CN103558991B - Data hierarchy storage processing method, device and storage device - Google Patents
Data hierarchy storage processing method, device and storage device Download PDFInfo
- Publication number
- CN103558991B CN103558991B CN201110460211.XA CN201110460211A CN103558991B CN 103558991 B CN103558991 B CN 103558991B CN 201110460211 A CN201110460211 A CN 201110460211A CN 103558991 B CN103558991 B CN 103558991B
- Authority
- CN
- China
- Prior art keywords
- migration units
- sub
- migration
- level disk
- units
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 22
- 230000005012 migration Effects 0.000 claims abstract description 168
- 230000000875 corresponding Effects 0.000 claims description 27
- 238000001514 detection method Methods 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims 2
- 229940116821 SSD Drugs 0.000 description 27
- 239000002699 waste material Substances 0.000 description 10
- 238000010438 heat treatment Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 3
- 230000001413 cellular Effects 0.000 description 2
- 230000001276 controlling effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000000034 method Methods 0.000 description 1
Abstract
The present invention provides a kind of data hierarchy storage processing method, device and storage device, wherein method includes: when the migration units at low-level disk moves to high-level disk, described migration units is split into many sub-migration units, and detects the data access frequency of each described sub-migration units respectively;When detecting that described sub-migration units, less than when setting threshold value, is migrated to described low-level disk by the data access frequency of described sub-migration units;When detecting that the plurality of sub-migration units all moves in described low-level disk, then the plurality of sub-migration units is merged into described migration units.The present invention improves the utilization rate of storage medium, and effectively controls the consumption of metadata storage resource.
Description
Technical field
The present invention relates to memory technology, particularly relate to a kind of data hierarchy storage processing method, device and deposit
Storage equipment.
Background technology
Bedding storage refers to existing storage device can be in the way of being arranged to disk layering, by difference
The disk of type is arranged in different levels, and according to the requirements for access of data, data can be at different layers
Between migrate.Such as, (solid state disk is called for short: SSD) belong to access speed relatively solid state hard disc
High high-level disk, (Hard Disk Drive, is called for short hard disk drive: HDD) belong to access speed
Spend relatively low low-level disk, in the realization of Bedding storage, will seldom be accessed for data storage visiting
Ask in slow disk layer such as HDD, and it is fireballing that the data often accessed store access
In disk layer such as SSD, so that when accessing frequently-used data, improving the reading speed of data and deposit
The access efficiency of storage equipment.
Disk is typically to store data with the form of data block, is provided with multiple equivalently-sized in disk
Memory element, each memory element is for one data block of storage.When data migrate between the different layers,
Also with data block bulk migration, even if this data block only having the visiting frequency of part data bring up to
Set threshold value (referred to as " dsc data ") and need to move to high-level disk, be the most also data block
Bulk migration, i.e. " cold data " (data that visiting frequency is the most relatively low) in data block follow " heat
Data " move together.Commonly referred to " migration units ", expression is with integral unit to above-mentioned data block
Form carry out migrating.Additionally, storage device also includes the metadata storage for storing metadata
Resource, this metadata is the information corresponding with migration units, such as, the storage address etc. of migration units,
And each migration units has a corresponding metadata, thus the consumption of metadata storage resource be with
The quantity of migration units is directly proportional.
The size of current " migration units " is the most all fixing, such as, be fixedly installed as 256MB.
But, there are the following problems for the mode of this set fixed dimension migration units: if setting migration units
Size relatively big, metadata storage resource consumption is less, but can there are above-mentioned cold data and follow hot number
According to the phenomenon of movement, the memory space of the high-level storage medium of costliness can be made to cause waste, space profit
Low by rate;If the size setting migration units is less, nature can reduce the waste of high-level storage medium
Phenomenon, but the metadata storage resource consumed is the most.
Summary of the invention
The first aspect of the invention is to provide a kind of data hierarchy storage processing method, it is therefore an objective to improves and deposits
The utilization rate of storage media, and control the reasonability of metadata storage resource consumption.
Another aspect of the present invention is to provide a kind of data hierarchy storage processing means, it is therefore an objective to improves and deposits
The utilization rate of storage media, and control the reasonability of metadata storage resource consumption.
