JP4496790B2 - Data storage device and method, and recording / reproducing system - Google Patents

Description

  The present invention relates to a data storage device and method for managing and storing various data such as AV data as files under the control of a host device, and a recording / reproducing system including the host device and the data storage device. The present invention relates to a data storage device and method having a recording medium such as a disk having different data transfer rates depending on the access position, and a recording / reproducing system.

  The FAT file system is a medium in a host device such as a PC (Personal Computer) such as a hard disk drive (HDD) or a non-volatile solid memory (Sony: Memory Stick (registered trademark), Toshiba); (Registered trademark), manufactured by SanDisk: compact flash (registered trademark), multimedia card, etc.).

  The FAT file system includes a FAT (File Allocation Table) used to indicate where each file is arranged on the recording medium, and a directory item indicating the attribute of the file and where the file exists in the directory. One data is used.

  Usually, an area for FAT and root directory is provided on a recording medium for exclusive use, and the PC transfers such information necessary for file access from a recording / reproducing apparatus such as an HDD via a PC interface (SCSI (Small Computer System). Interface), IDE (Integrated Drive Electronics), IEEE 1394, USB (Universal Serial Bus), etc.) and recording / reproduction control based on them.

  For example, when writing a file to the HDD, the data is written to an empty cluster, and when the writing is completed, information indicating which cluster is to be used next is written to the FAT item. When erasing a file, the written data is left as it is, and the FAT item corresponding to the used cluster is set as an empty cluster. Further, at the time of reading, the starting cluster address of the file is searched from the directory item, the FAT item corresponding to the file is read to know which cluster has the file data to be read, and reading from the medium is performed.

  In such data transfer between the host device and the HDD, a disk cache device is used, and the information written to the hard disk and the information read from the hard disk are temporarily cached, so that the appearance of hard disk access is apparent. You can increase the top speed. In addition, when a nonvolatile semiconductor memory such as flash EEPROM (electrically erasable and programmable Read Only Memory) is used for the cache as in the technique described in Patent Document 1, for example, the cache is not necessarily turned off before the power is turned off. This is convenient because it is not necessary to write the contents back to the hard disk.

Japanese Patent Laid-Open No. 11-45210

  However, as in the technique described in Patent Document 1 described above, configuring a storage device with only a non-volatile solid-state memory has a problem that costs increase.

  By the way, in such a recording / reproducing apparatus using the FAT file system, when the HDD is newly started or when the power saving mode is shifted to the normal mode, data transfer becomes possible. Takes about several seconds, and the HDD requires a time required for a seek operation in units of 10 ms. Therefore, there is a problem that writing or reading cannot be executed immediately. Furthermore, further improvement in transfer speed is required from the application side.

  In order to avoid such problems, methods such as increasing the rotational speed of the HDD, increasing the linear recording density, or simultaneously reading and writing with a plurality of heads can be considered, but they are used for consumer devices. It is difficult to realize these methods on an HDD that does not have such high capability.

  The present invention has been proposed in view of such a conventional situation, and is efficient even during a period until data is transferred and can be written to a recording medium, such as a seek time and a rotation waiting time of a disc. It is an object of the present invention to provide a data storage device and method, and a recording / reproducing system that further improve the transfer rate of data from a host device.

In order to achieve the above-described object, a data storage device according to the present invention includes a rotary recording medium having a first area in which data transferred from a host device is temporarily written, and data transferred from the host device. A non-volatile solid-state memory having a second area to be temporarily written, and a control means for controlling the data transferred from the host device to be written to the recording area of the rotary recording medium and the non-volatile solid-state memory, The recording area includes a first writing area composed of the first area and the second area in which data transferred from the host device is temporarily written, and data from the first area in the rotary recording medium. and a second write area including a third region transfer rate slow, the control means, of the data transferred from the host device, the top portion Data written to the second region, data other than the head portion is controlled to write parallel to the first region and the second region, written in the first write area at a predetermined timing The data is moved to the second write area including the third area and the first write area is released, and the predetermined timing is the FLUSH operation or mode change from the host device. Or when the access request from the host device has not been set for a predetermined time or more.

  According to the present invention, in a hybrid storage device having a recording medium and a nonvolatile solid-state memory as a recording area, a part of the recording medium (first area) having a high data transfer speed and the non-volatile solid-state memory A part of the area (second area) is used as a first writing area in which data transferred from the host device is temporarily written, and these data are transferred to other areas of the recording medium (at a predetermined timing). By moving to the second writing area including the third area) and releasing the first writing area, the first writing area can be used efficiently, so that a large amount of data from the host device can be obtained. It can also respond to write requests.

  Further, the control means controls to write the data of the head part in the data transferred from the host device to the second area and write the data other than the head part to at least the first area. Thus, it is possible to prepare for writing to the first area while writing the data of the head portion in the nonvolatile solid-state memory having a high access speed.

  Further, the control means can control to write data other than the head portion of the data transferred from the host device in parallel to the first area and the second area. In addition, data can be written at a higher speed.

  Furthermore, the recording medium is a rotating recording medium, and the control means controls to write transfer data from the host device in the second area as long as it is necessary for the seek operation of the recording medium. Thus, data can be written to the second area until data can be recorded on the recording medium such as seek time or rotation waiting time.

  The second write area includes a fourth area other than the second area of the non-volatile solid-state memory, and the control means has an access request from the host device for a predetermined time or longer. When it does not occur, the data written in the first write area can be moved to the second write area, and at that time, the data written in the second area can be erased. This eliminates the need for an erasing operation before writing data to the nonvolatile solid-state memory.

  Further, when moving the data, the control means can control to write the data of the head portion into the fourth area and write data other than the head portion into at least the third area. While the head portion data is being read from the accessible non-volatile solid-state memory, the recording medium can be prepared for reading.

  A memory for storing a table for managing data transferred from the host device as a file, wherein the table is recorded in a predetermined area of the second writing area; The control means reads the table recorded in the predetermined area into the memory in response to a predetermined mode switching command from the host device, refers to the read table, and records the recording medium and the nonvolatile solid-state memory For example, a FAT (File Allocation Table) system or the like can be used.

  Further, the file is composed of a plurality of super clusters, and the super cluster is an integral multiple of a cluster which is an access unit of the host device, and the control means includes the recording medium and the non-volatile solid-state memory. Depending on the writing speed, control is performed so that data from the host device is written to the recording medium and the non-volatile solid memory in units of super clusters, or is written to the recording medium and the non-volatile solid memory in units of clusters. A switching function for switching between control and control can be provided, and the data amount and data size allocated to the nonvolatile solid-state memory and the recording medium can be changed as appropriate according to the data transfer speed to the nonvolatile solid-state memory.

