there have been 215 read mounts since the DM with dm-id “dm01” was written, and 24 after the DM with dm-id “dm02” was written. Dm-id “dm03” DM was written 13 times read mount was written, dm-id “dm04” DM was written 156 times It is indicated that there was a read mount and that there were 79 read mounts since the DM with dm-id “dm05” was written.

  Also, in the Dm-count column, the DM with the dm-id “dm01” and the DM with the dm-id “dm02” have not been read once out of the past eight read mounts. The DM with id “dm03” has been read 4 times out of the past 8 read mounts, the DM with dm-id “dm04” has been read out once in the past 8 read mounts, It is shown that the DM with the dm-id “dm05” has been read 7 times out of the 8 read mounts in the past. These numbers are reflected in the two rightmost bits of Dm-readfreq.

  In addition, the Dm-readfreq column indicates DM read with 8 bits. Since the DM read count in the eight read mounts is represented by 2 bits, the figure shows the DM read count in the 25 to 32 read mounts. Note that “....” On the left side of the bit string of Dm-readfreq indicates that the bit value is omitted. For dm-id “dm02”, Ram-count is 24, so 6 bits is sufficient for Dm-readfreq. For dm-id “dm03”, Ram-count is 13, so Dm-readfreq is 4. Bits are sufficient, but in this case as well, the left side is filled with 0 bit strings and is represented by 8 bits.

Next, the update of these three counters will be described with a specific example.
FIG. 7 is a diagram showing how each counter stored in the cartridge 20 is updated when the tape drive 10 receives a Write command or a Read command. Although not mentioned in the above description, the Total Read counter is also shown here. This Total Read counter is a counter for each tape 23, not for each dm-id, and is a counter that is updated when the tape 23 is mounted for reading. On the other hand, in this figure, Dm-readfreq is omitted. In this figure, D1 to D3 are generalized DMs and are shown as data.

First, as shown in A, it is assumed that a Write command for D1 is sent. In this case, it is assumed that the Total Read counter is 0. D1 Ram-count and Dm-count are created, but their values are zero.
Next, as shown in B, it is assumed that a Read command for D1 is sent. In this case, the Total Read counter is 1. In addition, the Ram-count and Dm-count of D1 are also 1.
Next, as shown in C, it is assumed that a Read command for D1 is sent. In this case, the Total Read counter is 2. In addition, the Ram-count and Dm-count of D1 are also 2.

Also, as shown in D, it is assumed that a Write command for D2 is sent. In this case, the Total Read counter remains 2, and the Ram-count and Dm-count of D1 also remain 2. In addition, Ram-count and Dm-count of D2 are created, but their values are zero.
Next, as shown in E, it is assumed that a Read command for D1 is sent. In this case, the Total Read counter is 3. Also, the Ram-count of D1 is 3, and since D1 is read, the Dm-count of D1 is also 3. On the other hand, the Ram-count of D2 is 1, but since D2 has not been read, the Dm-count of D2 remains 0.
Next, as shown in F, it is assumed that a Read command for D2 is sent. In this case, the Total Read counter is 4. On the other hand, the Ram-count of D1 is 4, but since D1 has not been read, the Dm-count of D1 remains 3. In addition, the Ram-count of D2 is 2, and D2 is read, so the Dm-count of D2 is 1.

  Thereafter, when the Write command and the Read command are sent in G, H, and I, the total read counters, the Ram-count and Dm-count of D1, D2, and D3 are updated in the same manner as described above.

Next, the rearrangement based on the access frequency when data and metadata are transferred to another tape 23 will be described.
When the data and metadata on the tape 23 are migrated to another tape 23 (or the tape 23 itself), the access on the tape 23 starts from the one having the high access frequency in the dm access counter table shown in FIG. Data and metadata corresponding to the dm-id are recorded in a place advantageous in terms of performance.

As a method for determining the access frequency, a method by calculating a Dm-count / Total Read counter is the most basic (hereinafter, this determination method will be referred to as a “basic method”). However, when the access frequency is obtained by this basic method, the result varies depending on when the DM is written on the tape 23. For example, the DM first written on the tape 23 cannot be simply compared with the DM already written after the Ram-count reaches 100.
Therefore, in the present embodiment, the following access frequency determination method is adopted.

