(19)
(11)EP 3 223 132 A1

(12)EUROPEAN PATENT APPLICATION

(43)Date of publication:
27.09.2017 Bulletin 2017/39

(21)Application number: 16177688.5

(22)Date of filing:  14.12.2012
(51)Int. Cl.: 
G06F 3/06  (2006.01)
(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 31.12.2011 CN 201110460211

(62)Application number of the earlier application in accordance with Art. 76 EPC:
12861614.1 / 2746927

(71)Applicant: Huawei Technologies Co., Ltd.
Longgang District Shenzhen, Guangdong 518129 (CN)

(72)Inventors:
  • LIN, Yu
    Shenzhen Guangdong (CN)
  • LIANG, Shangdong
    Shenzhen Guangdong (CN)

(74)Representative: Kreuz, Georg Maria 
Huawei Technologies Duesseldorf GmbH Riesstrasse 8
80992 München
80992 München (DE)

 
Remarks:
Claims filed after the date of receipt of the application (Rule 68(4) EPC).
 


(54)METHOD AND APPARATUS FOR TIERED STORAGE PROCESSING OF DATA, AND STORAGE DEVICE


(57) The present invention provides a method and an apparatus for tiered storage processing of data, and a storage device. The method includes: when a migration unit of a lower-tier disk is migrated to a higher-tier disk, splitting the migration unit into multiple migration subunits, and detecting a data access frequency of each migration subunit respectively; migrating the migration subunit to the lower-tier disk when detecting that the data access frequency of the migration subunit is lower than a set threshold; and combining the multiple migration subunits into the migration unit when detecting that the multiple migration subunits are all migrated to the lower-tier disk. The present invention improves usage of storage media, and controls metadata storage resource consumption effectively.




Description


[0001] This application claims priority to Chinese Patent Application No. 201110460211.X, filed with the Chinese Patent Office on December 31, 2011 and entitled "METHOD AND APPARATUS FOR TIERED STORAGE PROCESSING OF DATA, AND STORAGE DEVICE", which is hereby incorporated by reference in its entirety.

TECHNICAL FIELD



[0002] The present invention relates to storage technologies, and in particular, to a method and an apparatus for tiered storage processing of data, and a storage device.

BACKGROUND



[0003] Tiered storage refers to setting an existing storage device as disk tiers and setting different types of disks onto different tiers so that data can be migrated between different tiers according to data access requirements. For example, a solid state disk (solid state disk, SSD for short) is a higher-tier disk characterized by a high access speed, and a hard disk drive (Hard Disk Drive, HDD for short) is a lower-tier disk characterized by a low access speed. In the practice of tiered storage, infrequently accessed data is stored into a disk tier characterized by a low access speed such as an HDD, and frequently accessed data is stored into a disk tier characterized by a high access speed such as an SSD, and therefore, in accessing frequently used data, the data access speed and the storage device access efficiency are enhanced.

[0004] A disk generally stores data in the form of data blocks. Multiple storage units of the same size are set in the disk, and each storage unit is configured to store one data block. When data is migrated between different tiers, the data block as a whole is migrated. Even if only a part of data (called "hot data") in the data block has an access frequency rising to a set threshold and needs to be migrated to a higher-tier disk, the data block as a whole is actually migrated. That is, the "cold data" (data whose access frequency is still low) in the data block moves together with "hot data". The data block is generally called a "migration unit", indicating that the data block is a unit of migration. In addition, the storage device further includes metadata storage resources used to store metadata, where the metadata is information corresponding to a migration unit, for example, a storage address of the migration unit, and each migration unit has corresponding metadata, and therefore, the consumption amount of the metadata storage resources is proportional to the number of migration units.

[0005] Currently, the size of the "migration unit" is generally fixed, such as 256 MB fixedly. However, the setting of the migration units of a fixed size brings the following problems: If the set size of the migration unit is large, few metadata storage resources are consumed, but the phenomenon that cold data moves together with the hot data will lead to waste of the storage space of the expensive higher-tier storage medium and low usage of space; and, if the set size of the migration unit is small, the waste of the higher-tier storage medium is reduced naturally, but more metadata storage resources are consumed.

SUMMARY



[0006] One aspect of the present invention provides a method for tiered storage processing of data to improve usage of storage media and control reasonableness of metadata storage resource consumption.