In the data recording method according to the present invention, the control means includes a first area predetermined in the recording area of the rotary recording medium and a second area predetermined in the recording area of the nonvolatile solid-state memory. when no write the data transferred to one of the writing area from the host device temporarily, of the data transferred from the host device, the data of the beginning portion is written to the second region, other than the above head portion And writing the data in the first area and the second area in parallel, and the control means writes the data written in the first writing area at a predetermined timing to the rotating recording medium. Moving the recording area to a second writing area including at least a third area having a data transfer rate slower than that of the first area to release the first writing area; Yes, and the above predetermined timing is when the access request was not more than a predetermined time set in advance or from the host device when ordered to FLUSH operation or mode change from the host device.

The recording / reproducing system according to the present invention is a recording / reproducing system having a host device and a data storage device that accesses a rotating recording medium according to the control of the host device and writes data and / or reads data. The data storage device includes a rotary recording medium having a first area in which data transferred from the host device is temporarily written, and a second area in which data transferred from the host device is temporarily written. A nonvolatile solid-state memory having control means for controlling the data transferred from the host device to be written in a recording area of the rotary recording medium and the nonvolatile solid-state memory, and the recording area is from the host device. and the first region and the first write area of a second region where the transfer data is temporarily written, the And a second write area including a third region transfer rate slower data than the first regions in the rolling recording medium, said control means, of the data transferred from the host device, the top portion data written to the second region, data other than the head portion is controlled to write parallel to the first region and the second region, to the first write area at a predetermined timing The written data is controlled to move to the second write area including the third area to release the first write area, and the predetermined timing is the FLUSH operation or mode from the host device. This is a case where a change is ordered or a case where an access request from the host device has not been set for a predetermined time or more.

  According to the recording / reproducing apparatus and method of the present invention, a hybrid storage device including a recording medium and a nonvolatile solid-state memory is provided, and data transferred from the host device is temporarily written in the recording medium and the nonvolatile solid-state memory. A first writing area comprising a first area and a second area is provided. Then, the data transferred from the host device is temporarily written in the first writing area, and the second writing including the third area where the data transfer speed is slower than the first area of the recording medium at a predetermined timing. By moving to the area and controlling the first writing area to be released, a nonvolatile solid-state memory capable of high-speed access and a recording medium having a high writing speed can be used in combination. Writing speed and reading speed can be increased. For example, when a disk is provided as a recording medium, the seek time or rotation waiting time of the disk can be efficiently used by writing the head portion of the transferred data into the nonvolatile solid-state memory when writing. .

  Also, providing a recording / playback system suitable for consumer equipment such as digital cameras, which has a significantly higher write ratio than data read by increasing the data write speed without placing a burden on the host device. can do.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. In this embodiment, the present invention is applied to a storage device (hereinafter referred to as a disk) having a recording medium (hereinafter referred to as a disk) and a non-volatile solid state memory (hereinafter referred to as a disk) having different transfer speeds depending on data recording positions. This is applied to a recording / reproducing system comprising a hybrid storage device) and a host device connected thereto.

  The hybrid storage device of the present embodiment has two storage areas (recording areas) of a disk and a semiconductor solid-state memory, and writes and reads data in parallel or simultaneously according to the situation, thereby transferring the data transfer rate. It is intended to improve. In the disk, an area on the outer periphery side of the disk having a high data transfer speed is allocated as a cache as a first area where data can be temporarily written, and an area on the inner periphery side as a third area is assigned to the system. It is allocated to the area and the user area. Non-volatile solid-state memory can be used as a cache as a second area for temporarily writing transfer data, or as a storage device for storing transfer data as a fourth area. It is used properly according to the usage. For example, during the seek time and rotation waiting time when data is written in the cache area of the disk, the nonvolatile solid-state memory is used as a cache. Further, by storing only the head data of the file in the nonvolatile solid-state memory, the data can be read from the nonvolatile solid-state memory during the seek time, rotation waiting time, etc. when reading the data from the disk. As described above, by using the nonvolatile solid-state memory together with the disk, high-efficiency data transfer is possible even in the disk seek time, rotation waiting time, and the like.

  Here, the nonvolatile solid-state memory is, for example, a NAND flash memory that is used for a memory card and exchanges data in units of blocks. In this embodiment, for example, a transfer speed for one nonvolatile individual memory chip is used. In the following description, it is assumed that the writing speed is slower than the disk and the reading speed is about the same as that of the disk.

  In this recording / reproducing system, the MS-DOS compatible FAT file system is adopted as the file system, but any system that manages data as a file can be applied. Further, in this embodiment, the hybrid storage device is described as having a magnetic recording medium such as a hard disk. However, if the storage medium is a random accessible storage medium, the same processing can be performed on an optical disk such as a CD or DVD. It goes without saying that it is possible.

  FIG. 1 is a block diagram showing a recording / reproducing system according to the present embodiment. As shown in FIG. 1, a recording / reproducing system 1 includes a host device 2 such as a PC (Personal Computer) or AV equipment, an IDE (Integrated Drive Electronics), a SCSI (Small Computer System Interface), and an FC (Fibre Channel). And a hybrid storage device 4 connected via an interface 3 such as a USB (Universal Serial Bus).

  The hybrid storage device 4 records various data output from the host device 2 in a disk 11 as a rotary recording medium such as a hard disk and a storage area in the nonvolatile solid-state memory 22. A spindle motor 12 that rotates and drives, a voice coil motor 13 that feeds a magnetic head (not shown) that reads and writes signals to and from the disk 11 in the radial direction of the disk 11, and the spindle motor 12 and the voice coil motor. 13 has a servo control unit 14 that performs drive control, and a read / write channel unit 15 that encodes data to be written to the disk 11 and decodes data read from the disk 11.

  Further, the hybrid storage device 4 includes a hybrid storage device control unit 16 that manages data on the disk 11 and the nonvolatile solid-state memory 22, data read from the disk 11 and the nonvolatile solid-state memory 22, and the disk 11 and the nonvolatile solid-state storage. A buffer memory 17 that buffers data to be written to the memory 22; an interface control unit 18 that constitutes an input / output circuit for data to be transmitted to and received from the host device 2; a control command; and the disk 11 or the nonvolatile solid-state memory 22 A memory 19 for storing the contents on the FAT area and a CPU (Central Processing Unit) 20 for controlling them are connected via a bus 21. Further, as described above, the nonvolatile solid-state memory 22 connected to the hybrid storage device control unit 16 is also used as a disk cache.

  The hybrid storage device 4 receives a command issued from the host device 3 via the interface 3 by the interface control unit 18, and the interface control unit 18 understands the content and notifies the CPU 20 in the hybrid storage device 4. The CPU 20 sets necessary commands and parameters for the hybrid storage device control unit 16, the read / write channel unit 15, and the servo control unit 14 based on the notification contents, and executes those operations.