(First determination method)
The first determination method is a method of deriving a relative access frequency for the DM by calculating Dm-count / Ram-count.
This will be described with reference to the example shown in FIG. However, in this case, it is assumed that an abbreviated portion “....” Of Dm-readfreq in FIG.
First, when the basic method is used, if the total read counter is 250, the access frequency of dm01 is 0.02 (= 5/250), the access frequency of dm02 is 0, and the access frequency of dm03 is 0.012 (= 3/250), the access frequency of dm04 is 0.016 (= 4/250), and the access frequency of dm05 is 0.016 (= 4/250). That is, the access frequency increases in the order of dm01, dm05, dm04, dm03, dm02 (if the points are the same, the access frequency that is more recent is defined as the higher access frequency).
On the other hand, when the first determination method is used, the access frequency of dm01 is 0.023 (= 5/215), the access frequency of dm02 is 0, the access frequency of dm03 is 0.231 (= 3/13), dm04 Access frequency is 0.026 (= 4/156), and dm05 access frequency is 0.051 (= 4/79). That is, the access frequency increases in the order of dm03, dm05, dm04, dm01, and dm02, and the order is different from the case where the basic method is used.

(Second determination method)
The second determination method is a method in which the Ram-count held for each DM is used as the criteria for the DM for calculating the access frequency.
For example, a recently written DM may be excluded from the access frequency calculation target. In this case, DMs with a Ram-count of 3 or less may be excluded.
In addition, it is also conceivable that past cartridge read mounts are limited to the number of times from the number of times, that is, the access frequency to the DM is measured only during a certain period. In this case, for example, the target DM for calculating the access frequency is limited to those with Ram-count of 100 to 70. Then, the number of accesses to the DM corresponding to the 100th to 70th times of the read mount is calculated from Dm-readfreq, and they are compared and arranged in the descending order of frequency.
Note that the Ram-count range considered in the second determination method is an example of a predetermined range for limiting a target whose access frequency is grasped.

When the access frequency is calculated in this way, the tape drive 10 migrates data from one tape to another based on this access frequency.
FIG. 8 is a diagram showing an example of data and metadata on the tape 23 after data migration. Here, as in FIG. 4, there are a plurality of wraps on the tape 23, and the white arrow indicates that the data writing direction is reversed, and the solid arrow indicates that the running direction of the tape 23 is reversed. ing. In this example, it is assumed that dm02, dm03, dm01, and dm04 are from the highest access frequency.
Here, first, DM with dm-id dm02 is written in the order of metadata and data in the forward direction with respect to wrap # 0. Next, metadata is written in the DM with the dm-id dm03 in the opposite direction to the wrap # 1. Thereafter, in the forward direction again with respect to lap # 2, data is written out of DMs with dm-id dm04, and DMs with dm-id dm04 are written in the order of metadata and data.
That is, after the data transfer, unlike the example of FIG. 4, data and metadata are recorded for each logical association shown in FIG. As a result, when data and metadata are sequentially read from the tape 23, it is not necessary to perform unnecessary alignment of the tape 23, thereby improving access performance.

In such a data migration method, data is arranged according to the access frequency calculated in the present embodiment, so that the access time to data with a truly high access frequency is shortened.
In addition, the data and metadata access performance can be improved by rearranging the data and metadata distributed on the tape 23 before migration according to a certain rule (in this embodiment, in order). .
A digital data archive system or the like may employ a format in which the tape 23 is used as a secondary or tertiary storage as a hierarchical storage mounting format for storing a large amount of data. In that case, even if it is an archive, data with high access frequency exists and is assumed to be actually accessed. This embodiment also contributes to improvement of access performance to such data.

By the way, in the above, the data and metadata distributed and arranged on the migration source tape 23 are sequentially stored in the migration destination tape 23 for each dm-id. However, the arrangement method on the transfer destination tape 23 is not limited to this.
Hereinafter, other arrangement methods will be described. In the arrangement method described here, when data is re-arranged on the tape 23, data with high access frequency is arranged at a position where the data can be accessed in a short time after the tape 23 is mounted in consideration of the characteristics of the medium called the tape. Is.

First, as a characteristic of the tape storage device, the following preparation operation is performed to read data.
1. Mount the cartridge on the storage device.
2. The position of the tape 23 is adjusted to the head position where the data to be read is stored.
That is, immediately after the cartridge is mounted, the head of the storage device is positioned at the head of the tape 23. Therefore, it can be said that the data arranged at a position near the head of the tape 23 has a higher access speed than the data arranged at a position near the end of the tape 23. This is because when data is recorded near the end of the tape 23, it is necessary to align the position by running the tape 23 so far.