[0007] Another aspect of the present invention provides an apparatus for tiered storage processing of data to improve usage of storage media and control reasonableness of metadata storage resource consumption.

[0008] Still another aspect of the present invention provides a storage device to improve usage of storage media and control reasonableness of metadata storage resource consumption.

[0009] The method for tiered storage processing of data provided in the present invention includes:

when a migration unit of a lower-tier disk is migrated to a higher-tier disk, splitting the migration unit into multiple migration subunits, and detecting a data access frequency of each migration subunit respectively;

migrating the migration subunit to the lower-tier disk when detecting that the data access frequency of the migration subunit is lower than a set threshold; and

combining the multiple migration subunits into the migration unit when detecting that the multiple migration subunits are all migrated to the lower-tier disk.



[0010] The apparatus for tiered storage processing of data provided in the present invention includes:

a resource adjusting unit, configured to split a migration unit into multiple migration subunits, and combine the multiple migration subunits into the migration unit;

a hotness identifying unit, configured to detect a data access frequency of each migration subunit respectively; and

a migration controlling unit, configured to: migrate a migration unit in a lower-tier disk to a higher-tier disk; migrate the migration subunit to the lower-tier disk when the hotness identifying unit detects that the data access frequency of the migration subunit is lower than a set threshold; and, detect whether the multiple migration subunits are all migrated to the lower-tier disk.



[0011] The storage device provided in the present invention includes a lower-tier disk and a higher-tier disk, and further includes an apparatus for tiered storage processing of data provided in the present invention, where the apparatus for tiered storage processing of data is connected to the lower-tier disk and the higher-tier disk respectively.

[0012] Technical effects of the method for tiered storage processing of data in the present invention are: When a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

[0013] Technical effects of the apparatus for tiered storage processing of data in the present invention are: When a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

[0014] Technical effects of the storage device in the present invention are: When a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

BRIEF DESCRIPTION OF DRAWINGS



[0015] 

FIG. 1 is a schematic flowchart of an embodiment of a method for tiered storage processing of data according to the present invention;

FIG. 2 is a schematic principle diagram of another embodiment of a method for tiered storage processing of data according to the present invention; and

FIG. 3 is a schematic structural diagram of an embodiment of an apparatus for tiered storage processing of data according to the present invention.


DESCRIPTION OF EMBODIMENTS


Embodiment 1



[0016] FIG. 1 is a schematic flowchart of an embodiment of a method for tiered storage processing of data according to the present invention. As shown in FIG. 1, the method in this embodiment may include the following steps:

101. When a migration unit of a lower-tier disk is migrated to a higher-tier disk, split the migration unit into multiple migration subunits, and detect a data access frequency of each migration subunit respectively.



[0017] The lower-tier disk refers to a disk characterized by a low access speed, such as an HDD; and the higher-tier disk refers to a disk characterized by a high access speed, such as an SSD, that is, the higher-tier disk provides a better performance and a higher speed than the lower-tier disk.

[0018] The migration unit and the migration subunit refer to data blocks stored in the disk and migrated as an integral unit. The migration unit of a lower-tier disk generally refers to a migration unit of a large size such as 256 MB; and, when the migration unit becomes hot and is migrated to a higher-tier disk, the migration unit is split into multiple migration subunits of a smaller size such as 32 MB. In this embodiment, the number of migration subunits obtained after the splitting is not limited. For example, a migration unit may be split into four migration subunits or eight migration subunits. In addition, the size of the migration subunits obtained after the splitting is user-definable.

[0019] Optionally, after the migration unit is split into multiple migration subunits, storage address information of the multiple migration subunits is recorded respectively to ensure correct data access subsequently. For example, assuming that the migration unit is split into four migration subunits, the specific physical storage addresses of the four migration subunits in the higher-tier disk need to be recorded respectively.

[0020] 102. Migrate the migration subunit to the lower-tier disk when detecting that the data access frequency of the migration subunit is lower than a set threshold.

[0021] If the data access frequency of the migration subunit is lower than a set threshold, it indicates that the access frequency of the migration subunit is low, that is, infrequently accessed, and becomes "cold data", and therefore, the migration subunit is migrated to the lower-tier disk in this embodiment.