  The servo control unit 14 performs drive control of the spindle motor 9 and the voice coil motor 11 to move the head with respect to predetermined tracks and sectors of the disk 11. The read / write channel section 15 encodes (modulates) the transmitted user data into a digital bit sequence suitable for the characteristics of the recording / reproducing system when writing to the disk 11. Also, at the time of reading, high frequency noise is removed from the reproduction signal read from the head, then conversion from analog signal to digital signal is performed, data is estimated using maximum likelihood decoding, and then demodulation is performed. Play user data.

  The hybrid storage device control unit 16 manages data exchange among the buffer memory 17, the nonvolatile solid-state memory 22, the read / write channel unit 15, and the interface control unit 18, and performs processing related to the data format. . At that time, encoding, error detection, and error correction processing are also performed.

  Data transferred from the host device 2 is stored in the hard disk 11 or the nonvolatile solid-state memory 22. The buffer memory 17 stores data temporarily at that time or at the time of reading, and is used to suppress a decrease in performance due to a difference in transfer speed.

  Correspondingly, the host device 2 sends various commands to the hybrid storage device 4 in accordance with an instruction from a host controller by processing of a central processing unit (CPU) 32 that secures a work area in the memory 31. To do.

  In this processing, the CPU 32 uploads the boot area, FAT area, and directory area data, which are management data recorded in the system entry area of the disk 11, to the memory 31 by executing a predetermined processing procedure when the power is turned on. Various commands are sent out by setting parameters based on the uploaded data.

  Next, a data management method in this embodiment will be described. Data recorded on the disk 11 is managed in different address spaces in the host device 2 and the hybrid storage device 4. 2A is a diagram showing an LBA space P indicating an address space viewed from the host device 2 and a logical sector space Q indicating an address space viewed from the hybrid storage device 4, respectively.

  The storage area in the hybrid storage device 4 is an area combining the disk 11 and the nonvolatile solid-state memory 22. In these storage areas, the smallest unit for recording data is called a sector, and its size is usually 512 bytes. In the hybrid storage device 4, the storage area of the disk 11 is managed in units of sectors, which are a physical address, a physical sector number, and a logical sector number. The physical address is composed of a surface number, a track number, and a sector number. Here, the physical sector number is an address in which all sectors are assigned in order from the outside to the inside of the disk 11, and the logical sector number is a substitute for a defective sector that cannot be read or written. An address is assigned later. The storage area of the nonvolatile solid-state memory 22 can also be divided and managed in units corresponding to the sector size, and logical sector numbers can be assigned and managed by disabling defects in the pre-use inspection. In the nonvolatile solid-state memory 22, the address on the memory changes even when writing to the same logical sector number for the purpose of Wear Leveling performed for the purpose of evenly using the storage area. It will not be mentioned in the explanation.

  Thereby, the usable area can be managed in an integrated manner. On the other hand, the host device 2 accesses the hybrid storage device 4 using not a logical sector number but a logical block address (Logical Block Address: LBA).

  A file system for managing files uses a plurality of sectors (N) as one cluster as a minimum unit for reading and writing. The FAT (File Allocation Table) is a table that records how files are stored in the cluster. A file system that is managed using the FAT is called a FAT file system. Hereinafter, it is assumed that the cluster address is obtained by simply dividing LBA by N, and in the description, N = 16 (1 cluster = 8 Kbytes).

  FIG. 2B shows an example of FAT. For example, the hybrid storage device 4 acquires the start cluster address (12340h) of File1 from the directory information A, and recognizes that the next data of File1 is recorded at the cluster address 12341h by referring to the FAT item. Until the end of the file, the cluster address (up to EOF) can be acquired and File1 can be read. In this FAT file system, files are managed by FAT.

  As shown in the right diagram of FIG. 2A, the storage area of the hybrid storage device 4 is divided into four areas on the logical sector space Q: a user area UA, a system area SA, a hybrid storage system area, and a cache area CA. Only the system area SA and the user area UA are allocated to the LBA space P viewed from the host device 2 as shown in the left diagram of FIG. 2A. When the area used by the host apparatus 2 is 5 Gbytes from the cluster address 000000 to 9FFFFFh in the LBA space, for example, the host apparatus 2 does not use addresses after that (F00000h and later).

  The system area SA stores a master boot record MBR, an IPL (Initial Program Loader) that is a program necessary for booting (starting up) the system, and FAT. The master boot record MBR is a sector having an LBA of 0 when viewed from the host device 2, and a bootstrap code and a partition table are recorded here.

  The user area UA includes a directory area for recording directory information for managing files and a data area for recording actual data. In the FAT file system before FAT32, the root directory belongs to the system area SA. In the directory area, for each directory (file), the file name, extension, attribute, latest update time, start cluster address, file size, and the like are stored as directory information.

  The system area for the hybrid storage device is used by the CPU 20 in the hybrid storage device 4 for storing boot codes and various tables, or as an area for secondary defect replacement processing. Normally, the host device 2 is not used. For this reason, it is omitted in FIG. 2A.

  The cache area CA is used for the purpose of temporarily storing data. In the present embodiment, as shown in the right diagram of FIG. 2A, in the logical sector space viewed from the hybrid storage device 4, the area (256 MByte) from 000000h to 07FFFFh in the logical sector number corresponds to it. Therefore, unlike the LBA space viewed from the host device 2, in the logical sector space, the logical sector number of the master boot record MBR is 080000h. A cluster address is also defined for this disk cache area CA.

  In the LBA space, this cache area CA is arranged in a space after the cluster address F0000h (super cluster address F000h) not used by the host device, and the host device 2 directly writes to this disk cache area CA. There has been no such thing.

  In recent years, a hard disk drive (HDD) divides a disk into a plurality of zones (for example, 10 to 20), and zone bit recording using the same write frequency and recording / reproducing parameters (coefficients of waveform equivalent filter, etc.) in each zone. Is generally done. When zone bit recording is performed, writing (reading) only within one zone will improve efficiency without changing parameters, so the size of the disk cache area will take into account the size of the zone. It will be decided.

  Here, the logical sector space viewed from the hybrid storage device 4 in the present embodiment will be described in more detail. Normally, in a HDD to which the FAT file system is applied, when data is written, data is continuously written to the outer area of the disk where the data transfer speed is as fast as possible. You have to use a slower area. In addition, if data is repeatedly written and erased, it is difficult to secure a continuous free space, and a single file is divided into a large number of non-consecutive clusters and recorded, resulting in a significant decrease in transfer speed (usually normal Called fragmentation).

  These states become a problem when, for example, continuous shooting is performed in a high image quality mode with a digital still camera, and a large amount of data is written on the disk in a short time.

  In order to avoid such a problem, a cache area is provided on the outer peripheral side of the disk having the fastest data transfer speed, and when the host device writes to the cache area, only the start cluster address is received from the HDD, and the subsequent FAT If the HDD is managed on the HDD side, the area on the outer peripheral side can be used efficiently even in HDDs with low performance such as consumer devices, and the transfer speed can be increased.