FIG. 9A shows normal data reading / writing on the tape 23 compliant with the LTO (Linear Tape-Open) standard.
First, after reading and writing data to lap # 0 in the right direction as indicated by arrow 201, the running direction of tape 23 is reversed as indicated by arrow 202. Next, after reading and writing data to the lap # 1 in the left direction, the running direction of the tape 23 is reversed as indicated by an arrow 203. Further, after reading and writing data to lap # 2 in the right direction, the running direction of the tape 23 is reversed as indicated by an arrow 204. Finally, as indicated by an arrow 205, data is read / written to the wrap # 55 in the left direction.

Thus, in the LTO, data is read and written while reciprocating a data storage area called a wrap defined on the tape 23 from the beginning to the end of the tape 23. In the case of performing sequential reading and writing, it reciprocates on the wrap in this way, but since the mechanical operation of switching the wrap by moving the head in the width direction of the tape 23 is completed in a short time, even in such an LTO format, The position near the head of the tape 23 can be accessed in a shorter time than the position near the end of the tape 23.
Therefore, an arrangement method in which data with high access frequency is arranged near the head of the tape 23 and data with low access frequency is arranged near the end of the tape 23 is preferable. Here, the top of the tape 23 is an example of a first end, and the end of the tape 23 is an example of a second end.

Some tape storage devices have a function called segmentation. Segmentation is a method in which the tape 23 is divided into a plurality of segments, and data is sequentially written from the first segment. Here, the segment is an example of a plurality of regions formed by dividing at least one position in the longitudinal direction.
FIG. 9B shows data reading and writing when segmentation is used.
First, after reading / writing data in the right direction in the segment # 1 portion of the wrap # 0 as indicated by an arrow 211, the running direction of the tape 23 is reversed as indicated by an arrow 212. Next, after reading and writing data to the segment # 1 portion of the wrap # 1 in the left direction, the running direction of the tape 23 is reversed as indicated by an arrow 213. Further, after the data is read / written to the segment # 1 portion of the wrap # 2 in the right direction, the running direction of the tape 23 is reversed as indicated by an arrow 214. Then, as indicated by an arrow 215, data is read from and written to the segment # 1 portion of the wrap # 55 in the left direction.

  Thereafter, as indicated by an arrow 220, the head position is moved to the segment # 2 portion of the lap # 0. Then, after reading / writing data in the right direction in the segment # 2 portion of the wrap # 0 as indicated by an arrow 221, the running direction of the tape 23 is reversed as indicated by an arrow 222. Next, after reading / writing data to the segment # 2 portion of the wrap # 1 in the left direction, the running direction of the tape 23 is reversed as indicated by an arrow 223. Further, after data is read / written to the segment # 2 portion of the wrap # 2 in the right direction, the running direction of the tape 23 is reversed as indicated by an arrow 224. Finally, as indicated by an arrow 225, data is read / written to the left in the segment # 2 portion of the wrap # 55.

Thus, segment # 1 has a faster access speed than segment # 2. If three or more segments are provided, it can be generally said that the segment closer to the head of the tape 23 has a higher access speed than the segment closer to the end of the tape 23.
Therefore, using this characteristic, when data is transferred to the tape 23, data and metadata with high access frequency are arranged in a segment near the head of the tape 23, and data and metadata with low access frequency are placed on the tape 23. It is preferable to arrange in a segment close to the end of. Here, the segment near the head of the tape 23 is an example of a first area, and the segment near the end of the tape 23 is an example of a second area.

FIG. 10 is a diagram illustrating an example of data and metadata on the tape 23 after data migration using segmentation. Here, as in FIG. 4, there are a plurality of wraps on the tape 23, and the white arrow indicates that the data writing direction is reversed, and the solid arrow indicates that the running direction of the tape 23 is reversed. ing. Also in this example, it is assumed that dm02, dm03, dm01, and dm04 are from the highest access frequency.
Here, first, the DM having the dm-id dm02 is written in the order of metadata and data in the forward direction with respect to the segment # 1 portion of the wrap # 0. Next, the remaining portion of the data in the DM with dm-id dm02 is written in the opposite direction to the segment # 1 portion of wrap # 1. Thereafter, DM with dm-id dm03 is written in the order of metadata and data again in the forward direction with respect to the segment # 1 portion of wrap # 2. If there is a continuation of the data whose d-id is d3, it is written in the segment # 1 portion of the wrap # 3, although not shown.

Thereafter, DM with dm-id dm01 is written in the order of metadata and data in the forward direction with respect to the segment # 2 portion of wrap # 0. Next, in the reverse direction to the segment # 2 portion of wrap # 1, DM with dm-id dm04 is written in the order of metadata and data.
That is, after the data transfer, unlike the example of FIG. 4, data and metadata are recorded for each logical association shown in FIG. As a result, when data and metadata are sequentially read from the tape 23, it is not necessary to perform unnecessary alignment of the tape 23, thereby improving access performance.

