JP2010134788A - Cluster storage device, and method of controlling same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a cluster storage system capable of allowing for different disk capacities and maintaining appropriate redundancy for the number of active disks. <P>SOLUTION: A cluster storage comprising an information recording controller and three or more storage devices is provided. When storing data, the information recording controller calculates data difference sequentially, and distributes and stores data to a plurality of storage devices. When outputting data, the information recording controller collects data differences from a plurality of storage devices to restore and output the data. Thus, different disk capacities are allowed, and redundancy in accordance with the number of active disks is secured. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a thruster storage apparatus and a method for controlling the same, and, for example, relates to a technique for ensuring redundancy in a plurality of storage apparatuses.

  In recent years, digital data has increased and HDD capacity has increased. The increase in digital data increases the importance of data protection, while the increase in HDD capacity increases the risk of loss at the time of failure, while enabling increasing digital data to be secured.

  In addition, terabyte-unit storage is now available at home, and it has become common to apply digital data protection that was previously used for large-scale servers for the purpose of facilitating maintenance. . A representative example of this digital data protection method is RAID (Redundant Arrays of Inexpensive Disks), which is disclosed in the following non-patent literature.

  RAID has variations from 0 to 6 depending on the redundancy and the disk capacity utilization rate. Of these, a system is generally often constructed by any one or a combination of 0, 1, and 5. Hereinafter, outlines of 0, 1, and 5 will be described.

  RAID0 is also called striping and has no redundancy, but the disk capacity utilization rate is 100%. The principle of RAID 0 is a mechanism in which data is distributed and stored in two or more disks of the same capacity, and the controller combines and restores the data in the read state when reading. As a result, the read / write bandwidth of the disk is increased in proportion to the number of disks.

  RAID 1 is also called mirroring, and has resistance against the failure of the number of disks connected to the system minus one. On the other hand, the utilization rate of the disk capacity is 1 / (number of disks in the system) × 100 [%], which is inversely proportional to the number of disks, and the lowest utilization rate among RAID variations. The principle of RAID 1 is a mechanism for storing the same data on two or more disks of the same capacity. Depending on the implementation, it is also possible to increase the read bandwidth in proportion to the number of disks by reading data piecewise from a plurality of disks and restoring it with a controller, as in RAID 0.

  RAID 5 is parity distribution recording in units of blocks, and has tolerance against failure of one disk, and the capacity utilization rate is (number of system disks −1) / (number of system disks) × 100 [%]. is there. The principle of RAID5 is a mechanism in which data is divided into blocks, error correction codes (parity) for each block are calculated, and distributed and stored on three or more disks of the same capacity. When one disk fails, the original data can be restored using the parity. Since the data is stored in a distributed manner, the read is as fast as RAID 0, but at the time of writing, there is a latency for parity calculation.

A case for redundant arrays of inexpensive disks (RAID). (SIGMOD ´88.)

  RAID is a system based on the premise that all disks have the same capacity in any variation. That is, it is desirable to prepare a disk having the same capacity as the capacity used in the system configuration when a disk failure occurs in RAID. This is because the file is managed by LBA (Logical Block Addressing), so even if a large capacity disk is used, the capacity of the area used depends on the small capacity disk, and the unused area is wasted.

  However, due to the recent rapid increase in disk capacity, it is often difficult to acquire a disk having the same capacity as that in the system configuration when a disk fails or disadvantageous in terms of the bit unit price. Due to this problem, when a disk failure occurs, the entire system is reconstructed or a part of a large capacity disk is unusable and replaced with a small capacity disk, resulting in a time loss or a monetary loss. In other words, if you consider using a hot-selling disk or an inexpensive disk (a disk with a different disk capacity from the initial one) later, the disk capacity will vary. It is desired to operate as RAID by making full use of.

  Also, due to the nature of RAID, it is not possible to ensure redundancy in a state where a disk has failed and the number of disks has decreased. This disadvantage is fatal in a home environment where the failure rate of a disk is higher than that of a data center or the like and it takes time to recover. That is, it means that RAID is not suitable for home systems.

  The present invention has been made in view of such a situation, and provides a technique for ensuring data redundancy even in the event of a disk failure while allowing disks with different disk capacities in a storage system.

  In order to solve the above-described problem, the present invention stores three or more storage devices and data provided from an external device in any of the three or more storage devices, and the three or more storage devices. In a cluster storage apparatus including an information recording / reproduction control apparatus that controls access to data stored in any of the above, an access interpreter executes data storage and access in units of objects. More specifically, the access interpreting means assigns an ID to the object unit data, manages it, receives an access request for data stored in three or more storage devices, and assigns an ID corresponding to the data to be accessed. Acquire and implement data access.

  In the information recording / reproducing control apparatus, when the difference information generation unit receives the update data generated by updating the original data stored in any of the three or more storage apparatuses, the update data and the original data And difference information is generated. Then, the access interpretation unit assigns the same ID as the original data to the difference information and stores it in any of the three or more storage devices.

  In the information recording / reproducing control apparatus, the original data search means further searches the original data based on the ID corresponding to the data to be accessed, and the difference information search means searches for difference information based on the ID. . Then, the restoration means restores the update data for each object using the original data and the difference information obtained by the search.

  In addition, the access interpreting means stores data in units of objects to which IDs are assigned in at least two of the three or more storage devices, thereby realizing a redundancy of 2 or more.

  Furthermore, the access interpreting means aggregates and stores one original data and all the difference information related thereto in at least one of the three or more storage devices. In this case, specifically, the original flag assigning means assigns an original flag indicating original data to the original data that is aggregated and stored in one storage device, and the access interpreting means The given original data and all the difference information related thereto are collected and stored in one storage device. Regarding the operation at the time of access, when the access interpreting means receives the access request, the access interpreting means acquires the ID of the corresponding object unit data, and the original data searching means obtains the original based on the ID corresponding to the access request data. The original data is preferentially searched from the storage device that stores the original data having the flag, and the difference information search means, based on this ID, from the storage device that stores the original data having the original flag, The difference information is searched, and the restoring means restores the data in object units using all the difference information related to the original data having the original flag.

  According to another aspect of the present invention, three or more storage devices and data provided from an external device are stored in any of the three or more storage devices, and any of the three or more storage devices. In a cluster storage device comprising an information recording / playback control device for controlling access to data stored in the storage device, the failure detection means detects a failure of the storage device, and the inspection means detects the failure. The data stored in an operable storage device other than the above is searched to check whether a predetermined redundancy is ensured, and when the data storage means does not ensure the predetermined redundancy, In order to ensure the degree of storage, the deficient data is stored in the operable storage device.