  Also, the transfer rate is lowered by the replacement process of defective sectors occurring in the cache area on the disk. On the other hand, a replacement area is provided on the disk adjacent to the cache area. When data is written to the cache area, the replacement process is not performed. After the data written in the cache area is moved to the user area, By performing the replacement process of the defective sector by the ripping / replacement method, it is possible to reduce the deterioration of the access performance due to the replacement process.

  By the way, even when the data transfer speed is improved by these methods, when the HDD is newly started up or when it is shifted from the power saving mode to the normal mode, data transfer becomes possible. Takes a few seconds. In addition, since the HDD requires a time required for a seek operation in units of 10 ms, writing or reading cannot be executed immediately. Therefore, in order to further improve the transfer speed, in this embodiment, the head is moved to a predetermined area (empty area) of the cache area of the disk 11 and data can be written, that is, the data is stored. In the preparation period necessary for a seek operation until writing, data is written into the cache area of the nonvolatile solid-state memory, so that data transfer is efficiently performed even in the preparation period.

  As described above, the hybrid storage device 4 according to the present embodiment includes the nonvolatile solid-state memory 22 that is used as a cache in addition to the disk 11 when data is transferred from the host device. The nonvolatile solid-state memory 22 has the same input / output performance regardless of whether it is a random access or a sequential access. However, in the disk 11, the data read / write performance is deteriorated in the case of random access. Further, sequential writing to the nonvolatile solid-state memory 22 is slower than sequential writing to the disk. Therefore, in consideration of these characteristics, in the logical sector space, the addresses of the nonvolatile solid-state memory 22 are arranged at the beginning of the cache area, the beginning of the system area, and the end of the user area. That is, the addressing of the nonvolatile solid-state memory 22 in the logical sector space in the present embodiment is as shown in FIG.

As shown in FIG. 3, in the logical sector space Q, among the cache areas CA constituted by 256 Mbytes from the logical sector numbers 000000h to 07FFFFh, the 16 Mbyte area CA ₁ from 000000h to 007FFFh is stored in the nonvolatile solid-state memory 22. the area CA ₂ of 240MByte from 007FFFh to 07FFFFh shall constitute at the disk 11. The system area and the user area are 5 Gbytes in total, and the nonvolatile solid-state memory 22 is configured with 16 MBytes from 080000h to 087FFFh at the beginning of the system area and 16 Mbytes at the end of the user area, for example. As described above, the disk 11 and the non-volatile solid-state memory 22 are assigned consecutive logical sector numbers, so that the data written to or read from the disk 11 and the non-volatile solid-state memory 22 can be managed, for example, as one. It can be managed in the same way as the data recorded on the disc.

Here, the cache area CA ₂ provided on the disk 11, the write speed and a large amount of data, to allow maximum stably be carried out at the transfer rate included in the hybrid storage device 4, the outermost circumferential side of the disk 11 It is preferable to provide in. Also, if there is a defect in a physical sector in the cache area CA _2, it may be provided adjacent the replacement area to prepare the alternate sector to this.

Thus, by using the outer peripheral side of the region which is a region capable of high-speed data transfer in the disk 11 as a cache area CA _2, it is possible to greatly improve the transfer rate when viewed from the host apparatus 2. That is, writes the data transferred from the host device 2 to the cache area CA ₁ and the cache area CA _2, in free time, from the cache area CA ₂ of the cache area CA ₁ and the disk 11 of the non-volatile solid-state memory 22, a nonvolatile Data can be transferred to the user area UA ₂ , which is a recording area on the inner peripheral side whose transfer speed is slower than the user area UA _{1 of the} volatile solid-state memory 22 and the cache area CA _{2 of} the disk 11, so By releasing the cache areas CA ₁ and CA ₂ and preparing for the writing of the next transfer data, it is possible to write a large amount of data at high speed.

That is, such as when that has been a large amount of data is transferred from the host device, as described above, the until ready to write to the cache area CA ₂ of at least the disc 11 is ready, the non-volatile solid-state memory 22 data is written into the cache area CA _1. Further, as will be described later, the writing speed can be further increased by writing data in parallel in the cache area CA ₁ and the cache area CA ₂ . Incidentally, if the transfer data from the host device 2 is not so large amount, the like does not necessitate much faster writing may be used only cache area CA ₂ of the disc 11.

Of the data recorded in the system area SA ₁ and the system area SA ₂ , data that needs to be read frequently, such as MBR, IPL, and FAT, is stored in the nonvolatile solid-state memory 22 that allows high-speed access to the data. By doing so, the reading speed can be increased.

Similarly, of the data to be written in the user area UA ₁ and the user area UA ₂ , only the head data such as the head cluster of the file or the super cluster can be stored in the user area UA ₁ provided in the nonvolatile solid-state memory 22. For example, in the hybrid storage device 4, when data is read in units of files or super clusters, the head is moved or the disk is awaited for rotation while the head data is being read from the nonvolatile solid-state memory 22. And the reading speed can be improved.

Next, an operation method of the hybrid storage device 4 when data is written using the cache area CA ₁ and the cache area CA ₂ will be specifically described with reference to FIGS. Normally, management of address information for managing files on the disk 11 is performed by the host device 2 (hereinafter referred to as PC mode), but in the present embodiment, the hybrid storage device 4 permits the host device 2 to be permitted. When it is obtained, an operation mode capable of executing FAT and its management is provided. Hereinafter, this operation mode is referred to as an AV mode. If the hybrid storage device 4 is in the AV mode, the hybrid storage device 4 itself can perform FAT and its management. Here, FAT and its management in the AV mode in the hybrid storage device 4 are the hybrid data shown in FIG. 1 for the FAT area data recorded on the system area SA (nonvolatile solid-state memory 22 in the present embodiment). The data is read into the memory 19 in the storage device 4 and updated on the memory 19, and the FAT on the system area SA is updated as necessary, and updated on the memory 19 when shifting to the PC mode. By writing back the FAT to the FAT area of the system area SA, the hybrid storage device 4 exclusively shares the FAT and its management with the host apparatus 2.

  That is, in order to manage the address information for managing the file, it is necessary to manage the FAT on the system area SA and its management, and the directory information and its management. By switching, the host device 2 and the hybrid storage device 4 are exclusively shared. In the latter case, the hybrid storage device 4 is not managed, and a dedicated command for allowing the host device 2 to directly access and manage the system area SA at any time, that is, in the PC mode or the AV mode. Shall be set.