  Here, the data transfer destination tape 23 is one, but a plurality of tapes 23 may be the transfer destination. In that case, use segment # 1 of the first tape 23, use segment # 1 of the second tape 23 when it is used up, and use segment # 1 of the third tape 23 when it is used up. This is repeated until the segment # 1 of all the transfer destination tapes 23 are used up. Then, when the segment # 1 of all the transfer destination tapes 23 is used up, the segment # 2 may be used in order from the first tape 23.

In this way, by preparing a plurality of migration destination tapes 23, a larger storage space can be provided, and the amount of data and metadata that can be written to segment # 1 with high access performance can be increased. That is, the amount of data and metadata that can be accessed in a short time increases.
Further, in this method, by changing the number of transfer destination tapes 23 and the method of segment definition, it is possible to provide a storage space according to the access frequency of data on the transfer source tape 23. For example, when there is little data with high access frequency, the first segment may be defined from the beginning of the tape 23. If there is a lot of frequently accessed data, it is conceivable to increase the migration destination tape 23 or to define the first segment larger as required. Furthermore, when there is a variation in the access frequency, it is conceivable to increase the number of segments, group data for each frequency, and write to each segment.

In such a data migration method, data is arranged according to the access frequency calculated in the present embodiment, so that the access performance is improved when individual access is made to data with a truly high access frequency.
A digital data archive system or the like may employ a format in which the tape 23 is used as a secondary or tertiary storage as a hierarchical storage mounting format for storing a large amount of data. In that case, even if it is an archive, data with high access frequency exists and is assumed to be actually accessed. The present embodiment contributes to the improvement of the access performance of data that has been frequently accessed in the past in a situation where these archived data are individually accessed.

  Note that, in the above, focusing on the access performance when reading data and metadata, the arrangement method of data and metadata has been considered. However, an arrangement method irrespective of access performance when reading data and metadata may be adopted.

  Next, the tape drive 10 that performs the operation as described above will be described in detail. Here, it is assumed that the dm access counter table is stored in the cartridge memory 24. 1 shows the tape drive 10 in which only one cartridge 20 can be loaded, but here, the tape drive 10 in which two cartridges 20 can be loaded is assumed, and one cartridge 20 to the other cartridge 20 is assumed. It is assumed that data migration can be performed by one tape drive 10 without replacing the cartridge 20.

FIG. 11 is a block diagram illustrating a functional configuration example of the controller 16 of the tape drive 10.
As illustrated, the controller 16 includes a command processing unit 61, a buffer management unit 62, a channel input / output unit 63, a cartridge memory input / output unit (hereinafter referred to as “CM input / output unit”) 64, and a frequency calculation unit. 65, a transition control unit 66, a head position management unit 67, and a tape running management unit 68.

  Among these, the command processing unit 61 receives a command from the host I / F 11. Here, as a command, a Write command that instructs to store data in the buffer 12, a synchronous command that instructs to write data stored in the buffer 12 to the tape 23, and an instruction to read data from the tape 23 There is a Read command and the like.

  The buffer management unit 62 prepares data in the buffer 12 when the command processing unit 61 receives a Write command. When the command processing unit 61 receives a synchronization command, the data is read from the buffer 12 and output to the channel input / output unit 63. Further, when the command processing unit 61 receives a Read command, if the target data is not in the buffer 12, it instructs the channel input / output unit 63 to read the data, and the target data is stored in the buffer 12. If it exists, the data is returned to the host 200 via the command processing unit 61.

  The channel input / output unit 63 outputs the data read from the buffer 12 by the buffer management unit 62 to the channel 13 and outputs the data received from the channel 13 to the buffer management unit 62.

  The CM input / output unit 64 writes information to the dm access counter table stored in the cartridge memory 24 via the CMI / F 19 and reads information from the dm access counter table via the CMI / F 19. In the present embodiment, the first information indicating the number of times any data is read from the first recording medium after each data is written to the first recording medium, and the first information from the first recording medium. As an example of an acquisition unit that acquires second information indicating the number of times data has been read, and an example of an update unit that updates the first information and the second information, a CM input / output unit 64 is provided. Provided.