  According to another aspect of the present invention, three or more storage devices and data provided from an external device are stored in any of the three or more storage devices, and the three or more storage devices In a cluster storage device comprising an information recording / playback control device that controls access to data stored in any of the storage devices, the failure detection means detects a failure of the storage device, and the inspection means detects the failure. Searches data stored in an operable storage device other than the storage device, checks whether a predetermined redundancy is ensured, and operates when the data reduction means does not ensure the predetermined redundancy A part of the data stored in the possible storage device is deleted, and the predetermined redundancy is lowered.

  Further features of the present invention will become apparent from the best mode for carrying out the present invention and the accompanying drawings.

  According to the cluster storage apparatus of the present invention, different disk capacities can be allowed, and redundancy according to the number of operating disks can be ensured. Since the use of disks with different capacities is allowed in this way, the cost can be lowest in terms of the unit price per bit when the system is repaired, and disk repair is not required as much as possible, and therefore an inexpensive system can be provided.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be noted that this embodiment is merely an example for realizing the present invention, and does not limit the technical scope of the present invention. In each drawing, the same reference numerals are assigned to common components.

1) First Embodiment The first embodiment relates to a basic system outline and redundancy. One of the important features of the present invention is that data is recognized as one “significant unit (object)”, thereby realizing a mixture of disks having different capacities that cannot be realized by the conventional RAID system. That is, in RAID, data is divided into unit capacities (blocks / bits) in the LBA space to provide redundancy, whereas in the first embodiment, data units are recognized as objects using means such as a database, and difference information Is also recognized as an object, and redundancy is ensured by multiplexing each.

<Configuration of cluster storage system>
FIG. 1 is a diagram showing a schematic configuration of a cluster storage system 1 according to the first embodiment of the present invention. The cluster storage system 1 is a virtual device that virtualizes a client device 10 that creates, changes, or plays back an object, three or more storage devices 14 that store and hold objects, and the storage device 14 as one virtual storage. And a riser device 12. Here, the client device 10 and the virtualizer device are connected via the network 2.

  In the virtualizer device 12, storage means logically formed by virtualizing a plurality of storage devices 14 is called a cluster 16. In order to maintain the same redundancy before and after the disk breaks based on RAID 5, three or more disk devices (may have different capacities) are required.

  The client device 10 is a device that includes a user interface 20, a CPU 22, a memory 24, and a communication interface 26.

  The user interface 20 includes an input device 202 such as a keyboard and an output device 204 such as a display. The user interface 20 is a device for a user to input a command or output an object to the user. This object is data digitized in an appropriate format.

  The CPU 22 is a device for reading and processing each means (each realized by a program) stored in the memory 24. Hereinafter, the means is read into the CPU 22 from the memory 24 and the operation performed by the CPU 22 is simply expressed as “the CPU 22 executes”. The memory 24 is a device for storing processing means used by the device and temporarily storing information.

  Next, each means in the memory 24 mounted on the client device 10 will be described. The object creation unit 500 is a unit that creates an object by digitizing information input by the user using the input device 202 in an appropriate format.

  The object updating unit 502 is a unit that changes the state by the user changing or updating an existing object using the input device 202 and the output device 204.

  Depending on the use of the client apparatus 10, an object playback means for outputting an object to the user via the output apparatus 204 may be provided, and further, either object creation or object update or both are not provided. May be.

  The object management means 504 is means for storing or accessing an object in the cluster 16 via the communication interface 26 (manages that an object from one address to another address is an object). The object management unit 504 is a general file system such as NTFS or ext3 when the virtualizer device 12 virtualizes the storage device 14 as a block storage accessible by the logical blocking address method. When virtualized as a network-attached storage that operates with a protocol that can access objects via the network 2, the object management means 504 is also a protocol interpretation means such as NFS. Further, when virtualizing as an object-based storage device based on search or the like, the object management unit 504 may be a unit that dynamically generates a search query according to a protocol.

  The communication interface 26 is a device for transmitting commands and information to other devices connected to the network 2 or receiving commands and information from other devices connected to the network 2.

  The storage device 14 includes a communication interface 26 and storage means 28. The storage means 28 is means for storing information received via the communication interface 26 in accordance with a command. Here, the storage device 14 is a block storage that stores information in accordance with a command from another device connected to the network 2. Note that the storage device 14 may be mounted so that the storage device 14 can manage objects by incorporating the CPU 22, the memory 24, and the object management unit 504 into the storage device 14.

  The virtualizer device 12 is a device that includes a communication interface 26, a CPU 22, and a memory 24. Hereinafter, each means in the memory 24 mounted on the virtualizer device 12 will be described.

  The access interpreting means 520 is a kind of file system and has a function of interpreting data sent from the client device (host) 10 according to a communication protocol, which part is an object. In other words, the access interpretation unit 520 is a unit that converts an object storage request or an access request to the object of the client device 10 via the network 2 into information that the object management unit 504 can understand. Furthermore, the access interpreting means 520 is a means for the virtualizer device 12 to determine the access method to the cluster 16, and is one of the most important means in this embodiment.

  The object is separately managed in a base part created by the object creating unit 500 and difference information that is a change difference added to the object after the base part creation. The base part and the difference information are managed by being assigned a unique base ID within the cluster 16. That is, the base part and the difference information related thereto have the same base ID. As described above, since the object is managed by the base ID, it is possible to allow use of disk devices having different capacities without being restricted by the LBA space.

  The access interpreting means 520 is implemented as means for assigning pointers to objects in various supported access methods to the base ID space using, for example, database means. The base ID management means 522 manages the base IDs so that they do not overlap.

  In this embodiment, the objects in the cluster 16 are a base ID management unit 522, a base part search unit 524, a difference information search unit 526, a difference information generation unit 528, a difference information confirmation unit 5210, an object restoration unit 5212, and a storage storage determination unit. 5214. Hereinafter, each function will be described.

  Base ID management means 522 is means for guaranteeing the uniqueness of the base ID in the cluster 16.

  Further, the base part searching unit 524 and the difference information searching unit 526 respectively search the base unit and the difference information from the storage device 14 based on the base ID converted by the access interpretation unit 520 to access of the client device 10.

  The difference information generation unit 528 is a unit that generates a difference between two objects, while the object restoration unit 5212 restores an object from the original object and the difference information generated by the difference information generation unit 528.