[0022] In this embodiment, to improve usage of storage space of the higher-tier disk, a large-sized migration unit is split into small-sized migration subunits, and therefore, hot data in the migration unit can be identified more precisely, and the cold data can be migrated back to the lower-tier disk, thereby preventing the cold data from wastefully occupying the expensive storage space in the higher-tier disk, and improving the disk usage.

[0023] 103. Combine the multiple migration subunits into the migration unit when detecting that the multiple migration subunits are all migrated to the lower-tier disk.

[0024] When it is detected that all migration subunits, which are obtained after the split migration unit migrated to the higher-tier disk in 101, are migrated back to the lower-tier disk, that is, all the migration subunits become cold again, the migration subunits are re-integrated into a large-sized migration unit in this embodiment.

[0025] The lower-tier disk integrates migration subunits into a migration unit in time, thereby reducing the number of migration units and controlling the metadata storage resource consumption effectively.

[0026] In the method for tiered storage processing of data in this embodiment, when a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

Embodiment 2



[0027] FIG. 2 is a schematic principle diagram of another embodiment of a method for tiered storage processing of data according to the present invention. In this embodiment, assuming that the lower-tier disk is an HDD and the higher-tier disk is an SSD, the method for tiered storage processing of data in the present invention is described in detail. The HDD and the SSD may be disks that are set on two different tiers in a storage device, where the storage device further includes a relevant controlling unit that is configured to control migration of data between the two tiers of disks, for example, a hotness identifying unit configured to detect the data access frequency, or a migration controlling unit configured to control data migration, or the like.

[0028] As shown in FIG. 2, the storage space in the HDD and the SSD is divided into multiple storage units at a large size granularity first. For example, the storage unit may be configured to store a migration unit 11 of a 256 MB size. The migration unit 11 is a data block that is stored and migrated as a whole.

[0029] After the storage space is divided into multiple storage units, a mapping relationship is created between each storage unit in the HDD and the SSD and a logical block address (Logical Block Address, LBA) of the storage device respectively. For example, the storage device includes an HDD and an SSD that are tiered, and logical block address information is maintained in the storage device. Data is stored in a physical storage address in the HDD and the SSD. However, when a read or write operation is performed for the data, the operation information carries LBA information. Then according to the mapping relationship between the LBA in the storage device and the physical storage address, the physical storage address of the data is searched out, and the data is found. Therefore, a mapping relationship between each storage unit and the LBA is created in the storage device. Specifically, each storage unit has a corresponding storage identifier, that is, a physical storage address, where the migration unit also has a data block identifier. The mapping relationship records the data block identifier, the logical block address LBA of the data block, and mapping relationship information between the physical storage addresses, so as to identify the storage location of the data block in the disk and ensure the correctness of data reading and writing.

[0030] The hotness identifying unit in the storage device may detect the data access frequency of the migration unit 11 in each storage unit, for example, detect IO statistic information of the migration unit. In addition, an access frequency threshold may be set. If the data access frequency of the migration unit is greater than the threshold, it indicates that the migration unit is frequently accessed and has become "hot data"; if the data access frequency of the migration unit is less than the threshold, it indicates that the migration unit is infrequently accessed and is "cold data".

[0031] After the migration unit becomes hot, the migration controlling unit in the storage device may control the migration unit to migrate to the SSD; and, after the migration to the SSD, the hotness identifying unit in the storage device in this embodiment may detect the data access frequency of each migration subunit in the migration unit respectively.

[0032] Specifically, the storage device may preset a size such as 32 MB for the migration subunit. Assuming that the size of the migration unit is 256 MB, the migration unit actually includes 8 migration subunits. After the migration unit is migrated to the SSD, the hotness identifying unit may detect the data access frequency of the 8 migration subunits respectively rather than detect the data access frequency of the entire migration unit as a whole just like the beginning, and therefore, hot data can be identified more easily. For example, only 64 MB data in the 256 MB migration unit becomes hot and the remaining is still cold data, but because the migration unit as a whole is a unit in the statistics of IO information, the migration unit is still regarded as hot and migrated to an upper tier; and then the IO information of each migration subunit in the migration unit is detected respectively to identify precisely 2 migration subunits that become hot and the remaining 6 migration subunits that are still cold data. In this way, the identification of hot data is more precise.