The frequency calculation unit 65 calculates the access frequency for each DM based on the information read by the CM input / output unit 64. In the present embodiment, a frequency calculation unit 65 is provided as an example of a grasping unit that grasps the access frequency of each data.
The migration control unit 66 controls data migration based on the access frequency calculated by the frequency calculation unit 65. That is, the buffer management unit 62 is instructed to read data and metadata from the migration source tape 23 and write them to the migration destination tape 23 in order of dm-id having the highest access frequency. In this way, since data and metadata corresponding to dm-id having a high access frequency are sequentially written to the migration destination tape 23, the migration control unit 66 determines the positions of the data and metadata on the migration destination tape 23. I can say that. That is, in the present embodiment, the shift control unit 66 is provided as an example of a determination unit that determines the recording position of each data on the second recording medium.

The head position management unit 67 outputs a signal for offsetting the positions of the write head 14 a and the read head 14 b with respect to the tape 23 in the width direction of the tape 23 to the head position control system 17. Further, information regarding the current position of the write head 14a and the read head 14b in the width direction on the tape 23 is acquired.
The tape running management unit 68 outputs a signal for running the tape 23 in the forward direction and a signal for running the tape 23 in the reverse direction to the motor driver 18.

Next, the operation of the tape drive 10 in the present embodiment will be described in detail.
First, the operation of the tape drive 10 when a write command is sent from the host 200 will be described.
The host 200 sends a Write command with a dm-id that associates data and metadata to the archive device 100. Then, in the archive device 100, the control mechanism 30 generates a d-id that identifies this data and an m-id that identifies this metadata, and associates the dm-id, the d-id, and the m-id. And register as dm-a. Further, the cartridge 20 to which data and metadata are written is determined, the d-id, the m-id, and the cartridge 20 are associated and registered as dm-l, and the accessor 40 is instructed to load the cartridge 20 into the tape drive 10. To do. Then, the control mechanism 30 sends a Write command with dm-id, d-id, m-id, data, and metadata to the tape drive 10. Thereby, the operation of the tape drive 10 is started.

FIG. 12 is a flowchart showing an operation example of the controller 16 of the tape drive 10 at this time.
In the controller 16, first, the command processing unit 61 receives a Write command accompanied by dm-id, d-id, m-id, data, and metadata via the host I / F 11 (step 601). At this time, dm-id, d-id, m-id, data, and metadata are transferred to the buffer management unit 62 and accumulated in the buffer 12 by the buffer management unit 62.

Next, the controller 16 writes the data and metadata to the tape 23 of the cartridge 20 loaded in the tape drive 10 (step 602).
Specifically, the head position management unit 67 that has received an instruction from the buffer management unit 62 controls the head position control system 17 to align the write head 14 a in the width direction of the tape 23 and from the buffer management unit 62. The tape running management unit 68 that has received the instruction controls the motor driver 18 to run the tape 23. On the other hand, the buffer management unit 62 reads the data and metadata from the buffer 12 and passes them to the channel input / output unit 63. The channel input / output unit 63 outputs the data and metadata to the write head 14a via the channel 13.
At this time, the channel input / output unit 63 receives information on the storage position in the longitudinal direction of the data and metadata on the tape 23 from the servo head via the channel 13, and the head position management unit 67 Information on the storage positions of the data and metadata in the width direction is received from the head position control system 17. The information on these recording positions is transferred to the buffer management unit 62, and the buffer management unit 62 outputs the d-id and m-id read from the buffer 12 and the information on these recording positions to the CM input / output unit 64. To do. Then, the CM input / output unit 64 writes the correspondence information between the d-id and the data recording position and the correspondence information between the m-id and the metadata recording position into the cartridge memory 24 via the CMI / F 19. As a result, it is possible to read data using d-id as a key and to read metadata using m-id as a key.

Thereafter, the controller 16 creates an entry corresponding to dm-id in the dm access counter table stored in the cartridge memory 24 (step 603).
Specifically, the buffer management unit 62 reads the dm-id from the buffer 12 and passes it to the CM input / output unit 64, and the CM input / output unit 64 sends a record including the dm-id to the cartridge memory 24 via the CMI / F 19. Write.

Next, the operation of the tape drive 10 when a Read command is sent from the host 200 will be described. Here, it is assumed that the cartridge 20 to be read is mounted on the tape drive 10 whenever there is a read request for data and metadata from the host 200.
The host 200 sends a Read command with dm-id to the archive device 100. Then, in the archive device 100, the control mechanism 30 specifies d-id and m-id associated with dm-id based on dm-a. Further, the cartridge 20 associated with d-id and m-id is specified based on dm-l, and the accessor 40 is instructed to load the cartridge 20 into the tape drive 10. Then, the control mechanism 30 sends a Read command with dm-id, d-id, and m-id to the tape drive 10. Thereby, the operation of the tape drive 10 is started.