  When the client device 10 changes the object using the object update unit 502 and saves it in the cluster 16, difference information is generated by the difference information generation unit 528, and is stored in the storage device 14 separately from the object. This operation is performed every time the object is updated. That is, the latest object can be restored by preparing the base part and all the difference information and then restoring the object sequentially by the object restoration means 5212. Furthermore, it is possible to trace all the object changing processes, and the state before the latest can be restored.

  The difference information confirmation unit 5210 confirms the aggregation of all the difference information with the base part (confirms whether all the difference information exists).

  The storage storage determining unit 5214 determines the storage device 14 that stores the base unit or the difference information in accordance with the policy set by the user or the system administrator. This policy is a setting for determining whether the storage device 14 is based on the capacity load of the storage means 28 or the access frequency to the storage device 14, the fault tolerance number of the storage device 14 connected to the cluster 16, That is, it represents the setting of redundancy. Further, the difference information of the object is stored in the storage device 14 that can be stored by the storage storage determination unit 5214. That is, even when storage units 28 having different capacities are mixed, the storage storage determining unit 5124 can allocate data corresponding to the capacities, and as a result, the storage device 14 having different capacities can be allowed while providing redundancy. The secured cluster storage system 1 can be provided.

  In this embodiment, the redundancy of the cluster 16 is realized as a tolerance for the failure of one storage device 14 by storing the base unit and the difference information in two storages, respectively. In order to increase the redundancy, it is only necessary to further increase the multiplicity of the base part and the difference information. The number of storage devices 14 required for this increase is redundancy + 1. When there are three storage devices, the base unit and the difference information are stored in the two storage devices as described above. However, when the number of storage devices further increases and the redundancy is increased, three units are stored. You may make it preserve | save a base part and difference information in the above storage devices, respectively.

  The above is the system configuration example of the first embodiment of the present invention. In the present invention, not limited to the configuration shown in FIG. 1, the client device 10 is directly connected to the virtualizer device 12, and the virtualizer device 12 and each storage device 14 have another communication interface 26 included in the virtualizer device 12. The storage device 14 may be connected to the storage device 14, or the virtual device 12 may be replaced with the storage device 14. Further, the client device 10 or the client device 10 and the storage device 14 share the devices and means of the virtualizer device 12 so that the client device 10 is directly connected to the storage device 14 and virtualized.

<Information storage format>
Hereinafter, the storage format of the base unit and difference information in the storage apparatus 14 will be described.

  FIG. 2 is a diagram illustrating an example of a storage format of the base unit 62. The base unit 62 is stored in the storage apparatus 14 with the base ID 70 assigned to the initial object 72. The base ID 70 is an integer value assigned to each object, and is guaranteed to be unique in the cluster by the base ID management means 522. The storage apparatus 14 recognizes the data string as one object based on this base ID. The initial object 72 is an object generated by the client device 10 using the object creation means 500.

  Note that the storage method of the base unit 62 is not limited to the above, and may be managed separately using, for example, a database unit, instead of a format in which the base ID 70 is directly assigned to the new object 72. Further, instead of directly associating with the new object 72, for example, a format in which the head and tail LBA in which the object is stored is stored together may be used.

  FIG. 3 is a diagram illustrating an example of a storage format of the difference information 64. The difference information 64 is stored in the storage device 14 in a format in which the base ID 70 and the version number 74 are added to the object difference 76. The base ID 70 is assigned in order to associate the object difference 76 detected by the difference information generating unit 528 with the base unit 62 of the same object. The version number 74 is an integer value for recording the number of times the object has been updated. The version number 74 increases from 0 in the order in which the object is updated with the base unit 62 as the starting point.

  The object difference 76 is information generated by the difference information generating unit 528 from the original object and the updated object. The object difference 76 may be a difference between the initial object 72 and the latest object, or may be a difference between the previous version and the latest object. In the former case, an area for storing an object can be saved. On the other hand, the latter can acquire the update history of the object sequentially. In the present embodiment, the latter method of sequentially acquiring the object difference 64 will be described below.

  Further, the base ID 70 or the version number 74 given to the base unit 62 and the difference information 64 may be separately managed by a database unit or the like instead of being given to the new object 72 and the object difference 76.

  At the time of object restoration, in the base ID 70 obtained by the access interpretation means 520, all the version numbers 74 existing in the cluster 16 are totaled, and it is confirmed whether all the difference information 64 up to the latest difference information is prepared. Restoration is performed.

<Operation sequence>
An example of the operation procedure of the client device 10 and the cluster 16 will be described. FIG. 4 is a diagram illustrating an example of an operation sequence of the client device 10 and the cluster 16 according to the first embodiment.

  In FIG. 4, C indicates the client device 10, and V indicates the virtualizer device 12. S1, S2, and S3 represent the storage apparatus 14, respectively. FIG. 4 shows the respective processes in time series (80, 82, 84, 86) in order from the top. Reference numeral 90 denotes a failure occurrence of the storage apparatus 14. The failure 90 in the present embodiment is defined as a state that can be recovered within a certain period of time, for example, communication disconnection or power interruption.

  Further, in FIG. 4, the procedure surrounded by a dotted rectangle is a new object creation procedure 80, an object access procedure 82 when a failure 90 occurs in the storage device 14, and an object update procedure 84 when a failure 90 occurs in the storage device 14. And an object access procedure 86 accompanied by object restoration when a failure 90 occurs in the storage apparatus 14. Hereinafter, the new object creation procedure 80 will be described in order.

  i) In the new object creation procedure 80, the client device 10 creates a new object 601 using the object creation means 500. Then, the object management unit 504 issues a save request for the generated new object 601 and transmits it to the cluster 16 via the communication interface 26. This storage request is received by the virtualizer device 12. Then, the base ID 70 is assigned to the new object by the access interpretation means 520, and the base unit 62 is generated by combining the new object 601 and the base ID 70.

  Thereafter, the virtualizer device 12 determines two storage devices 14 (here, S1 and S2) by the storage storage determination means 5214, transmits the base unit 62 to each of them, and stores them.

  ii) Next, an object access procedure 82 when a failure occurs in the storage apparatus 14 storing the base unit 62 will be described.

  The client device 10 uses the object management unit 504 to issue an access request to the object 601 requested by the user (for example, issued when the user clicks a file name), and sends it to the cluster 16 via the communication interface 26. Send. The virtualizer device 12 receives this access request and starts an object access response operation 82v.

  FIG. 5 is a flowchart for explaining in detail the object access response operation 82v by the virtualizer device 12 in response to the object access request of the client device 10. Each annotation at the end of the sentence indicates a reference process in the figure.