An additional aspect of the present invention is to provide a kind of storage device, it is therefore an objective to improve the utilization of storage medium
Rate, and control the reasonability of metadata storage resource consumption.
The data hierarchy storage processing method that the present invention provides, including:
After the migration units of low-level disk is moved to high-level disk, described migration units is split into
Many sub-migration units, and detect the data access frequency of each described sub-migration units respectively;
When detecting that described son, less than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit and migrate to described low-level disk;
When detecting that the plurality of sub-migration units all moves in described low-level disk, then by institute
State many sub-migration units and merge into described migration units.
Another data hierarchy storage processing method that the present invention provides, including:
Migration units in low-level disk is split into many sub-migration units;Detect each described son respectively
The data access frequency of migration units;
When detecting that described son, higher than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit to migrate in high-level disk;
Detect the data access frequency of sub-migration units described in described high-level disk;
When detecting that described son, less than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit and migrate to described low-level disk;
When detecting that the plurality of sub-migration units all moves in described low-level disk, by described
Many sub-migration units merge into described migration units.
The data hierarchy storage processing means that the present invention provides, including:
Resource adjustment unit, for described migration units being split into many sub-migration units, and, use
In the plurality of sub-migration units is merged into described migration units;
Temperature recognition unit, for detecting the data access frequency of each described sub-migration units respectively;
Migrate control unit, for the migration units of low-level disk is moved to high-level disk;And,
When described temperature recognition unit detects that the data access frequency of described sub-migration units is less than setting threshold value
Time, described sub-migration units is migrated to described low-level disk;And, it is used for detecting the plurality of son
Migration units all moves in described low-level disk.
The another kind of data hierarchy storage processing means that the present invention provides, including:
Migrate control unit, for the migration units of low-level disk is moved to high-level disk;
Resource adjustment unit, for moving to high-level disk by described migration units from low-level disk
After described migration units is split into many sub-migration units, and, for by the plurality of sub-migration list
Unit merges into described migration units;
Temperature recognition unit, for detecting the data access frequency of each described sub-migration units respectively;
Migrate control unit, be additionally operable to when described temperature recognition unit detects the number of described sub-migration units
According to visiting frequency less than when setting threshold value, described sub-migration units is migrated to described low-level disk;With
And, it is used for detecting the plurality of sub-migration units and all moves in described low-level disk.
The storage device that the present invention provides, including low-level disk and high-level disk, also includes the present invention
Described data hierarchy storage processing means;Described data hierarchy storage processing means respectively with described low layer
Level disk, high-level disk connect.
Having the technical effect that by moving to height in migration units of data hierarchy storage processing method of the present invention
During level disk, detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data,
And when identifying cold data, sub-migration units corresponding for these cold data can be migrated to low-level magnetic
Dish, thus decrease the waste of storage space phenomenon of high-level disk so that high-level disk only stores heat
Data, improve the utilization rate of storage medium;And, additionally it is possible to all move in each sub-migration units
Integrate during low-level disk, each sub-migration units is merged into migration units, thus decreases in time
The quantity of migration units, effectively controls the consumption of metadata storage resource.
Having the technical effect that by moving to height in migration units of data hierarchy of the present invention storage processing means
During level disk, detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data,
And when identifying cold data, sub-migration units corresponding for these cold data can be migrated to low-level magnetic
Dish, thus decrease the waste of storage space phenomenon of high-level disk so that high-level disk only stores heat
Data, improve the utilization rate of storage medium;And, additionally it is possible to all move in each sub-migration units
Integrate during low-level disk, each sub-migration units is merged into migration units, thus decreases in time
The quantity of migration units, effectively controls the consumption of metadata storage resource.
Storage device of the present invention have the technical effect that by when migration units moves to high-level disk,
Detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data, and can be
When identifying cold data, sub-migration units corresponding for these cold data is migrated to low-level disk, thus subtracts
Lack the waste of storage space phenomenon of high-level disk so that high-level disk only stores dsc data, improved
The utilization rate of storage medium;And, additionally it is possible to all move to low-level disk in each sub-migration units
Shi Jinhang integrates, and each sub-migration units is merged into migration units, thus decreases migration units in time
Quantity, effectively controls the consumption of metadata storage resource.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of data hierarchy storage processing method one embodiment of the present invention;
Fig. 2 is the principle schematic of data hierarchy another embodiment of storage processing method of the present invention;
Fig. 3 is the structural representation of data hierarchy of the present invention storage processing means embodiment.