FIG. 13 is a flowchart showing an operation example of the controller 16 of the tape drive 10 at this time.
In the controller 16, first, the command processing unit 61 receives a Read command with dm-id, d-id, and m-id via the host I / F 11 (step 621). At this time, dm-id, d-id, and m-id are passed to the buffer management unit 62.

Next, the controller 16 reads data corresponding to d-id and metadata corresponding to m-id from the tape 23 of the cartridge 20 loaded in the tape drive 10 (step 622).
Specifically, the head position management unit 67 that has received an instruction from the buffer management unit 62 controls the head position control system 17 to align the read head 14 b in the width direction of the tape 23 and from the buffer management unit 62. The tape running management unit 68 that has received the instruction controls the motor driver 18 to run the tape 23. On the other hand, the data and metadata read by the read head 14 b are acquired by the channel input / output unit 63 via the channel 13 and passed to the buffer management unit 62, and the buffer management unit 62 accumulates the data and metadata in the buffer 12. . The data and metadata stored in the buffer 12 are transmitted to the host 200.

Thereafter, the controller 16 updates the dm access counter table stored in the cartridge memory 24. This update process starts when the CM input / output unit 64 reads the dm access counter table from the cartridge memory 24 into its own memory area via the CMI / F 19.
In the update process, first, the CM input / output unit 64 reads one record from the dm access counter table (step 623). Then, it is determined whether or not the Ram-count included in the read record is a multiple of 8 (step 624). As a result, if Ram-count is a multiple of 8, a new read mount unit starting from the current read mount with 8 as one unit starts. Therefore, initialization is performed by substituting 0 into Dm-count, and “00” is added to the right end of Dm-readfreq (step 625). On the other hand, if Ram-count is not a multiple of 8, the process proceeds to the next process.

  Next, the CM input / output unit 64 adds 1 to the Ram-count included in the read record (step 626). Then, it is determined whether or not the dm-id included in the read record matches the dm-id specified by the Read command (step 627). As a result, if the dm-ids match, 1 is added to Dm-count, and the two rightmost bits of Dm-readfreq are updated according to the result (step 628). For example, it is assumed that the relationship between the value of Dm-count and 2 bits of Dm-readfreq is as shown in FIG. In this case, if Dm-count after adding 1 is 0, the rightmost 2 bits of Dm-readfreq are updated to be “00”, and Dm-count after adding 1 is 1 or 2 Then, the rightmost 2 bits of Dm-readfreq are updated to be “01”, and if the Dm-count after adding 1 is 3 to 5, the rightmost 2 bits of Dm-readfreq is “ If the Dm-count after adding 1 is 6-8, the rightmost 2 bits of Dm-readfreq are updated so as to be “11”. Note that the rules for updating Dm-readfreq according to the value of Dm-count may be coded directly in the program or defined in a table outside the program and acquired by the program referring to the table. May be.

  Thereafter, the CM input / output unit 64 determines whether or not the record read from the dm access counter table is the last record (step 629). If it is not the last record, the processing in steps 623 to 628 is repeated for the next record. If it is the last record, the process is terminated.

At any time after the dm access counter table is updated in this manner, the tape drive 10 performs processing related to data migration.
FIG. 14 is a flowchart showing an operation example of the controller 16 of the tape drive 10 at this time. Here, the description will be made on the assumption that the data and metadata are arranged as shown in FIG.

In the tape drive 10, the CM input / output unit 64 first reads the dm access counter table from the cartridge memory 24 via the CMI / F 19 and transfers it to the frequency calculation unit 65 (step 641).
Then, the frequency calculation unit 65 calculates the access frequency for each dm-id using the first determination method or the second determination method described above, rearranges the records in the dm access counter table in descending order of access frequency, The rearranged dm access counter table is transferred to the migration control unit 66 (step 642).

Thereby, the transfer control unit 66 transfers data from one tape 23 to the other tape 23.
That is, first, the migration control unit 66 reads one record from the rearranged dm access counter table and transmits it to the buffer management unit 62 (step 643). Then, the buffer management unit 62 reads data and metadata corresponding to the dm-id included in the read record from the migration source tape 23 (step 644).
Specifically, the head position management unit 67 that has received an instruction from the buffer management unit 62 controls the head position control system 17 to align the read head 14 b in the width direction of the tape 23 and from the buffer management unit 62. The tape running management unit 68 that has received the instruction controls the motor driver 18 to run the tape 23. On the other hand, the data and metadata read by the read head 14 b are acquired by the channel input / output unit 63 via the channel 13 and passed to the buffer management unit 62, and the buffer management unit 62 accumulates the data and metadata in the buffer 12. . It should be noted that d-id and m-id, which are keys for reading data and metadata corresponding to dm-id, may be specified by inquiring the control mechanism 30, for example.