  First, the virtualizer device 12 specifies the base ID 70 by the access interpretation unit 520. Based on the identified base ID 70, the virtualizer device 12 searches the base unit for which an access request has been made from each storage device 14 managed by itself using the base unit search unit 524. (S822)

  If the base unit 62 cannot be found by the above operation, the access interpretation unit 520 notifies the client 10 that the object 601 does not exist, and ends the object access operation 82v (No in step S8221 → S826 → S8214).

  When the base unit 62 is found, the difference information search unit 526 searches for the difference information 64 having the same base ID 70 as the found base unit 62 from all the storage apparatuses 14 (S1 and S3) that can respond (S8221). Yes).

  Thereafter, the virtualizer device 12 confirms whether or not all the difference information 64 necessary for object restoration exists for each difference information 64 returned from the storage device 14 using the difference information confirmation unit 5210 (S824).

  If it is determined that all the difference information 64 exists and the object 601 can be restored, the object restoration unit 5212 restores the object 601 (Yes in S8241).

  In this case, since the difference information 64 does not exist and is only the base unit 62, the base ID 70 is removed from the base unit 62 using the object restoration unit 5212, and is transmitted to the client device 10 as the object 601 to access the object. The operation 82v is terminated (S828 → S8212 → S8214).

  If all the difference information necessary for restoration does not exist and the object cannot be restored in the latest state, the restoreable object is notified to the client, and the object access operation ends (No → S8210 in S8241).

  The operation 8210 when the restoration could not be performed is determined by the user policy set in the cluster 16. For example, the object access operation is terminated by notifying the object access failure 826 without performing this presentation. May be.

  iii) Next, the update procedure 84 of the object 601 when a failure occurs in the storage device 14 (S2) storing the base unit 62 will be described. The client device 10 uses the object update unit 502 to update the object 601 received from the cluster 16. Then, the object management unit 504 issues a save request for the updated object 602 to the cluster 16. The virtualizer device 12 receives the save request from the client device 10 and starts an object save operation 84v.

  FIG. 6 is a flowchart for explaining in detail the object saving operation 84v by the virtualizer device 12 in response to the object saving request of the client device 10.

  The virtualizer device 12 specifies the base ID 70 of the object 601 that is the basis of the updated object 602 by using the access interpretation unit 520, and restores the object 601 from the storage device 14 using the specified base ID 70. In an aspect having sufficient memory 24, the existing object 601 may be obtained by storing the object 601 restored at the time of object access 82 from the client device 10 (S841).

  If the existing object 601 could not be acquired, the new object creation procedure 80 is executed, the object 602 is newly saved as the new object 72, and the object saving operation 84v ends (No in S8411 → S80 → S846).

  When acquiring the existing object 601, the virtualizer device 12 generates an object difference (not shown) with the update object 602 received from the client device 10 using the object difference information generation unit 528. After the object difference is generated by the object difference information generating unit 528, the difference information 64 is obtained by adding the base ID 70 and the latest version number 74 (1 in this case) specified by the access interpreting unit 520 to the object difference (not shown). (Yes in 8411 → S842).

  Thereafter, the virtualizer device 12 selects two storage devices 14 using the storage storage determination unit 5214 (S8440).

  In the present embodiment, S1 and S3 that can be used are selected, the difference information 64 is saved for each of them, and the object saving operation 82v ends (S8441 → S8442 → S846).

  iv) Further, a description will be given of an object access procedure 86 accompanied by restoration when a failure occurs in the storage unit 14 (S1) storing the base unit 62 and the difference information 64. FIG.

  The client device 10 uses the object management unit 504 to issue an access request to the object 602 requested by the user and transmits it to the cluster 16 via the communication interface 26. The virtualizer device 12 that has received this access request specifies the base ID 70 using the access interpretation means 520. Thereafter, the virtualizer device 12 issues an access request to the base unit 62 and the difference information 64 using the base unit 62 search unit 524 and the difference information 64 search unit 526 based on the base ID 70. Then, the access request is transmitted to all the storage apparatuses 14 (S2 and S3) belonging to the cluster 16 that can respond.

  Each storage device 14 that has received an access request to the base unit 62 and the difference information 64 transmits information requested by the virtualizer device 12. The virtualizer device 12 analyzes each piece of information received from the storage device 14 using the difference information confirmation unit 5210. In the present embodiment, the latest difference information 64 is version 1, and all the difference information 64 up to version 1 can be aggregated. After the difference information confirmation unit 5210 confirms that all the difference information 64 up to the latest difference information 64 is available, the object restoration unit 5212 restores the object 602 in the latest state from the acquired difference information 64. Then, the restored object 602 is transmitted to the client device 10, and the access procedure 86 to the object accompanying the restoration is completed.

  Even if the capacity of the storage unit 28 of the cluster storage device 14 is different, the storage storage determining unit 5214 appropriately selects the storage unit for the base unit 62 or the difference information 64 so that redundancy can be achieved. It is in the point that can be kept.

  Therefore, the cluster storage 1 according to the present invention can provide a cluster storage system that allows storage devices 14 having different disk capacities to realize redundant management of objects including the history management function.

2) Second Embodiment In the second embodiment, data is collected as much as possible in one storage device in units of objects. The user can intuitively imagine that this object is in some storage. That is, when acquiring an object, there is a possibility that only one storage needs to be activated (the possibility of repeating activation and deactivation one by one as in the case of conventional storage).

  Considering such a situation, the method according to the first embodiment may not provide all the worst data.

  Therefore, a storage device that is a parent device of the object is determined, and a mechanism is provided such that the storage device (child device) for when the parent device cannot acquire the difference is provisionally received. With this operation, the user can sufficiently acquire information even in the above situation, and can intuitively use the system.

  The advantages of the distributed cluster storage device are two points: high-speed data access by distributed arrangement and easy recovery from storage failure, that is, maintainability. As for data access speed, data transfer is generally a performance bottleneck. This is because data retrieval or data restoration is faster than data transfer.

  In the present invention, the client apparatus 10 includes a difference information confirmation unit 5210 and an object restoration unit 5212 in order to realize high-speed data access using the storage apparatus 14. When the storage device 14 searches the base unit 62 or the difference information 64 and transmits it to the client device 10, the data transfer rate is determined by the network 2. Generally this is the maximum transfer rate.

  On the other hand, with respect to improving maintainability, a plurality of storage devices 14 are not required when restoring the object 60, and the storage device 14 having the object 60 can be copied / moved by making it possible to restore the object 60 by one unit. There is a method that can be realized by only one storage device 14. This method can be realized by moving one storage apparatus 14 that is aggregated even when the object 60 is physically carried, and convenience is also improved.