  In the present embodiment, data writing using the cache area CA is performed in the AV mode in which management of address information for managing files is performed in the hybrid storage device 4. As described above, the recording / reproducing system 1 shares the FAT information between the host device 2 and the hybrid storage device 4, but leaves the FAT and its management to the hybrid storage device 4. Accordingly, in order to write data, the AV mode is first shifted to. In order to shift to the AV mode, as shown in FIG. 4, the host apparatus 2 updates the FAT existing in the system area SA of the hybrid storage apparatus 4, that is, the FAT area FA shown in FIG. A command for shifting from the PC mode to the AV mode is issued to the storage device 4 (step S1). The CPU 20 in the hybrid storage device 4 knows the contents via the interface control unit 18, reads the FAT on the system area SA on the memory 19 (step S2), sets a flag indicating the AV mode, and sets the host The apparatus 2 is informed of the transition to the AV mode (step S3).

  Next, a case where the host apparatus 2 performs writing after shifting to the AV mode will be described. Here, a case will be described as an example where an operation for requesting writing of the content data is performed from the host device 2 when images are continuously shot by a digital camera. The host device 2 determines the file name (File1, File2) of the content data, creates a new directory item, and writes the directory item to the hybrid storage device 4 (step S4). The directory item at this time is incomplete because the start cluster address and the latest update time are unknown. The hybrid storage device 4 searches for a free cluster from its own FAT, selects an appropriate cluster from the free clusters, and performs writing. After completion, the host apparatus 2 is notified of the start cluster address shown in the directory information A in FIG. 2A.

  Next, the host device 2 sets parameters necessary for writing content data. Specifically, the start cluster address of the file and the access size of the file are determined. The host device 2 entrusts the cluster address management to the hybrid storage device 4, but it is at least necessary to share the start cluster address in order to identify the file. FIG. 5A and FIG. 5B are diagrams showing an example of a command (Set Write File Parameter) for executing the command, and show the register contents at the time of command issuance and the register contents at the end of the command, respectively. FIG. The command is defined as an AT vendor unique command and is executed only once before writing a file. The hybrid storage device 4 that has executed the command informs the host device 2 of the cluster address to start the writing.

  When the cache area CA is used, the hybrid storage device 4 is notified that the cache area CA may be used on a file basis. Therefore, the command shown in FIG. 5A is set to 1 in the CP (Cache Position) bit and the command is executed (steps S6 and S7). At this time, the hybrid storage device 4 searches the FAT in the memory 19 for a free cluster from the head of the user area UA, and notifies the host device 2 of the found cluster address as a recording start cluster address. In the case of File 1 in FIG. 2A, the recording start cluster address is 12340h.

  At the same time, an empty cluster is searched in the cache area CA. The hybrid storage device 4 records the selected free cluster address as the head cluster address (super cluster address) together with the recording start cluster address in the user area UA notified to the host device 2.

In FIG. 3, the recording start address of File 1 is F00000h for LBA, 000000h for logical sector number, F0000h for cluster address (cluster number), and F000h for super cluster address (super cluster number). Similarly, the recording start address of File 2 is F00500h for LBA, 000500h for logical sector number, F0050h for cluster address, and F005h for super cluster address. That is, the first cluster of File 1 and File 2 is the disk cache area CA ₁ on the nonvolatile solid-state memory 22.

Data sent from the host device 2 is first stored in the buffer memory 17 shown in FIG. Thereafter, writing is started in an access size unit, in this embodiment, in a super cluster unit in an empty cluster in the cache area CA ₁ of the nonvolatile solid-state memory 22 selected as described above. It starts moving the head relative to the same time writing position of the cache area CA ₂ of the disc 11. Thus, writing to the nonvolatile solid-state memory 22 is continued until the head reaches the writing position.

Head when it reaches the writing position, to start writing to the cache area CA ₂ of the disc 11. The hybrid storage device control unit 16 formats the data on the buffer memory 17 and sends it to the read / write channel unit 15 to continuously write the data on the disk 11 in units of super clusters, which is the access size in the present embodiment.

Then, it is determined whether or not the written super cluster is the final cluster (EOF: end of file) of the file (step S8), and data writing is continued until the final cluster. That is, if it is not the last cluster, the FAT in the memory 19 is updated using the cluster address of the cache area CA mapped on the LBA space every time an access unit, ie, writing of one cluster is completed ( Step S9) and return to step S8. At this time, if the CPU 20 of the hybrid storage device 4 or the hybrid storage device control unit 16 has sufficient capacity, compression is performed to reduce the amount of writing to the cache area CA ₂ of the disk 11 and to improve the transfer rate as viewed from the host device 2. You may plan. On the other hand, when finished all the writing to the cache area CA ₂ of the disk 11 of the file, the host device 2, for the latest update time and start cluster address to update the directory entry on the system area SA. Thereafter, the hybrid storage device 4 is instructed to write the FAT on the memory 19 onto the disk 11 (step S10), and the writing to the disk cache area CA is terminated.

FIG. 6 is a schematic diagram showing an example of the disk cache area CA after the writing is performed in this way. FIG. 6 shows a case where the head can reach the planned position of the cache area CA ₂ of the disk 11 in the time for writing one super cluster to the cache area CA ₁ of the nonvolatile solid-state memory 22. That is, the movement of the writing position of the cache area CA ₂ of the head of the disk 11 is completed while writing the first super cluster of File1 super cluster address (super cluster number) F000h cache area CA ₁ of the nonvolatile memory 22 writes the second to fourth super cluster of File1 super cluster address of the cache area CA ₂ F090h, ···, the F092h, while writing the first super cluster Faile2 the cache area CA _1, movement of the head is completed and, it shows super cluster address F093h of the cache area CA _2, ···, an example of writing the second to sixth super clusters of File2 to F097h. When writing a plurality of superclusters into the nonvolatile solid-state memory 22, the writing speed of the nonvolatile solid-state memory 22 does not change even when accessed randomly, so there is no need to write a plurality of superclusters continuously. .

  At this time, writing to the recording start cluster address in the user area is not performed, but a used flag is set on the FAT.

  Thereafter, it is determined whether a change request from the AV mode to the PC mode has been made and whether or not a predetermined time has elapsed (steps S11 and S12), and the host device 2 is instructed to perform a FLUSH operation or mode change. If the writing or reading has not been performed for a preset time or longer, the data in the cache area CA is copied to a normal area, that is, for example, the area C in the user area UA shown in FIG. 2 (step S13). Since the start cluster address has already been notified to the host device 2, recording is started from the start class address notified to the host device 2. Thereafter, the FAT in the memory 19 is referred to, and as soon as an empty cluster is searched from the head of the user area UA and found, data is written into the empty cluster. In the case of FIG. 2, an example is shown in which the cluster address 12340h is the recording start address and File1 is recorded in the clusters from address 12341h to 1237Fh. If compression is being performed, decompression processing is performed and writing is performed after returning to the original state.