Detailed description of the invention
Embodiment one
Fig. 1 is the schematic flow sheet of data hierarchy storage processing method one embodiment of the present invention, such as Fig. 1 institute
Showing, the method for the present embodiment may include that
101, when the migration units at low-level disk moves to high-level disk, described migration units is divided
Split for many sub-migration units, and detect the data access frequency of each described sub-migration units respectively;
Wherein, described low-level disk, refer to the disk that access speed is slower, for example, HDD;
Described high-level disk, refers to access speed disk faster, for example, SSD;That is, high-level
Disk is more preferable than the performance of low-level disk, and speed is faster.
Wherein, described migration units and sub-migration units, refer to as an integral unit in disk
Storage the data block of bulk migration.The large-size that the migration units of described low-level disk is often referred to
Migration units, for example, 256MB;When this migration units heating, when above adjourning high-level disk, will
It is split into the sub-migration units of multiple reduced size, for example, 32MB.In the present embodiment, divided
The quantity of sub-migration units does not limits, for example, it is possible to be split into four sub-migration units or eight sons
Migration units etc.;Further, the size of the sub-migration units divided can independently set.
Optionally, after migration units is split into many sub-migration units, also described by recording respectively
The storage address information of many sub-migration units, to ensure the correct access of follow-up data.For example, it is assumed that
Migration units is split into four sub-migration units, then need to record these four sub-migration units respectively at high level
The concrete physical storage address of level disk.
102, when the data access frequency described sub-migration units being detected is less than and sets threshold value, by described
Sub-migration units migrates to described low-level disk;
Wherein, if the data access frequency of described sub-migration units is less than setting threshold value, then it represents that this son moves
The visiting frequency moving unit is relatively low, is the most infrequently accessed, becomes " cold data ";Then the present embodiment will
This sub-migration units migrates to low-level disk.
In the present embodiment, in order to improve the memory space utilization rate of high-level disk, by large-sized migration
Cellular spliting is undersized sub-migration units, it is possible to identify the heat in this migration units more accurately
Data, and cold Data Migration is returned low-level disk, it is to avoid the costliness of cold data waste high-level disks
Memory space, improve the utilization rate of disk.
103, when detecting that the plurality of sub-migration units all moves in described low-level disk, then
The plurality of sub-migration units is merged into described migration units;
Wherein, when each height detecting that the migration units moving to high-level disk in 101 is split into
When migration units the most all migrates back in low-level disk, i.e. each above-mentioned sub-migration units is the most complete
Portion turns cold, then this little migration units is reintegrated by the present embodiment is a large-sized migration units.
This low-level disk is by being integrated into a migration units by each sub-migration units in time, it is possible to reduce
The quantity of migration units, effectively controls the consumption of metadata storage resource.
The data hierarchy storage processing method of the present embodiment, by moving to high-level disk in migration units
Time, detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data, and energy
Enough when identifying cold data, sub-migration units corresponding for these cold data is migrated to low-level disk, from
And decrease the waste of storage space phenomenon of high-level disk so that high-level disk only stores dsc data,
Improve the utilization rate of storage medium;And, additionally it is possible to all move to low-level in each sub-migration units
Integrate during disk, each sub-migration units is merged into migration units, thus it is single to decrease migration in time
The quantity of unit, effectively controls the consumption of metadata storage resource.
Embodiment two
Fig. 2 is the principle schematic of data hierarchy another embodiment of storage processing method of the present invention, this enforcement
Example is the data hierarchy with low-level disk as HDD, as a example by high-level disk is SSD, to the present invention
Storage processing method elaborates.Wherein, described HDD and SSD can be certain storage device
Two level disks that middle layering is arranged, also include in this storage device for controlling data at the two layer
The associated control element migrated is carried out, such as, for detecting the temperature of data access frequency between level disk
Recognition unit, for controlling the migration control unit etc. of Data Migration.