  In the second embodiment, a method for improving maintenance by aggregating the base unit 62 and the difference information 64 in one storage device 14 will be described.

<Configuration of cluster storage system>
FIG. 7 is a diagram showing a schematic configuration of a cluster storage system according to the second embodiment of the present invention. The difference from the first embodiment is that the virtualizer device 12 further has original management means 5216.

  In the present embodiment, “original” is defined as one special base unit 62 having a higher priority than others among the plurality of stored base units 62. Further, the storage apparatus 14 that owns the original is expressed as “the native of the object”.

  The main point of the present embodiment is to realize the restoration of the object 60 with only one storage device 14 as much as possible by aggregating the difference information 64 natively.

<Example of data structure of base part>
FIG. 8 is a diagram illustrating an example of the configuration of the base unit 62 used in the second embodiment.

  The difference from the first embodiment in the configuration of the base portion is a portion to which the original flag 78 is added in addition to the base ID 70. The original flag 78 is information for determining whether or not the base unit 62 is original, and takes either a true or false state. Similarly to the base ID 70, the original flag 78 is not directly assigned to the base unit 62, but may be managed by association with the base unit 62 by a database means or the like.

<Operation sequence>
FIG. 9 is a diagram illustrating an example of an operation sequence of the client device 10 and the cluster 16 in the second embodiment. In FIG. 9, C indicates the client device 10, and V indicates the virtualizer device 12. S1, S2, and S3 represent the storage apparatus 14, respectively. In FIG. 9 as well, the processing is shown in time series (810, 812, 814) in order from the top. Further, 90 represents the occurrence of a failure in the storage apparatus 14. Also in the present embodiment, a failure is defined as a state that can be recovered within a certain period of time, for example, communication disconnection or power supply disconnection. Further, the procedure enclosed by the dotted rectangle includes the new object creation procedure 810 with the original flag 78, the object update procedure 812 during native operation, and the object restoration and native difference information 64 when the storage device 14 fails. The access procedure 814 to the object accompanied by duplication is shown.

  i) First, a new object creation procedure 810 with the original flag 78 will be described.

  The client device 10 creates a new object 72 using the object creation means 500. Then, the object management unit 504 issues a request to save the newly created new object 72 and transmits it to the cluster 16 via the communication interface 26.

  The virtualizer device 12 receives this storage request and starts a new object storage operation.

  FIG. 10 is a flowchart for explaining in detail the new object saving operation 810v of the virtualizer device 12 in response to the new object saving request of the client device 10.

  The access interpretation unit 520 assigns the base ID 70 to the new object 601 received from the client device 10, and generates the base unit 62 by combining the new object 72 and the base ID 70 (S8100).

  Thereafter, the virtualizer device 12 determines the two storage devices 14 (S1 and S2 in this example) using the storage storage determination unit 5214 (S8240). At this time, the original management unit 5216 sets the original flag 78 to true for the base unit 62 selected as the first storage device 14 (S8421), and sets the base unit 62 selected as the second storage device 14 to the base unit 62 selected as the second storage device 14. On the other hand, the original flag 78 is set as false (S8422).

  In addition, the original management unit 5216 sets the original flag 78 other than the first storage device 14 as false in the mode of determining three or more storage devices 14 for the purpose of increasing redundancy.

  Thereafter, the base unit 62 is transmitted to each of them and stored. As a result, the new object saving operation 810v ends (S8241 → S8242 → S844).

  ii) Next, an object update procedure 812 during native operation will be described. Similar to the object update procedure 84 in the first embodiment, the client apparatus 10 updates the object 601 using the object update unit 502. In addition, the object management unit 504 issues a storage request for the updated object, and transmits the storage request to the cluster 16 via the communication interface 26.

  FIG. 11 is a flowchart for explaining in detail the object storing operation 812v of the virtualizer device 12 that has received the storing request.

  The virtualizer device 12 that has received the save request uses the base ID management means 522 to restore the object from the specified base ID 70 from the storage device 14, or in the form of having sufficient memory 24, the client device The existing object 601 is acquired by storing the restored object at the time of object access from 10 (S821).

  When the existing object 601 can be acquired, the object difference information generation unit 528 generates an object difference (not shown) between the acquired existing object 601 and the update object 602 received from the client device 10, and the access interpretation unit 520 The difference information 64 is generated by assigning the identified base ID 70 and the latest version number 74 (in this case, 1) to the object difference (Yes in S8211 → S822).

  On the other hand, if the existing object 601 could not be acquired, a new object creation procedure 810 (see FIG. 10) with the original flag 78 is executed, the updated object 602 is newly saved as a new object, and the object saving operation 812v ends ( No in S8211 → S84).

  Thereafter, the virtualizer device 12 determines (selects) two storage devices 14 using the storage storage determination unit 5214 (S8240).

  Further, the object restoration unit 5212 detects the original flag 78 when the storage device is determined (S8240). That is, when the virtualizer device 12 can access the native (S1), the storage storage determining unit 5214 preferentially selects the native as the first storage (S8250 → S8251 → S8252).

  Thereafter, the difference information 64 is transmitted to each of the selected storage devices 14 and stored. As a result, the object saving operation ends (S8241 → S8242 → S826).

  iii) Next, an object access procedure 814 that accompanies object restoration and copy of the difference information 64 when a failure occurs in the storage apparatus 14 will be described. In this example, it is assumed that the native storage apparatus does not have all the difference information 64 necessary for restoration (that is, some data for restoration is missing) due to a failure of the storage apparatus 14. is doing.

  In this state, the virtualizer device 12 replicates the difference information 64 natively when access to the object 603 occurs, and returns to a state where the object 603 can be restored only by the native storage device.

  FIG. 12 is a flowchart for explaining in detail the object access response operation 814v of the virtualizer device 12 that has received the object access request from the client device 10.

  The virtualizer device 12 that has received the object access request from the client device 10 specifies the base ID 70 of the target object using the access interpreter 520. Further, the virtualizer device 12 searches the storage unit 14 belonging to the cluster 16 for the base unit 62 that matches the base ID 70 using the base unit search means 524 based on the base ID 70 (S822).

  If there is no matching base unit 62, the virtualizer device 12 notifies the client device 10 of the object access failure. As a result, the object access response operation 814v ends (No in S8221 → S826 → S8214).