  When copying is completed, the FAT of the FAT area FA in the system area SA is finally updated at that time (step S14). Then, if necessary, the mode is changed from the AV mode to the PC mode (step S15). If the host instructs to shift from the AV mode to the PC mode during copying, the processing is completed. Notifying the host device 2 that the operation mode has not been changed, the operation mode switching is stopped.

As described above, in the present embodiment, when the host device performs writing, only the start cluster address is received from the hybrid storage device 4 and the subsequent FAT management is performed by the hybrid storage device 4 side. It is possible to use the outer peripheral area (cache area CA ₂ ) and use the non-volatile solid-state memory 22 as a cache during preparation for recording in the cache area CA ₂ , so that a large capacity file is extremely efficient. It is possible to write continuously well, and without providing a large buffer in the host device 2, for example, it is possible to improve the transfer speed for reducing the waiting time after continuous shooting of the digital camera.

  Next, the transfer rate to the buffer memory 17 is sufficiently high, data can be written to the nonvolatile solid-state memory 22 and the disk 11 at the same time, and data that can be applied when writing to the nonvolatile solid-state memory 22 is relatively fast. The transfer method will be described with reference to FIGS.

FIG. 7 and FIG. 8 are diagrams showing the arrangement status of File 1 composed of 10 super clusters written to the cache area CA, and the write timing at that time. First, the head is moved to the cache area CA ₂ of the disk 11 while the first super cluster is written to the cache area CA ₁ of the nonvolatile solid-state memory 22, and the second super cluster is written to the cache area CA ₂ of the disk 11. The time required for writing to the nonvolatile solid-state memory 22 (Program time) varies, and when the writing speed is the slowest, that is, when the worst value is assumed (hereinafter also referred to as the latest time), the nonvolatile solid-state memory 22 is used. Is slower than the writing speed of the disk 11. In the example shown in FIG. 8, for the sake of simplicity, the ratio of the writing speed to the cache area CA ₁ of the nonvolatile solid-state memory 22 and the writing speed to the cache area CA ₂ of the disk 11 is 1: 3. Show. Actually, the writing speed of the nonvolatile solid-state memory 22 varies depending on the size of data to be written, the difference in device (solid), the difference in writing location, the operating temperature, the number of times of writing, and the like. However, there are variations due to the difference in recording position such as the inner side or the outer side, the difference in devices (solid), the continuity of the empty area, and the like.

In FIG. 7, when the writing to the nonvolatile solid-state memory 22 is completed, that is, when the writing of the first super cluster is completed, the movement of the head to the disk 11 is completed, and the second super cluster is recorded. Shows the case of writing. In this case, when the writing of the first super cluster to the cache area CA ₁ of the nonvolatile solid-state memory 22 is completed, the super cluster (second super cluster) that is writing to the cache area CA ₂ of the disk 11 at that time. The third super cluster which is the next super cluster is selected, and the third super cluster is written to the cache area CA ₁ .

That is, carried out at the time T1 shown in FIG. 8 the writing of the first super cluster SC1 to the cache area CA ₁ of the nonvolatile solid-state memory 22 is completed, the write cache area CA ₂ of the disc 11 of the second super cluster SC2 since it is, the cache area CA ₁ of the nonvolatile solid-state memory 22, while the third start writing super cluster SC3, are writing a third super cluster of the cache area CA ₁ of the nonvolatile solid-state memory 22 , the cache area CA ₂ of the disc 11, and writes the fourth supercluster SC4~ sixth supercluster SC5.

Then, at time T2 the third write superclusters SC3 to the cache area CA ₁ of the nonvolatile solid-state memory 22 is completed, similarly, the sixth super cluster SC6 that writing to the cache area CA ₂ of the disc 11 picked the next seventh super cluster SC7 of the cache area CA ₁ of the nonvolatile solid-state memory 22 writes supercluster SC7 of the seventh, during which the cache area CA ₂ of the disc 11 eighth supercluster Write the tenth super cluster and complete the process.

As described above, when parallel writing is possible, until the write preparation to the cache area CA ₂ of the disk 11 is completed, that is, from when the data is transferred from the host device 2 until the disk 11 is accessed. Although writing to the cache area CA ₁ of the nonvolatile solid-state memory 22, by subsequent write data in parallel, the seek time of the head, it is possible to take advantage of time, such as rotational latency efficiently, the head is a predetermined After reaching the writing position, the writing speed is increased by adding the writing speed of data to the nonvolatile solid-state memory 22 to the writing speed of data to the disk 11.

  FIG. 9 is a diagram illustrating the timing of the FLUSH operation. As described above, when the FLUSH operation or mode change is instructed from the host device 2, or when writing or reading has not been performed for a preset time (IDLE), the data written in the cache area CA is stored in the user area. A process of moving to the area UA is performed.

When performing this FRUSH operation, as shown in FIG. 9, data is simultaneously read from the nonvolatile solid-state memory 22 and the cache areas CA ₁ and CA _{2 of the} disk 11 and copied to the user area UA. In the meantime, the erasing operation of the nonvolatile solid-state memory 22 is performed. That is, while reading the second 1,3,7 superclusters SC1,3,7 written to the cache area CA ₁ of the nonvolatile solid-state memory 22, a head to read data from the cache area CA ₂ of the disc 11 moves Let Then, the _second , fourth to sixth, 8-10 superclusters SC2, 4-6, 8-10 written in the cache area CA2 of the disk 11 at the time when the movement of the head is finished are read, and a predetermined area in the user area UA is read. The data in the disk cache area CA ₁ of the nonvolatile solid-state memory 22 is erased while the head is moved to the empty cluster position, and 10 super clusters are sequentially written in the user area of the disk 11.

Here, when data is transferred to the user area UA, for example, the _first supercluster SC1 of File1 writes to the user area UA1 on the nonvolatile solid-state memory 22 shown in FIG. 4, and the subsequent superclusters SC2 to SC10 are discs. 11 is written in the user area UA ₂ , when reading the data of the file ₁ , the head is moved to the disk 11 while the head super cluster SC _{1 is} being read from the user area UA ₁ of the nonvolatile solid-state memory 22. can be moved to the position of the second super cluster SC2 in the user area UA _2, also in the data read processing, it is possible to speed up the reading speed.

Next, a data transfer method applicable when the writing speed to the nonvolatile solid-state memory 22 is relatively slow will be described. In this case, one super cluster can be divided and recorded in the nonvolatile solid-state memory 22 and the cache areas CA ₁ and CA ₂ of the disk 11. FIGS. 10A and 10B are a schematic diagram and a logical sector space showing the usage ratio in the recording area of the nonvolatile solid-state memory 22 and the disk 11 in one super cluster in that case.