As in figure 2 it is shown, the memory space in HDD and SSD will be first according to large scale granularity division it is
Multiple memory element, such as, this memory element may be used for the migration units that sizes of memory is 256MB
11.Described migration units 11 is the data block that entirety carries out storing and migrating.
After being divided into multiple memory element, set up each storage in HDD and SSD the most respectively single
(Logical Block Address, is called for short: right between LBA) for unit and the LBA of storage device
Should be related to.Such as, storage device includes HDD and SSD of layering setting and safeguards there is logic
Block address information, the physical storage address that data are stored in HDD and SSD, but at logarithm
According to when being written and read operation, operation information can be carried LBA information, further according in storage device
Corresponding relation between LBA and described physical storage address, finds the physical storage address of data, from
And find data.Therefore, in this storage device, foundation has between each above-mentioned memory element and LBA
Mapping relations;Concrete, each memory element can have a corresponding storage mark i.e. physical store ground
Location, migration units therein also has a data block identifier, and in these mapping relations, record has data block mark
Know, correspondence relationship information between LBA LBA and the physical storage address of data block, mark
Know data block deposit position in disk, to ensure the correctness of reading and writing data.
Temperature recognition unit in storage device can detect the migration units 11 in each memory element
Data access frequency, such as, detect the IO statistical information of this migration units.And it is possible to set an access
Frequency threshold, if the data access frequency of migration units is more than this threshold value, then it represents that this migration units is by warp
Frequentation is asked, heating becomes " dsc data ";If the data access frequency of migration units is less than this threshold value,
Then represent that this migration units is the most accessed, be " cold data ".
Migration control unit in storage device can after migration units heating, control this migration units to
SSD migrates;Further the temperature identification, after moving in SSD, in the storage device of the present embodiment
Unit can detect the data access frequency of each sub-migration units in this migration units respectively.
Concrete, storage device can preset the size of a sub-migration units, such as 32MB;
The size assuming migration units is 256MB, then this migration units substantially includes 8 sub-migration units.
After migration units moves in SSD, temperature recognition unit can detect these 8 sons respectively and migrate single
The data access frequency of unit, and detecting this with migration units generally unit and whole move when being no longer initial
Move the data access frequency of unit, be so easier to identify dsc data.Such as, the migration of originally 256MB
Being the data heating of only 64MB in fact in unit, other remain cold data, but owing to being to move
Move unit generally unit and add up its IO information, so still will be considered that this migration units is heating, will
Move in this migration units;Then by detecting the IO information of wherein each sub-migration units respectively, permissible
It is accurately identified the sub-migration units of 2 heating, and other 6 sons remaining cold data migrate single
Unit, the identification of dsc data is more accurate.
Optionally, the above-mentioned IO for each sub-migration units adds up, and can be to divide it in migration units
Before carry out, the most now migration units does not also divide, and remains an entirety, simply the most first each to its inside
Sub-migration units carries out dsc data identification;If detecting the most all dsc datas, then this migration units can
Dividing, remain in that an entirety, in the memory element of the large-size being stored in SSD;If
Detect and which includes cold data, then control migration units division, and be also to set in advance according to described
The size of fixed sub-migration units carries out dividing division.
Or, optionally, if SSD does not has the bigger storage corresponding with migration units size of free time
During unit, this migration units can first be split into many sub-migration units, and sub-migration units is moved to SSD
In less memory element corresponding to size in, then the IO of the detection respectively statistics described in performing.That is, move
Shifting unit is before division, for storing the storage sky that the memory element of this migration units is large scale granularity
Between;After migration units divides, memory space also can produce division, becomes multiple small size granularity
Memory space, many sub-migration units after storing division.Therefore, temperature identification and migration units
The order of division does not do strict restriction, if each the sub-migration units being capable of detecting when in this migration units
IO statistical information in SSD and identify dsc data.