  After detecting the base unit 62, the virtualizer device 12 uses the difference information search unit 526 to search the storage device 14 connected to the network 2 for the difference information 64 necessary for restoring the object 603 (S824).

  If the difference information 64 necessary for restoration exists and the base unit 62 search means 524 confirms the existence of the original, that is, if there is a native storage device, the difference information confirmation means 5210 uses the native storage device for object restoration. It is confirmed whether all necessary differences are held (Yes in S8241 → S8250 → Yes in S8251 → S8140).

  When all the difference information 64 necessary for restoration does not exist, a restoreable object is presented and the object is restored (No in S8241 → S8210 → S828).

  If the native storage device does not hold all the differences necessary for object restoration, the virtualizer device 12 replicates the missing difference information 64 natively (No in S8141 → S8142). If the native storage device holds all the differences necessary for object restoration, the process proceeds to S828.

  Thereafter, the virtualizer device 12 restores the object using the object restoration unit 5212 and transmits it to the client device 10. As a result, the object access operation ends (S828 → S8212 → S8214).

  As described above, the operations S810, S812, and S814 maintain a state where the object can be restored only by the native storage apparatus. As a result, the duplication / movement of the object 60 can be realized by only the native storage device and the other one storage device 14, thereby improving the maintainability as a result. Furthermore, even when the object is physically carried, it can be realized by moving only the native storage device, and convenience is improved.

3) Third Embodiment The third embodiment performs redundancy maintenance by adding a new storage, and relates to a process for repairing a storage failure. In the first and second embodiments, it is assumed that the storage apparatus 14 in which the failure 90 has occurred can be restored after a certain period.

  However, there is a case where the failure 90 that cannot actually be recovered, that is, the information stored in the storage device 14 becomes unavailable due to the failure.

  In view of this, the present embodiment provides a method for ensuring redundancy for failures in the cluster storage system 1.

  As the cluster storage system, the configuration shown in FIGS. 1 and 7 can be adopted.

<Operation sequence for ensuring redundancy>
FIG. 13 is a diagram illustrating an example of an operation sequence of a redundancy restoration operation for restoring redundancy by newly adding a storage device 141 to the cluster 16 in which redundancy cannot be ensured due to a failure of the storage device 14.

  Each element drawn under the storage device 14 identification symbols (S1, S2, and S3) indicates an object base 62 and difference information 64 stored by the client device 10 (not shown). 92 indicates a failure of S3.

  Hereinafter, an example of the redundancy restoration operation for restoring the redundancy lost due to the failure of S3 by adding the new storage device 141 will be described.

  When the virtualizer device 12 detects the addition of the new storage device 141, the virtualizer device 12 starts the redundancy restoration operation. This addition detection of the new storage device 141 collects information of the storage device 14 connected to the network 2 periodically using the storage storage determination unit 5214 for the purpose of speeding up the storage selection by the storage storage determination unit 5214. It is realized by doing.

  The virtualizer device 12 searches the base part 62 and the difference information 64 by using the base part search means 524 and the difference information search means 526 for all the base IDs 70 managed by the base ID management means 522.

  As a result of the search, the redundancy is restored by storing the number of base units 62 and difference information 64 that are less than the redundancy determined by the storage storage determination unit 5214 in the new storage device 141.

  By this operation, the cluster storage system 1 can realize restoration of lost redundancy.

4) Fourth Embodiment The fourth embodiment relates to redundancy maintenance in a cluster storage apparatus configured with redundancy + two or more.

  Redundancy that can be provided by the cluster storage apparatus 14 according to the present invention is a failure of up to (the number of clusters 16-2).

  However, in order to place importance on the capacity rather than the redundancy, the cluster 16 may be configured with redundancy + two or more. For example, if one disk fails in the cluster 16 operating with redundancy +3, the disk may be restored and the redundancy may be restored according to the third embodiment, but the cluster 16 still has redundancy +2 It is configured by a single cluster storage device 14, and can be operated as it is if the redundancy is appropriately adjusted.

  In this embodiment as well, the configuration shown in FIGS. 1 and 7 can be adopted as the cluster storage system.

  In the present embodiment, a redundancy adjustment operation for appropriately adjusting redundancy information and continuing operation for a failure in a cluster 16 that was originally operated with a number of redundancy having a margin will be described.

  FIG. 14 shows one cluster storage device in an operation in which redundancy is provided for the failure of one disk on a cluster 16 composed of one virtualizer device 12 and four cluster storage devices 14. It is a figure which shows an example of the sequence of the redundancy adjustment operation | movement when 14 (S3) fails 92. FIG.

  When the virtualizer apparatus 12 detects the failure 92 of the cluster storage apparatus 14 (S3) using the storage storage determination unit 5214, it starts the redundancy adjustment operation.

  First, the virtualizer device 12 uses the base unit search unit 524 or the difference information search unit 526 to search the base unit 62 and the difference information 64 of all the data that the accessible cluster storage device 14 has, The redundancy is checked using the difference information confirmation unit 5210.

  As a result, when the detected data 66 that does not maintain redundancy is detected, the storage storage determination unit 5214 determines each cluster storage device 14 that does not have the same data, and each has redundancy. Save data that is not kept.

  With this redundancy adjustment operation, the cluster storage apparatus 14 can maintain as much redundancy as possible by the storages that make up the cluster 16. This feature is particularly effective when the mean time between failures is long. That is, it can be said that the cluster storage device 14 is suitable in a home environment where spare disks are not always prepared.

5) Fifth Embodiment The fifth embodiment relates to maintenance of appropriate object redundancy by deleting redundant information. Also in this embodiment, the configuration shown in FIGS. 1 and 7 can be adopted as the cluster storage system.

  Repairing storage at home is cumbersome and should be avoided if possible. In other words, even if the number decreases, it can be said that a system that can continue to operate as it is is valuable.

  According to the cluster storage system 1 of the present invention, the object redundancy can be changed between 1 and the number of cluster storage devices 14 constituting the cluster 16 -2. This feature causes excessive redundancy information to be retained when the number of units decreases due to a failure 92 of the storage apparatus 14 and the redundancy required by the user cannot be secured. Such excessive redundant information squeezes the storage means 28 of the storage device 14 in capacity and reduces the utilization efficiency of the cluster storage system 1.

  In view of this, the present embodiment provides a procedure for restoring the excessive redundancy due to the failure 92 of the storage apparatus 14 to an appropriate redundancy.

  FIG. 15 is a diagram illustrating an example of a redundancy return sequence of the cluster 16 that has excessive redundant information due to the failure 92 of the storage apparatus 14.