Next, the buffer management unit 62 determines whether data and metadata can be written on the current wrap of the migration destination tape 23 (step 645). Specifically, the channel input / output unit 63 receives information on the storage position in the longitudinal direction of the data and metadata on the tape 23 from the servo head via the channel 13, and thus makes a determination based on the information.
As a result, if it is determined that data can be written on the current wrap, the data and metadata are directly written on the migration destination tape 23 (step 649).
Specifically, the head position management unit 67 that has received an instruction from the buffer management unit 62 controls the head position control system 17 to align the write head 14 a in the width direction of the tape 23 and from the buffer management unit 62. The tape running management unit 68 that has received the instruction controls the motor driver 18 to run the tape 23. On the other hand, the buffer management unit 62 reads the data and metadata from the buffer 12 and passes them to the channel input / output unit 63. The channel input / output unit 63 outputs the data and metadata to the write head 14a via the channel 13.

On the other hand, if it is determined that the current lap cannot be written, it is determined whether the current lap is the last lap (step 646). Specifically, since the head position management unit 67 receives information on the storage position in the width direction of the data and metadata on the tape 23 from the head position control system 17, the determination is made based on the information.
If the current lap is not the last lap, the lap is switched to the next lap (step 647), and the data and metadata are written to the destination tape 23 (step 649). The lap switching is performed by the head position management unit 67 controlling the head position control system 17.
If the current lap is the last lap, the lap # 1 of the next segment is switched (step 648), and the data and metadata are written to the destination tape 23 (step 649). The segment switching is performed by the tape running management unit 68 controlling the motor driver 18, and the lap switching is performed by the head position management unit 67 controlling the head position control system 17.

  Thereafter, the migration control unit 66 determines whether the record read from the dm access counter table is the last record (step 650). If it is not the last record, the processes in steps 643 to 649 are repeated for the next record. If it is the last record, the process is terminated.

  In this operation example, the data transfer is performed without replacing the two cartridges 20 in the tape drive 10 that can be loaded with two cartridges 20 at the same time. However, such control of data migration may be performed by the control mechanism 30. In this case, the data transfer may be performed by the control mechanism 30 controlling the two tape drives 10 in which only one cartridge 20 can be loaded.

  Here, the present invention may be realized entirely by hardware or entirely by software. It can also be realized by both hardware and software. The present invention can be realized as a computer, a data processing system, and a computer program. This computer program may be stored and provided on a computer readable medium. Here, the medium may be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (apparatus or equipment), or a propagation medium. Examples of computer-readable media include semiconductors, solid state storage devices, magnetic tape, removable computer diskettes, random access memory (RAM), read-only memory (ROM), rigid magnetic disks, and optical disks. The Current examples of optical disks include compact disk-read only memory (CD-ROM), compact disk-read / write (CD-R / W) and DVD.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the said embodiment. It will be apparent to those skilled in the art that various modifications and alternative embodiments can be made without departing from the spirit and scope of the invention.

It is the block diagram which showed the structure of the archive device with which embodiment of this invention is applied. It is a figure for demonstrating the correlation of dm-id, d-id, and m-id in the embodiment of the present invention. It is the figure which showed the specific example of correlation of dm-id, d-id, and m-id in embodiment of this invention. It is the figure which showed the state by which data and metadata were distributed and arranged on the tape before moving data in the embodiment of the present invention. It is a figure for demonstrating concretely about the counter used by embodiment of this invention. It is the figure which showed an example of the dm access counter table used in embodiment of this invention. It is a figure for demonstrating concretely about the update of the counter in embodiment of this invention. It is the figure which showed the state by which data and metadata were arrange | positioned on the tape in order of access frequency after migrating data by embodiment of this invention. It is a figure for demonstrating the part with a quick access frequency in a tape, and a slow part. It is the figure which showed the modification of the state in which data and metadata were arrange | positioned in order of access frequency on the tape after data transfer in embodiment of this invention. It is the block diagram which showed the function structural example of the controller in embodiment of this invention. It is the flowchart which showed the operation example of the controller at the time of Write command reception in embodiment of this invention. It is the flowchart which showed the operation example of the controller at the time of Read command reception in embodiment of this invention. It is the flowchart which showed the operation example of the controller at the time of the data transfer in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Tape drive, 11 ... Host I / F, 12 ... Buffer, 13 ... Channel, 14a ... Write head, 14b ... Read head, 15 ... Motor, 16 ... Controller, 17 ... Head position control system, 18 ... Motor driver, 19 ... CMI / F, 20 ... cartridge, 30 ... control mechanism, 40 ... accessor, 50 ... cartridge slot, 100 ... archive device, 200 ... host