Assume to have identified the sub-migration units that migration units includes that cold data are corresponding, then migration units
11 will be split into many sub-migration units 12, for example, 32MB;And migrate control unit and will control it
In sub-migration units corresponding to cold data move back in HDD so that these cold data no longer take SSD
In expensive memory space, improve the memory space utilization rate of SSD.Further, will always in SSD
Perform above-mentioned dsc data recognition method, detect the IO statistical information of each sub-migration units the most respectively, when
It is found to have original dsc data and turns cold when becoming cold data, just these cold data are moved back in HDD.
In the present embodiment, after migration units 11 is split into many sub-migration units 12, in order to ensure
The correctness of data access, needs to update mapping relation information, i.e. now migration units 11 is on disk
Physical storage address have occurred and that change, and be to be split into many sub-migration units 12, the plurality of son
Migration units 12 may be stored on each address in dispersion.Accordingly, it would be desirable to record migration units, described many
Corresponding relation between individual sub-migration units, and, record the plurality of sub-migration units at current thing
Reason storage address.Such as, between the LBA at the data block identifier of this migration units 11, data block place
Corresponding relation be constant, when searching this migration units 11 by this LBA, this migration will be inquired
Unit 11 is split into many sub-migration units 12, and the sub-migration units of a portion 12 is stored in SSD
Physical storage address a1, a2 ..., the physics that the sub-migration units of another part 12 is stored in HDD
Storage address b1, b2 ....Such as can organize addressing to use array mode in implementing, array becomes
The sub-migration units that member is respectively directed in HDD and SSD.
Optionally, when the migration units in HDD moves in SSD, will preferentially use the free time in SSD
The large scale memory element corresponding with migration units size;When the sub-migration units turned cold in SSD moves to
Time in HDD, also the small size corresponding with sub-migration units size idle in HDD will be preferentially used to deposit
Storage unit;When moving to the sub-migration units heating again in HDD when turning cold, preferentially move to SSD hollow
In the not busy small size memory element corresponding with sub-migration units size.
In the present embodiment, the sub-migration units moved back to HDD, will enter in units of this sub-migration units
Row IO adds up.Storage device is also that each transfer unit safeguards a count value, is used for obtaining in SSD
The quantity (being properly termed as the first quantity) of sub-migration units, or, obtain the son in HDD and migrate single
The quantity (being properly termed as the second quantity) of unit;If the first quantity is zero, or, the second quantity is equal to moving
Move the number of the sub-migration units of cellular spliting, it is determined that many sub-migration units all move to HDD
In.Such as, migration units is split into four sub-migration units, then can record this migration units and currently exist
The number of the sub-migration units in SSD;Initially this number is four, then moves back to HDD when data turn cold,
Number just deducts 1;Gradually, when this number vanishing, then it represents that this migration units split into four
Individual sub-migration units is adjourned in HDD under the most all turning cold and being whole.
When adjourning in HDD under above-mentioned sub-migration units is whole, flow process is integrated in triggering, by described many
It is a migration units that individual sub-migration units reconsolidates.Assume the sub-migration units that migration units splits into
Under when adjourning in HDD, initial large-sized memory element of HDD will be split into undersized sub-storage
Unit, for storing described sub-migration units;Such as, a memory element is split into four son storages
Unit, under a sub-migration units moving be stored therein in a sub-memory element;Three moved under follow-up
Individual sub-migration units is likely to be stored in the sub-memory element of other memory element division of HDD, each
Sub-migration units all takies a part of metadata storage resource, and IO statistics is also to be with sub-migration units
Unit is added up.When detect move under four sub-migration units are all time, by by sub-migration units weight
Newly merging into a migration units, it is a large-sized storage that i.e. four sub-memory element are also reintegrated
Unit, now reduces due to migration units quantity, and metadata storage resource consumption reduces, and controls unit's number
According to storage resource consumption, and IO adds up in units of also beginning to the migration units after this integration and unites
Meter.Additionally, the storage address information of each sub-migration units recorded when migration units is in splitting status
The most again becoming the storage address information of this migration units, the son of the migration units division safeguarded migrates single
The count value of unit takies resource and also can discharge, and no longer safeguards.