  First, when the virtualizer device 12 detects a failure 92 of the storage device 14, the virtualizer device 12 starts a redundancy return sequence. In addition, the detection of the failure 92 of the storage device 14 is performed by using the storage storage determination unit 5214, and information on the storage device 14 periodically connected to the network 2 for the purpose of speeding up storage selection by the storage storage determination unit 5214. It is realized by collecting.

  Then, the virtualizer device 12 searches the base unit 62 and the difference information 64 for all the base IDs 70 managed by the base ID management unit 522 using the base unit search unit 524 and the difference information search unit 526.

  As a result of the search, the redundancy is returned to an appropriate value by deleting the base unit 62 and the difference information 64 that exceed the redundancy determined by the storage storage determination unit 5214. At this time, the storage device 14 whose information is to be deleted is determined so that the loads of all the storage devices 14 belonging to the cluster 16 are uniform according to the policy of the storage storage determination means 5214.

  By such an operation, the redundancy of the object 60 in the cluster 16 is constantly set to an appropriate value. Therefore, the cluster storage system 1 can maintain appropriate redundancy information with an appropriate disk capacity according to the number of storage devices 14 constituting the cluster 16.

6) Summary In this embodiment, since RAID 5 is taken into consideration, a cluster storage system in which three or more storage devices are installed is employed. Then, the information recording control device (virtualizer device) has an access method conversion means (access interpretation means) for converting the access method from the external device to the data stored in the storage device into the access method into the cluster storage. I have. This conversion of the access method is to manage by assigning an identifier (ID) to the data in the object unit at the time of data storage, and at the time of data access (when an access request is made from an external device), An identifier (ID) corresponding to the object unit data is acquired, and object data corresponding to the ID is acquired. In this way, data is managed in units of objects, and IDs are assigned to stored data to realize data storage and access, so that three or more storage devices can be allowed to have different capacities. . In the conventional RAID system, data management by LBA is performed, so that disk devices having different capacities cannot be allowed. However, in the present invention, data management by object ID is performed, so that disk devices having different capacities can be used. It will be possible.

  In addition, if there is an update to the data (original data or base part) that has already been stored, difference information between the update data and the original data is generated, and the difference information is the same as the ID assigned to the original data Is assigned and stored in the storage device. In this way, even when data (original data and difference information) is distributed and stored in a plurality of storage devices, all the related data can be acquired. Moreover, the data amount of the original data and the difference data varies, and the data can be stored while taking into account the capacity of each storage device. Accordingly, various storage capacities can be allowed for each storage device.

  In another aspect, all of the original data (original data with a flag indicating that) and difference information related to data in a certain object unit and difference information are stored and managed in one storage device. As a result, the user can quickly acquire desired data using only a specific storage device. More specifically, the storage device selection unit specifies one storage device. Then, the data difference confirmation unit checks whether or not the specified storage device has all the data (difference information) necessary to restore the requested data. If the specified storage device does not have the difference information necessary to restore the requested data, the difference information that the data duplicating means does not have from the other storage device Acquire and copy information is stored in the specified storage device. In this way, all data (original data and difference information) is collected in a specific storage device.

  In another aspect, when a failed storage device is replaced with a new storage device, the new storage device is detected, data is acquired from another storage device, and the data stored in the failed storage device is stored. Port to a new storage device. However, even in this case, the data to be handled is in units of objects. When data is stored in each storage device, the same ID is assigned to the original data and the difference information (difference between the update data and the original data) and managed. It is good to have. By doing so, it is possible to accept any capacity as a new storage apparatus, and it becomes easy to ensure redundancy.

  Furthermore, in another aspect, when a storage device failure is detected, the data stored in the failed storage device is ported to another storage device. However, even in this case, the data handled is in units of objects, and when storing data in each storage device, the same ID is assigned to the original data and difference information (difference between update data and original data) and managed. It is good to be. In this way, redundancy can be ensured, and there is no problem even if the capacity of the failed storage apparatus is different from that of the other storage apparatus (not affected by the storage apparatus with the smallest capacity).

  In another aspect, when a storage device failure is detected, it is possible to detect data stored in an operable storage device other than the failed storage device and reduce the redundancy before the failure. In such a case, redundant data is deleted from the operable storage device to reduce the redundancy. However, even in this case, the data handled is in units of objects, and when storing data in each storage device, the same ID is assigned to the original data and difference information (difference between update data and original data) and managed. It is good to be. By doing so, when there is a surplus in the number of storage devices, it becomes possible to respond to an emergency without forcibly maintaining a high degree of redundancy. In particular, in a storage system used at home, priority is given to operation even when a storage device fails, and a certain amount of time is required until a new storage device is replenished. The storage system can be operated without any problem even during the period until the replenishment.

  The present invention can also be realized by a program code of software that realizes the functions of the embodiments. In this case, a storage medium in which the program code is recorded is provided to the system or apparatus, and the computer (or CPU or MPU) of the system or apparatus reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the program code itself and the storage medium storing the program code constitute the present invention. As a storage medium for supplying such program code, for example, a flexible disk, CD-ROM, DVD-ROM, hard disk, optical disk, magneto-optical disk, CD-R, magnetic tape, nonvolatile memory card, ROM Etc. are used.

  Also, based on the instruction of the program code, an OS (operating system) running on the computer performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing. May be. Further, after the program code read from the storage medium is written in the memory on the computer, the computer CPU or the like performs part or all of the actual processing based on the instruction of the program code. Thus, the functions of the above-described embodiments may be realized.

  Also, by distributing the program code of the software that realizes the functions of the embodiment via a network, the program code is stored in a storage means such as a hard disk or memory of a system or apparatus, or a storage medium such as a CD-RW or CD-R And the computer of the system or apparatus (or CPU or MPU) may read and execute the program code stored in the storage means or the storage medium when used.

1 is a diagram showing a schematic configuration example of a cluster storage system according to a first embodiment of the present invention. FIG. It is a figure which shows the data structural example of a base part. It is a figure which shows the data structural example of difference information. It is a figure which shows the operation | movement sequence of the cluster storage system by 1st Embodiment. It is a flowchart for demonstrating the object access response operation | movement of a virtualizer apparatus. It is a flowchart for demonstrating the object preservation | save operation | movement of a virtualizer apparatus. It is a figure which shows the schematic structural example of the cluster storage system by the 2nd Embodiment of this invention. It is a figure which shows the example of a data structure of the base part by 2nd Embodiment. It is a figure which shows the operation | movement sequence of the cluster storage system by 2nd Embodiment. It is a flowchart for demonstrating the new object preservation | save operation | movement of a virtualizer apparatus. It is a flowchart for demonstrating the object preservation | save operation | movement of a virtualizer apparatus. It is a flowchart for demonstrating the object access response operation | movement of a virtualizer apparatus. It is a figure which shows the operation | movement sequence of a night cluster storage system in 3rd Embodiment. It is a figure which shows the operation | movement sequence of a night cluster storage system in 4th Embodiment. It is a figure which shows the operation | movement sequence of a night cluster storage system in 5th Embodiment.