Claims (14)

An apparatus for supporting the transfer of data recorded on a first recording medium to a second recording medium,
For each data recorded on the first recording medium, a first indicating the number of times any data is read from the first recording medium after the data is written to the first recording medium. And an acquisition unit for acquiring the second information indicating the number of times each piece of data has been read from the first recording medium;
A grasping unit for grasping an access frequency of each data using the first information and the second information obtained by the obtaining unit;
A determination unit configured to determine a recording position of each data on the second recording medium based on the access frequency grasped by the grasping unit.   When specific data is read from the first recording medium, an update unit is further included that updates the first information for each data and updates the second information for the specific data. The apparatus of claim 1. The specific data includes main data and metadata that is additional information of the main data,
The apparatus according to claim 2, wherein the update unit determines that the specific data has been read when at least one of the main data and the metadata is read from the first recording medium.   According to the first information, each data is written on the first recording medium according to the number of times the first recording medium is mounted on the storage device after each data is written on the first recording medium. The apparatus of claim 1, wherein the apparatus indicates the number of times any data has been read from the first recording medium.   The apparatus according to claim 1, wherein the grasping unit grasps, as the access frequency, a ratio of the number of times indicated by the second information to the number of times indicated by the first information.   The apparatus according to claim 1, wherein the grasping unit grasps the access frequency using the second information for data in which the number of times indicated by the first information is within a predetermined range.   The second information is information representing a history of reading each data with a plurality of bit values, and the number of times each data is read out of M times of reading data from the first recording medium. 2. The apparatus of claim 1, wherein is represented by N (N <M) bit values of the plurality of bit values.   The determination unit records the data with high access frequency at a first recording position on the second recording medium, and the data with low access frequency on the first recording position on the second recording medium. The apparatus according to claim 1, wherein the recording position of each data on the second recording medium is determined so that the data is recorded at the second recording position that takes longer to read data than the data. The second recording medium is a tape medium from which data reading is started from the first end toward the second end,
9. The apparatus of claim 8, wherein the second recording position is closer to the second end than the first recording position. The second recording medium starts reading data from the first end toward the second end, and divides a plurality of regions formed by dividing at least one position in the longitudinal direction. A tape medium having
The first recording position is a position in a first area of the plurality of areas,
9. The apparatus according to claim 8, wherein the second recording position is a position in a second area of the plurality of areas that is closer to the second end than the first area. An apparatus for supporting data rearrangement on a recording medium,
For each data recorded on the recording medium, first information indicating the number of times any data is read from the recording medium after each data is written to the recording medium; An acquisition unit for acquiring second information indicating the number of times each data is read;
A grasping unit for grasping an access frequency of each data using the first information and the second information obtained by the obtaining unit;
A determination unit configured to determine a recording position of each data on the recording medium based on the access frequency grasped by the grasping unit. An apparatus for transferring data recorded on a first tape medium to a second tape medium,
For each data recorded on the first tape medium and including main data and metadata indicating accompanying information of the main data, the first data after the data is written on the first tape medium. First information indicating the number of times the tape medium is mounted on the tape drive for reading any data, and second information indicating the number of times each piece of data is read from the first tape medium; An acquisition unit for acquiring
A grasping unit for grasping, as an access frequency of each data, a ratio of the number of times indicated by the second information to the number of times indicated by the first information acquired by the acquiring unit;
And a writing unit that writes data to the second tape medium in order from the data with the highest access frequency grasped by the grasping unit. A method for supporting the transfer of data recorded on a first recording medium to a second recording medium,
For each data recorded on the first recording medium, a first indicating the number of times any data is read from the first recording medium after the data is written to the first recording medium. Obtaining the second information indicating the number of times each piece of data has been read from the first recording medium;
Using the acquired first information and the second information to grasp the access frequency of each data;
Determining a recording position of each data in the second recording medium based on the grasped access frequency. A program that causes a computer to function as a device that supports the transfer of data recorded on a first recording medium to a second recording medium,
The computer,
For each data recorded on the first recording medium, a first indicating the number of times any data is read from the first recording medium after the data is written to the first recording medium. And an acquisition unit for acquiring the second information indicating the number of times each piece of data has been read from the first recording medium;
A grasping unit for grasping an access frequency of each data using the first information and the second information obtained by the obtaining unit;
A program that functions as a determination unit that determines a recording position of each data in the second recording medium based on the access frequency grasped by the grasping unit.

Images