The data hierarchy storage processing method of the present embodiment, by moving to high-level disk in migration units
Time, detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data, and energy
Enough when identifying cold data, sub-migration units corresponding for these cold data is migrated to low-level disk, from
And decrease the waste of storage space phenomenon of high-level disk so that high-level disk only stores dsc data,
Improve the utilization rate of storage medium;And, additionally it is possible to all move to low-level in each sub-migration units
Integrate during disk, each sub-migration units is merged into migration units, thus it is single to decrease migration in time
The quantity of unit, effectively controls the consumption of metadata storage resource.
Embodiment three
Fig. 3 is the structural representation of data hierarchy of the present invention storage processing means embodiment, and this device is permissible
Perform the data hierarchy storage processing method of any embodiment of the present invention, the present embodiment the knot to this device
Structure briefly describes, and concrete operation principle can be in conjunction with seeing described in embodiment of the method.
As it is shown on figure 3, this device may include that resource adjustment unit 31, temperature recognition unit 32, moves
Move control unit 33, address record unit 34;Wherein,
Resource adjustment unit 31, for described migration units being split into many sub-migration units, and,
For the plurality of sub-migration units is merged into described migration units;
Temperature recognition unit 32, for detecting the data access frequency of each described sub-migration units respectively;
Migrate control unit 33, for the migration units of low-level disk is moved to high-level disk;With
And, when described temperature recognition unit detects that the data access frequency of described sub-migration units is less than setting threshold
During value, described sub-migration units is migrated to described low-level disk;And, it is used for detecting the plurality of
Sub-migration units all moves in described low-level disk.
Further, it is also possible to include that address records unit 34, specifically for migration units being divided described
After splitting for many sub-migration units, described many sub-migration units are merged into described migration units before,
Record the storage address information of the plurality of sub-migration units respectively, such as, record described migration units,
Corresponding relation between the plurality of sub-migration units;And, record the plurality of sub-migration units current
Physical storage address.
Further, migrate control unit 33, be additionally operable to detect in described high-level disk whether exist with
The memory element that described migration units size is corresponding, and when testing result is for being, by described migration units
Move to described memory element;Resource adjustment unit 31, is additionally operable to there is data access frequency less than setting
When determining the sub-migration units of threshold value, described migration units is split into many sub-migration units.
Further, migrate control unit 33, whether exist specifically for detecting in described high-level disk
The memory element corresponding with described migration units size;Resource adjustment unit 31, specifically for migrating control
When the testing result of unit processed is no, described migration units is split into many sub-migration units.
Further, migrate control unit 33 to include: statistical counting subelement 331 and state-detection sub-list
Unit 332.Wherein, statistical counting subelement 331, moves for the described son obtained in described high-level disk
Move the first quantity of unit, or, obtain second of described sub-migration units in described low-level disk
Quantity;State-detection subelement 332, being used in described first quantity is zero, or, described second number
When amount is equal to the number of the sub-migration units that described migration units divides, it is determined that multiple described sub-migration lists
Unit all moves in described low-level disk.
The data hierarchy storage processing means of the present embodiment, by moving to high-level disk in migration units
Time, detect the data access frequency of each sub-migration units respectively, it is easier to identify dsc data, and energy
Enough when identifying cold data, sub-migration units corresponding for these cold data is migrated to low-level disk, from
And decrease the waste of storage space phenomenon of high-level disk so that high-level disk only stores dsc data,
Improve the utilization rate of storage medium;And, additionally it is possible to all move to low-level in each sub-migration units
Integrate during disk, each sub-migration units is merged into migration units, thus it is single to decrease migration in time
The quantity of unit, effectively controls the consumption of metadata storage resource.
Embodiment four
The invention provides a kind of storage device, this equipment includes low-level disk and high-level disk, also
Processing means is stored including the data hierarchy described in any embodiment of the present invention;At described data hierarchy storage
Reason device is connected with described low-level disk, high-level disk respectively.
Data hierarchy storage processing means in this storage device, can move to high-level in migration units
During disk, detect the data access frequency of each sub-migration units respectively, and cold data can identified
Time, sub-migration units corresponding for these cold data is migrated to low-level disk, thus decreases high-level magnetic
The waste of storage space phenomenon of dish so that high-level disk only stores dsc data, improves storage medium
Utilization rate;Can also integrate when each sub-migration units all moves to low-level disk, Jiang Gezi
Migration units merges into migration units, thus decreases the quantity of migration units in time, effectively controls unit
The consumption of data storage resource.