As shown in FIG. 10A, the leading portion of the super cluster, here, 32 KB, which is 1/4 of the super cluster data, is written into the cache area CA ₁ of the nonvolatile solid-state memory 22 and the remaining portion 3 / 4 shows an example in which 96 KB, which is data No. 4, is recorded in the cache area CA ₂ of the disk 11. Therefore, unlike FIG. 4, in the logical sector number space Q, as shown in FIG. 10B, the nonvolatile solid-state memory 22 is divided and arranged separately. How much data of the super cluster is allocated to the nonvolatile solid-state memory 22 may be determined according to the standard transfer rate and the transfer rate of the hybrid storage device 4.

Even in this case, when the FRUSH operation is performed as described above, the nonvolatile solid-state memory 22 starts to access the nonvolatile solid-state memory 22 and the cache area CA of the disk 11 at the same time and moves data to the user area UA. Can be erased. In addition, when data is written to the user area UA, some clusters arranged at the head of the super clusters are moved to the user area UA ₁ on the nonvolatile solid-state memory 11, and the subsequent data is transferred to the user on the disk 11. by keeping moved to the area UA _2, the data read speed is faster.

In the present embodiment, when data transferred from the host device 2 is temporarily written into the cache area CA, the nonvolatile memory is ready for writing data to the disk 11 such as a seek operation or waiting for the disk 11 to rotate. By writing the data of the head super cluster or the head part of each super cluster to the cache area CA ₁ of the volatile solid-state memory 22, the overhead associated with the seek operation or the like can be reduced. In addition, after the seek operation of the disk 11 is completed, if data is written in parallel to the nonvolatile solid-state memory 22 and the cache areas CA ₁ and CA _{2 of the} disk 11 (parallel operation), the data at the time of writing Transfer rate can be further improved.

Further, for example, in a nonvolatile solid-state memory such as a NAND flash memory, an erase operation is required before writing, but when data is moved from the cache area CA to the user area UA, the cache area in the nonvolatile solid-state memory 22 is Since data erasing of CA ₁ is executed, it is not necessary to perform an erasing operation before writing to the cache area CA ₁ of the nonvolatile solid-state memory 22.

Further, when data is moved from the cache area CA to the user area UA, the top supercluster of the file, the data at the top of the supercluster, etc. are moved to the nonvolatile solid-state memory 22 and the remaining data are moved to the disk 11. Thus, the read process can be used efficiently even when preparations for reading data from the disk 11 are completed, such as a seek operation and waiting for the disk 11 to rotate. In addition, when data is written to the user area UA, if data is written in parallel to the user areas UA ₁ and UA _{2 of} the nonvolatile solid-state memory 22 and the disk 11, data can be moved to the user area at high speed. Can be

  Next, the effect of this embodiment will be described in more detail. In the NAND flash memory, writing (program) and reading are performed in units of pages, and erasing is performed in units of blocks. Originally, the block size was 16 Kbytes, and the page size was 528 bytes (data area 512 bytes + error correction code 16 bytes, actual data size 0.5 Kbyte). However, in recent years, the capacity of the memory has increased, and a new 128 Kbyte block size, 2112 bytes ( The page size (data area 512 bytes + error correction code 16 bytes) x 4, actual data size 2 Kbyte) is used in some memories with a capacity of 1 Gbit or more. For this reason, the newer one is called Large Block, Large Page, and the old one is called Small Block, Small Page. You cannot mix new and old ones like Large Block and Small Page.

  11A and 11B are diagrams summarizing the operation time when a NAND flash memory is used as a nonvolatile solid-state memory, and parameters using two different blocks and page sizes shown in FIG. 11A are used. FIG. 11B shows the calculated data transfer rate. From FIG. 11B, it can be seen that Large Brock with a larger access unit has improved operation speed.

  That is, the average erase time of the Large Block and Small Block is 2000 μs and 2000 μs, respectively, and the erase time (latest time) when the conditions are worst is 4000 μs and 10000 μs, respectively. Then, the access time of the Large Block, that is, the sum of the program time for one page and the time for transferring data for one page between the memory cell and the register, the average time is 305.9 μs and 130.9 μs, respectively, the latest time Are 805.9 μs and 130.9 μs, respectively. The sum of the program time for one page of the small block and the time for transferring data for one page between the memory cell and the register is 226.7 μs and 51.7 μs for the average time, and 1026.7 μs for the latest time, respectively. 51.7 μs.

  As shown in FIG. 11B, the operation time when such a NAND flash memory is used takes 7.26 ms (rounded up to the third decimal place) for 16 KByte writing (in the case of a Small Page). On the other hand, in the case of a hard disk, if the number of sectors per track is 200 and the rotational speed is 3600 rpm, it takes 2.67 ms to write 16 Kbytes (rounded up to the third decimal place).

  In the case of FIG. 9, since it takes an average of 58.1 ms (in the case of Small Page) to write 1 super cluster 128 Kbytes to the NAND flash memory, 3 super clusters (64 ms) are written to the hard disk using this time. be able to.

  When the data transfer speed is the slowest, the writing speed to the NAND flash memory takes 263 ms (in the case of Small Page) to write 1 super cluster 128 Kbytes, so 13 super clusters (278 ms) can be written to the hard disk. . When the writing to the NAND flash memory is completed, if the super cluster written to the hard disk is recognized and the next super cluster is written to the NAND flash memory, the overall transfer speed is the same as the hard disk drive transfer speed and the NAND flash memory. It becomes almost equal to the sum of the transfer rate to the memory.

As shown in FIG. 10, in one super cluster, the average time when 32 Kbytes are written to the NAND flash memory and 96 Kbytes are written to the hard disk using Small Page is as follows:
2 × 7.26 ms = 14.52 ms <16 ms = 6 × 2.67 ms
Thus, almost complete parallel writing is realized.

  When data is accessed in units of super clusters, data can be output with little delay for the top super cluster by writing the top of the super cluster in the nonvolatile solid-state memory 22. . Meanwhile, the hybrid storage device performs a seek operation on the disk 11, moves the head to a position where the remaining data (super cluster) exists, reads the data, and transfers the data to the host device 2. It is possible to efficiently utilize the seek operation that has been performed and to increase the data reading speed.

When the writing speed to the nonvolatile solid-state memory 22 is low, control is performed to reduce the allocation (occupation ratio) of data to be written to the nonvolatile solid-state memory 22. For example, when the super cluster is increased from 128 Kbytes to 256 Kbytes, and the allocation to the nonvolatile solid-state memory is set to 16 Kbytes, the allocation of 32 Kbytes is reduced to 16 Kbytes.
32.9 ms <40 ms = 15 × 2.67 ms
Thus, the data transfer speed is slightly reduced, but similarly efficient parallel writing can be realized.