[0033] Optionally, the IO statistics for each migration subunit may be performed before the splitting of the migration unit, that is, when the migration unit is still a whole without being split, and the hot data of each internal migration subunit is identified beforehand; if all the migration subunits are detected as hot data, the migration unit may remain as a whole without being split, and is stored in the large-sized storage unit of the SSD; and, if some are detected as cold data, the migration unit is controlled to split according to the preset size of migration subunits.

[0034] Alternatively, optionally, if no idle large storage unit corresponding to the size of the migration unit exists in the SSD, the migration unit may be split into multiple migration subunits first, and the migration subunits are migrated to the small storage unit of a corresponding size in the SSD, and then the IO statistics are detected respectively. That is, before the migration unit is split, the storage unit configured to store the migration unit is a storage space of a large size granularity; and, after the migration unit is split, the storage space is also split into multiple storage spaces of a small size granularity to store the multiple migration subunits after the splitting. Therefore, the order between the hotness identification and the migration unit splitting is not limited strictly, and any order is appropriate as long as the IO statistic information of each migration subunit of the migration unit in the SSD can be detected and the hot data can be identified.

[0035] Assuming that it is already identified that the migration unit includes migration subunits corresponding to cold data, the migration unit 11 is split into multiple migration subunits 12 such as 32 MB; and the migration controlling unit controls the migration subunits corresponding to the cold data to migrate back to the HDD so that the cold data no longer occupies the expensive storage space in the SSD, which improves usage of the SSD storage space. In addition, the hot data identification is performed in the SSD all along, that is, the IO statistic information of each migration subunit is detected respectively, and, once it is found that hot data changes to cold data, the cold data is migrated back to the HDD.

[0036] In this embodiment, after the migration unit 11 is split into multiple migration subunits 12, to ensure data access correctness, the mapping relationship information needs to be updated. That is, in this case, the physical storage address of the migration unit 11 on the disk has changed, and the migration unit has been split into multiple migration subunits 12, where the multiple migration subunits 12 may be stored in various addresses in a distributed manner. Therefore, the mapping relationship between the migration unit and the multiple migration subunits needs to be recorded, and the current physical storage addresses of the multiple migration subunits need to be recorded. For example, the mapping relationship between the data block identifier of the migration unit 11 and the LBA of the data block remains unchanged. In searching for the migration unit 11 according to the LBA, the multiple migration subunits 12 obtained after the split migration unit 11 can be found. Some of the migration subunits 12 are stored in the physical storage addresses a1, a2, ..., in the SSD, and other migration subunits 12 are stored in the physical storage addresses b1, b2, ..., in the HDD. In specific implementation, addressing may be performed by means of arrays, and an array member points to a migration subunit in the HDD and the SSD respectively.

[0037] Optionally, when the migration unit in the HDD is migrated to the SSD, the idle large-sized storage unit corresponding to the size of the migration unit in the SSD is preferred; when a cold migration subunit in the SSD is migrated to the HDD, the idle small-sized storage unit corresponding to the size of the migration unit in the HDD is preferred; and, when the cold migration subunit in the HDD becomes hot again, the migration subunit is preferably migrated to the idle small-sized storage unit corresponding to the size of the migration subunit in the SSD.

[0038] In this embodiment, the migration subunit migrated back to the HDD undergoes IO statistics by using the migration subunit as a unit. The storage device maintains a count value for each migration unit, and the count value is used to obtain the number (which may be referred to as a first number) of migration subunits in the SSD, or obtain the number (which may be referred to as a second number) of migration subunits in the HDD; and, if the first number is zero or the second number is equal to the number of migration subunits after the splitting, it is determined that all the multiple migration subunits are migrated to the HDD. For example, if the migration unit is split into four migration subunits, the number of migration subunits obtained after the split migration unit and currently located in the SSD may be recorded; if the number is 4 at the beginning, when the data becomes cold, one of the migration subunits is migrated back to the HDD, and the number decreases by 1; and, gradually, when the number becomes zero, it indicates that all the four migration subunits obtained after the split migration unit have become cold and migrated down to the HDD.