Explanation of symbols

1: Cluster storage system 2: Network 10: Client device 12: Virtualizer device 14: Storage device 20: User interface 22: CPU
24: Memory 26: Communication interface 28: Storage means 500: Object creation means 502: Object update means 520: Access interpretation means 522: Base ID management means 524: Base part search means 526: Difference information search means 528: Difference information generation means 5210: Difference information confirmation unit 5212: Object restoration unit 5214: Storage storage determination unit

Claims (19)

Store data provided from three or more storage devices and an external device in any of the three or more storage devices, and access to data stored in any of the three or more storage devices An information recording / reproducing control device for controlling the storage device,
The information recording / reproducing control apparatus comprises an access interpreting means for executing storage and access of the data in units of objects.   2. The cluster storage apparatus according to claim 1, wherein the access interpreting means manages the data by assigning an ID to the object unit data.   The access interpreting means receives an access request for data stored in the three or more storage devices, acquires an ID corresponding to the data to be accessed, and realizes access to the data. The cluster storage device described in 1. The information recording / reproduction control device further receives the update data generated by updating the original data stored in any of the three or more storage devices, and updates the update data and the original data. A difference information generating means for taking a difference and generating difference information;
4. The cluster storage according to claim 3, wherein the access interpretation unit assigns the same ID as the original data to the difference information and stores the difference information in any of the three or more storage devices. apparatus. The information recording / reproducing control device further includes:
Based on an ID corresponding to the access target data, original data search means for searching the original data;
Difference information search means for searching for the difference information based on the ID;
5. The cluster storage apparatus according to claim 4, further comprising a restoration unit that restores the update data in units of objects using the original data and the difference information.   3. The cluster storage apparatus according to claim 2, wherein the access interpreting unit stores data in units of objects to which the ID is assigned in at least two of the three or more storage apparatuses. The information recording / reproduction control device further receives the update data generated by updating the original data stored in any of the three or more storage devices, and updates the update data and the original data. A difference information generating means for taking a difference and generating difference information;
The cluster according to claim 6, wherein the access interpretation unit assigns the same ID as the original data to the difference information and stores the difference information in at least two of the three or more storage devices. -Storage device.   5. The access interpreting unit aggregates and stores one original data and all the difference information related to the original data in at least one of the three or more storage devices. The cluster storage device described in 1. The information recording / reproducing control apparatus further includes original flag adding means for adding an original flag indicating that the original data is original data to the original data stored in a centralized manner in the one storage device,
The said access interpretation means aggregates and stores the said original data to which the said original flag was given, and all the said difference information relevant to it in the said one storage apparatus. The cluster storage device described. When the access interpreting unit receives the access request, the access interpreting unit obtains the ID of the corresponding object unit data,
The information recording / reproducing control device further includes:
Original data search means for preferentially searching the original data from the storage device storing the original data having the original flag based on the ID corresponding to the data of the access request;
Based on the ID, difference information search means for searching all the related difference information from the storage device that stores the original data having the original flag;
10. The cluster storage apparatus according to claim 9, further comprising a restoration unit that restores the data in units of objects using the original data having the original flag and all the related difference information. Store data provided from three or more storage devices and an external device in any of the three or more storage devices, and access to data stored in any of the three or more storage devices An information recording / reproducing control device for controlling the storage device,
The information recording / reproducing control apparatus comprises:
Failure detection means for detecting a failure of the storage device;
Search means for searching data stored in an operable storage device other than the storage device in which the failure is detected, and checking whether a predetermined redundancy is secured;
Data storage means for storing insufficient data in the operable storage device so that the predetermined redundancy is ensured when the predetermined redundancy is not ensured;
A cluster storage apparatus comprising: Store data provided from three or more storage devices and an external device in any of the three or more storage devices, and access to data stored in any of the three or more storage devices An information recording / reproducing control device for controlling the storage device,
The information recording / reproducing control apparatus comprises:
Failure detection means for detecting a failure of the storage device;
Search means for searching data stored in an operable storage device other than the storage device in which the failure is detected, and checking whether a predetermined redundancy is secured;
Data organizing means for deleting a part of data stored in the operable storage device and reducing the predetermined redundancy when the predetermined redundancy is not secured;
A cluster storage apparatus comprising: Store data provided from three or more storage devices and an external device in any of the three or more storage devices, and access to data stored in any of the three or more storage devices A method of controlling a cluster storage device comprising:
In the information recording / reproducing control apparatus, the access interpreting unit executes storage and access of the data in units of objects.   14. The method according to claim 13, wherein the access interpreting unit manages the data by assigning an ID to the object unit data.   15. The access interpreting unit receives an access request for data stored in the three or more storage devices, acquires an ID corresponding to data to be accessed, and realizes data access. The method described in 1. In the information recording / reproducing control apparatus, when the difference information generating means receives the update data generated by updating the original data stored in any of the three or more storage apparatuses, the update data And the difference between the original data and generate difference information,
16. The method according to claim 15, wherein the access interpreting unit assigns the same ID as the original data to the difference information and stores the difference information in any of the three or more storage devices.   17. The access interpreting unit stores one piece of the original data and all the difference information related thereto in a concentrated manner in at least one of the three or more storage devices. The method described in 1. In the information recording / reproducing control apparatus,
An original flag assigning unit assigns an original flag indicating that the original data is original data to the original data collected and stored in the one storage device,
18. The access interpreting unit aggregates and stores the original data to which the original flag is assigned and all the difference information related to the original data in the one storage device. The method described. When the access interpretation means receives the access request, it obtains the ID of the corresponding object unit data,
In the information recording / reproducing control apparatus,
Based on an ID corresponding to the access request data, the original data search means preferentially searches for the original data from the storage device storing the original data having the original flag,
The difference information search means searches all the related difference information from the storage device storing the original data having the original flag based on the ID,
19. The method according to claim 18, wherein the restoration means restores the data in units of objects using the original data having the original flag and all the related difference information.

Images