It is a block diagram which shows the recording / reproducing system in embodiment of this invention. The left diagram and the right diagram are respectively an LBA space P indicating an address space viewed from the host device 2 and a logical sector space Q indicating an address space viewed from the hybrid storage device 4. It is a figure which shows an example of FAT. It is a schematic diagram which shows the addressing of the non-volatile solid-state memory 22 in the logical sector space in embodiment of this invention. 6 is a flowchart showing an operation method of the hybrid storage device 4 in the embodiment of the present invention when data is written using a cache area CA. (A) And (b) is a figure which shows one example (Set Write File Parameter) of the command which performs it, Comprising: It is a figure which shows the register content at the time of command issuing, and the register content at the time of command completion, respectively. It is a schematic diagram which shows an example of the disk cache area | region CA after writing was performed. It is a figure which shows the arrangement | positioning condition of File1 which consists of 10 super clusters written to cache area | region CA. It is a figure which shows the write timing at the time of writing File1 which consists of 10 super clusters to cache area | region CA. It is a figure which shows the timing of FLUSH operation | movement. (A) And (b) is the figure which shows the usage rate in the recording area of the non-volatile solid-state memory 22 and the disk 11 in 1 super cluster in that case, and a figure which shows a logical sector space. It is a figure which shows the parameter of two different blocks and page sizes in NAND flash memory. It is a figure which shows the data transfer rate calculated using the case where the parameter shown to FIG. 11A is used.

Explanation of symbols

  2 host device, 1 recording / reproducing system, 3 interface, 4 hard disk device, 11 disk, 12 spindle motor, 13 voice coil motor, 14 servo control unit, 15 read / write channel unit, 16 hybrid storage device control unit, 17 buffer memory, 18 Interface control unit, 19, 21 memory, 20, 32 CPU, 21 bus, 22 Non-volatile solid-state memory

Claims (17)

A rotary recording medium having a first area in which data transferred from the host device is temporarily written;
A nonvolatile solid-state memory having a second area in which data transferred from the host device is temporarily written;
Control means for controlling to write the data transferred from the host device to the recording area of the rotating recording medium and the nonvolatile solid-state memory,
The recording area includes a first writing area composed of the first area and the second area in which data transferred from the host device is temporarily written, and data from the first area in the rotary recording medium. And a second writing area including a third area where the transfer speed is slow,
Of the data transferred from the host device , the control means writes the data of the head portion into the second area, and the data other than the head portion is parallel to the first area and the second area. controlled to write to, the data written in the first write area at a predetermined timing is moved to the second write area including the third region to open the first write area The data storage device is controlled and the predetermined timing is when a FLUSH operation or mode change is instructed from the host device or when an access request from the host device is not longer than a predetermined time . The first area of the rotational recording medium is provided on the outer peripheral side, and the third area is provided on the inner peripheral side.
It said control means, while necessary for the seek operation of the rotary recording medium, the data storage device according to claim 1 that controls so as to write the data transferred from the host device to the second region. Said second write area, the data storage device according to the non-volatile solid-state the fourth area other than the second region of memory including claim 1. The control means, when moving the data, the data storage device according to claim 3 you erase the data written in the second region. The control means, when moving the data, the data of the beginning portion is written to the fourth region, according to claim 3 data other than the head portion that controls so as at least to write to the third region Data storage device. The control means, when moving the data, the data other than the head portion data storage device according to claim 5 that controls to write parallel to the third region and the fourth region. A memory for storing a table for managing data transferred from the host device as a file;
The table is recorded in a predetermined area of the second writing area,
The control means, in response to a predetermined mode switching instruction from the host device, to the memory read the table recorded in the predetermined area, by referring to the read table, the rotary recording medium and a non-volatile data storage device according to claim 3 that searches an empty region of the solid-state memory. The file is composed of a plurality of super clusters, and the super cluster is an integral multiple of a cluster which is an access unit of the host device,
Said control means is responsive to the write speed of the rotary recording medium and a non-volatile solid-state memory, the data from the host apparatus, and controls to write to the rotary recording medium and a nonvolatile solid-state memory in the supercluster units or, data storage device according to claim 7 that have a switching function for switching or controlling to write on the rotary recording medium and a nonvolatile solid-state memory in the cluster unit. The control means, compresses the data written to the first area, the data storage device according to claim 1 that restore the compressed data before moving the data to the third region. Control means receiving, from the host device to the first write area of a second region to a predetermined one of the first region and the non-volatile solid storage area of the memory to a predetermined one of the recording area of the rotary recording medium and when no temporarily write data is, among the data transferred from the host device, the data of the beginning portion is written to the second region, data other than the top portion of the first region and A writing step of writing to the second region in parallel ;
The control means includes at least a third area in which data written to the first writing area at a predetermined timing is slower than the first area in the recording area of the rotating recording medium. is moved to the second write area possess a moving step of opening the first write area,
The predetermined timing is a data storage method in which a FLUSH operation or mode change is instructed from the host device or an access request from the host device has not been set for a predetermined time or more . The first area of the rotational recording medium is provided on the outer peripheral side, and the third area is provided on the inner peripheral side.
In the write process, said control means, while necessary for the seek operation of the rotary recording medium, a data storage method according to write non claim 10 write data transferred to said second region from said host device . Said second write area, the data storage method according to the fourth region other than the second area of the nonvolatile solid-state memory to including claim 10. Above moving step, said control means, when moving the data, the data storage method according to claim 10 you erase the data written in the second region. Above moving step, said control means, when moving the data, the data of the beginning portion is written to the fourth region, the first data other than the portion written to at least the third region write non claim 10 The data storage method described in 1. In the moving process, said control means, when moving the data, the data storage method according to parallel write No claim 14 to data other than the top portion of the third region and the fourth region . It said control means, said recorded in a predetermined region of the second write area, a table for managing data transferred from the host device in a file, according to a predetermined mode switching instruction from the host apparatus Reading process to read into memory,
It said control means, by referring to the table read into the memory, data storage method according to claim 12 and a step that searches an empty region of the rotating recording medium and a non-volatile solid-state memory. In a recording / reproducing system having a host device and a data storage device that accesses a rotating recording medium according to the control of the host device and writes data and / or reads data,
The data storage device
A rotary recording medium having a first area in which data transferred from the host device is temporarily written;
A nonvolatile solid-state memory having a second area in which data transferred from the host device is temporarily written;
Control means for controlling the data transferred from the host device to be written in a recording area of the rotating recording medium and the nonvolatile solid-state memory,
The recording area includes a first writing area composed of the first area and the second area in which data transferred from the host device is temporarily written, and data from the first area in the rotary recording medium. And a second writing area including a third area where the transfer speed is slow,
Of the data transferred from the host device , the control means writes the data of the head portion into the second area, and the data other than the head portion is parallel to the first area and the second area. controlled to write to, the data written in the first write area at a predetermined timing is moved to the second write area including the third region to open the first write area The recording / reproducing system in which the predetermined timing is controlled when a FLUSH operation or a mode change is instructed from the host apparatus or when an access request from the host apparatus has not been set for a predetermined time or more .

Images