[0039] When all the migration subunits are migrated down to the HDD, an integration process is triggered so that the multiple migration subunits are re-integrated into a migration unit. It is assumed that, when the migration unit is split into migration subunits and migrated down to the HDD, the initial large-sized storage unit of the HDD is split into small-sized storage subunits for storing the migration subunits. For example, a storage unit is split into four storage subunits, and one migration subunit migrated down is stored in one of the storage subunits; and the subsequent three migration subunits migrated down may be stored in other storage subunits obtained after the split storage unit of the HDD, each migration subunit occupies a part of metadata storage resources, and the IO statistics are performed by using the migration subunit as a unit. When it is detected that all the four migration subunits are migrated down, the migration subunits will be re-integrated into a migration unit. That is, the four storage subunits are also re-integrated into a large-sized storage unit. In this case, due to decrease of the number of migration units, the consumption of metadata storage resources decreases, and the consumption of the metadata storage resources is controlled, and the IO statistics begin to be performed by using the integrated migration unit as a unit. In addition, the storage address information of each migration subunit, which is recorded when the migration unit is in a split state, changes back to the storage address information of the migration unit, and the resources occupied by the maintained count value of the migration subunits obtained after the split migration unit are released and the count value is not maintained any longer.

[0040] In the method for tiered storage processing of data in this embodiment, when a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

Embodiment 3



[0041] FIG. 3 is a schematic structural diagram of an embodiment of an apparatus for tiered storage processing of data according to the present invention. The apparatus can perform the method for tiered storage processing of data according to any embodiment of the present invention. This embodiment only gives a brief description about the structure of the apparatus. For detailed working principles of the apparatus, reference may be made to the description in the method embodiment.

[0042] As shown in FIG. 3, the apparatus may include a resource adjusting unit 31, a hotness identifying unit 32, a migration controlling unit 33, and an address recording unit 34.

[0043] The resource adjusting unit 31 is configured to split a migration unit into multiple migration subunits, and combine the multiple migration subunits into the migration unit.

[0044] The hotness identifying unit 32 is configured to detect a data access frequency of each migration subunit respectively.

[0045] The migration controlling unit 33 is configured to: migrate a migration unit in a lower-tier disk to a higher-tier disk; migrate the migration subunit to the lower-tier disk when the hotness identifying unit detects that the data access frequency of the migration subunit is lower than a set threshold; and, detect whether the multiple migration subunits are all migrated to the lower-tier disk.

[0046] Further, the apparatus may further include an address recording unit 34, which is specifically configured to: after the migration unit is split into multiple migration subunits, and before the multiple migration subunits are combined into the migration unit, record storage address information of the multiple migration subunits respectively, for example, record a mapping relationship between the migration unit and the multiple migration subunits; and record current physical storage addresses of the multiple migration subunits.

[0047] Further, the migration controlling unit 33 is further configured to detect whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; and, if a result of the detection is yes, migrate the migration unit to the storage unit; and the resource adjusting unit 31 is further configured to split the migration unit into multiple migration subunits if a migration subunit exists whose data access frequency is lower than a set threshold.

[0048] Further, the migration controlling unit 33 is specifically configured to detect whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; and the resource adjusting unit 31 is specifically configured to split the migration unit into multiple migration subunits if a result of the detection of the migration controlling unit is no.

[0049] Further, the migration controlling unit 33 includes a counting subunit 331 and a state detecting subunit 332. The counting subunit 331 is configured to obtain a first number of the migration subunits in the higher-tier disk, or obtain a second number of the migration subunits in the lower-tier disk; and the state detecting subunit 332 is configured to determine that the multiple migration subunits are all migrated to the lower-tier disk if the first number is zero or the second number is equal to the number of migration subunits obtained after the split migration unit.

[0050] With the apparatus for tiered storage processing of data in this embodiment, when a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively so that hot data can be identified more easily; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

Embodiment 4



[0051] The present invention provides a storage device, where the device includes a lower-tier disk and a higher-tier disk, and further includes an apparatus for tiered storage processing of data provided in any embodiment of the present invention, where the apparatus for tiered storage processing of data is connected to the lower-tier disk and the higher-tier disk respectively.

[0052] With the apparatus for tiered storage processing of data in the storage device, when a migration unit is migrated to a higher-tier disk, a data access frequency of each migration subunit is detected respectively; and when cold data is identified, the migration subunit corresponding to the cold data can be migrated to a lower-tier disk, thereby reducing waste of the storage space of the higher-tier disk, so that the higher-tier disk stores only hot data, and improving the usage of storage media; in addition, when all migration subunits are migrated to the lower-tier disk, integration can be performed by combining all the migration subunits into a migration unit, thereby reducing the number of migration units in time and controlling metadata storage resource consumption effectively.

[0053] Persons of ordinary skill in the art may understand that, all or a part of the steps of the foregoing method embodiments may be implemented by a program instructing relevant hardware. The foregoing programs may be stored in a computer readable storage medium. When the program runs, the steps of the forgoing method embodiments are performed. The foregoing storage medium includes various mediums capable of storing program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

[0054] Finally, it should be noted that the foregoing embodiments are merely intended for describing the technical solutions of the present invention, rather than limiting the present invention. Although the present invention is described in detail with reference to the foregoing embodiments, persons of ordinary skill in the art should understand that they may still make modifications to the technical solutions described in the foregoing embodiments, or make equivalent replacements to some or all the technical features thereof, as long as the modifications and replacements do not cause the essence of corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0055] Further embodiments of the present invention are provided in the following. It should be noted that the numbering used in the following section does not necessarily need to comply with the numbering used in the previous sections.

Embodiment 1. A method for tiered storage processing of data, comprising: when a migration unit of a lower-tier disk is migrated to a higher-tier disk, splitting the migration unit into multiple migration subunits, and detecting a data access frequency of each migration subunit respectively; migrating the migration subunit to the lower-tier disk when detecting that the data access frequency of the migration subunit is lower than a set threshold; and combining the multiple migration subunits into the migration unit when detecting that the multiple migration subunits are all migrated to the lower-tier disk.

Embodiment 2. The method for tiered storage processing of data according to embodiment 1, wherein: after the splitting the migration unit into multiple migration subunits, and before the combining the multiple migration subunits into the migration unit, the method further comprises: recording a mapping relationship between the migration unit and the multiple migration subunits; and recording current physical storage addresses of the multiple migration subunits.

Embodiment 3. The method for tiered storage processing of data according to embodiment 1, wherein: when a migration unit of a lower-tier disk is migrated to a higher-tier disk, the splitting the migration unit into multiple migration subunits, and the detecting a data access frequency of each migration subunit respectively, comprise: detecting whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; if the storage unit exists, migrating the migration unit to the storage unit, and detecting the data access frequency of each migration subunit respectively; and splitting the migration unit into multiple migration subunits if a migration subunit exists whose data access frequency is lower than a set threshold.

Embodiment 4. The method for tiered storage processing of data according to embodiment 1, wherein: when a migration unit of a lower-tier disk is migrated to a higher-tier disk, the splitting the migration unit into multiple migration subunits, and the detecting a data access frequency of each migration subunit respectively, comprise: detecting whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; and if the storage unit does not exist, splitting the migration unit into multiple migration subunits; and, after the migration unit is split into the multiple migration subunits, detecting the data access frequency of each migration subunit respectively.

Embodiment 5. The method for tiered storage processing of data according to embodiment 1, wherein: the detecting that the multiple migration subunits are all migrated to the lower-tier disk comprises: obtaining a first number of the migration subunits in the higher-tier disk, or obtaining a second number of the migration subunits in the lower-tier disk; and determining that the multiple migration subunits are all migrated to the lower-tier disk if the first number is zero or the second number is equal to the number of migration subunits obtained after the split migration unit.

Embodiment 6. An apparatus for tiered storage processing of data, comprising: a resource adjusting unit, configured to split a migration unit into multiple migration subunits, and combine the multiple migration subunits into the migration unit; a hotness identifying unit, configured to detect a data access frequency of each migration subunit respectively; a migration controlling unit, configured to: migrate a migration unit in a lower-tier disk to a higher-tier disk; migrate the migration subunit to the lower-tier disk when the hotness identifying unit detects that the data access frequency of the migration subunit is lower than a set threshold; and, detect whether the multiple migration subunits are all migrated to the lower-tier disk.

Embodiment 7. The apparatus for tiered storage processing of data according to embodiment 6, further comprising: an address recording unit, configured to: after the migration unit is split into multiple migration subunits, and before multiple migration subunits are combined into the migration unit, record a mapping relationship between the migration unit and the multiple migration subunits, and record current physical storage addresses of the multiple migration subunits.

Embodiment 8. The apparatus for tiered storage processing of data according to embodiment 6, wherein: the migration controlling unit is specifically configured to detect whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; and, if a result of the detection is yes, migrate the migration unit to the storage unit; and the resource adjusting unit is specifically configured to split the migration unit into multiple migration subunits if a migration subunit exists whose data access frequency is lower than a set threshold.

Embodiment 9. The apparatus for tiered storage processing of data according to embodiment 6, wherein: the migration controlling unit is specifically configured to detect whether a storage unit corresponding to a size of the migration unit exists in the higher-tier disk; and the resource adjusting unit is specifically configured to split the migration unit into multiple migration subunits if a result of the detection of the migration controlling unit is no.

Embodiment 10. The apparatus for tiered storage processing of data according to embodiment 6, wherein the migration controlling unit comprises: a counting subunit, configured to obtain a first number of the migration subunits in the higher-tier disk, or obtain a second number of the migration subunits in the lower-tier disk; and a state detecting subunit, configured to determine that the multiple migration subunits are all migrated to the lower-tier disk if the first number is zero or the second number is equal to the number of migration subunits obtained after the split migration unit.

Embodiment 11. A storage device, comprising a lower-tier disk and a higher-tier disk, and further comprising an apparatus for tiered storage processing of data according to any one of embodiments 6 to 10, wherein the apparatus for tiered storage processing of data is connected to the lower-tier disk and the higher-tier disk respectively.




Claims

1. A method for tiered storage processing of data, comprising:

monitoring a data access frequency of a migration unit in a lower-tier disk, wherein the migration unit comprises multiple migration subunits;

determing that the data access frequency of the migration unit has reached a access threshold value;

moving the migration unit to a higher-tier disk after the data access frequency of the migration unit reaches above a access threshold value;

performing IO statistics for each of migration subunits in the migration unit before splitting the migration unit;

identifying one or more migration subunits which become cold by detecting the IO statistics for each of migration subunit respectively;

splitting the migration unit into the multiple migration subunits when the one or more migration subunits become cold; and

moving the one or more migration subunits to the lower-tier disk.


 
2. The method according to claim 1, further comprising:

creating a mapping between the migration unit and each of the multiple migration subunits.


 
3. The method according to claim 1, further comprising:

storing a physical address of the migration unit after the migration unit is moved to the higher-tier disk.


 
4. The method according to claim 1, wherein moving the migration unit to a higher-tier disk comprises:

determining whether the higher-tier disk includes a storage unit having a size that is greater than a size of the migration unit; and

moving the migration unit to the storage unit when the higher-tier disk includes the storage disk having the size that is greater than a size of the migration unit.


 
5. The method according to claim 1, further comprising combining all of the multiple migration subunits into the migration unit after each of the multiple migration subunits is moved to the lower-tier disk when each of the multiple migration subunits becomes cold.
 
6. An apparatus for tiered storage processing of data, comprising:

a hotness identifying unit (32), configured to monitor a data access frequency of a migration unit in a lower-tier disk, wherein the migration unit comprises multiple migration subunits; determine that the data access frequency of the migration unit has reached a access threshold value;

a migration controlling unit (33), configured to move the migration unit to a higher-tier disk after the data access frequency of the migration unit reaches above a access threshold value;

the hotness identifying unit (32), further configured to perform IO statistics for each of migration subunits in the migration unit before splitting the migration unit and identify one or more migration subunits which become cold by detecting the IO statistics for each of migration subunit respectively; and

a migration controlling unit (33), configured to split the migration unit into the multiple migration subunits when the one or more migration subunits become cold; and

the migration controlling unit (33), further configured to move the one or more migration subunits to the lower-tier disk.


 
7. The apparatus according to claim 6, wherein the migration controlling unit (33) is further configured to create a mapping between the migration unit and each of the multiple migration subunits.
 
8. The apparatus according to claim 6, further comprising:

an address recording unit (34), configured to store a physical address of the migration unit after the migration unit is moved to the higher-tier disk.


 
9. The apparatus according to claim 6, wherein the migration controlling unit (33) is configured to determine whether the higher-tier disk includes a storage unit having a size that is greater than a size of the migration unit; and
move the migration unit to the storage unit when the higher-tier disk includes the storage disk having the size that is greater than a size of the migration unit.
 
10. The apparatus according to claim 6, wherein the migration controlling unit (33) is further configured to combine all of the multiple migration subunits into the migration unit after each of the multiple migration subunits is moved to the lower-tier disk when each of the multiple migration subunits becomes cold.
 




Drawing

















REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description