One of ordinary skill in the art will appreciate that: realize all or part of step of above-mentioned each method embodiment
Suddenly can be completed by the hardware that programmed instruction is relevant.Aforesaid program can be stored in a computer can
Read in storage medium.This program upon execution, performs to include the step of above-mentioned each method embodiment;And
Aforesaid storage medium includes: ROM, RAM, magnetic disc or CD etc. are various can store program code
Medium.
Last it is noted that various embodiments above is only in order to illustrate technical scheme, rather than right
It limits;Although the present invention being described in detail with reference to foregoing embodiments, this area common
Skilled artisans appreciate that the technical scheme described in foregoing embodiments still can be modified by it,
Or the most some or all of technical characteristic is carried out equivalent;And these amendments or replacement, and
The essence not making appropriate technical solution departs from the scope of various embodiments of the present invention technical scheme.
Claims (12)
1. a data hierarchy storage processing method, it is characterised in that including:
After the migration units of low-level disk is moved to high-level disk, described migration units is split into
Many sub-migration units,
Detect the data access frequency of each described sub-migration units respectively;
When detecting that described son, less than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit and migrate to described low-level disk;When detecting that the plurality of sub-migration units all moves to institute
When stating in low-level disk, then the plurality of sub-migration units is merged into described migration units.
Data hierarchy storage processing method the most according to claim 1, it is characterised in that described
After migration units is split into many sub-migration units, described many sub-migration units are merged into described
Before migration units, also include:
Record the corresponding relation between described migration units, the plurality of sub-migration units;
Record the physical storage address that the plurality of sub-migration units is current.
Data hierarchy storage processing method the most according to claim 2, it is characterised in that described inspection
Measure the plurality of sub-migration units all to move in described low-level disk, particularly as follows:
Obtain the first quantity of described sub-migration units in described high-level disk, or, obtain described
Second quantity of the described sub-migration units in low-level disk;
If described first quantity is zero, or, described second quantity is equal to the son of described migration units division
The number of migration units, it is determined that the plurality of sub-migration units all moves in described low-level disk.
4. a data hierarchy storage processing method, it is characterised in that described method includes:
Migration units in low-level disk is split into many sub-migration units;Detect each described son respectively
The data access frequency of migration units;
When detecting that described son, higher than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit to migrate in high-level disk;
Detect the data access frequency of sub-migration units described in described high-level disk;
When detecting that described son, less than when setting threshold value, is moved by the data access frequency of described sub-migration units
Move unit and migrate to described low-level disk;
When detecting that the plurality of sub-migration units all moves in described low-level disk, by described
Many sub-migration units merge into described migration units.
Data hierarchy storage processing method the most according to claim 4, it is characterised in that described
After migration units is split into many sub-migration units, described method also includes:
Record the corresponding relation between described migration units, the plurality of sub-migration units;
Record the physical storage address that the plurality of sub-migration units is current.
Data hierarchy storage processing method the most according to claim 4, it is characterised in that described inspection
Measure many sub-migration units and all move in described low-level disk, including:
Obtain the first quantity of described sub-migration units in described high-level disk, or, obtain described
Second quantity of the described sub-migration units in low-level disk;
If described first quantity is zero, or, described second quantity is equal to the son of described migration units division
The number of migration units, it is determined that the plurality of sub-migration units all moves in described low-level disk.
7. a data hierarchy storage processing means, it is characterised in that including:
Resource adjustment unit, for described migration units being split into many sub-migration units, and, use
In the plurality of sub-migration units is merged into described migration units;
Temperature recognition unit, for detecting the data access frequency of each described sub-migration units respectively;
Migrate control unit, for the migration units of low-level disk is moved to high-level disk;And,
When described temperature recognition unit detects that the data access frequency of described sub-migration units is less than setting threshold value
Time, described sub-migration units is migrated to described low-level disk;And, it is used for detecting the plurality of son
Migration units all moves in described low-level disk.
Data hierarchy the most according to claim 7 storage processing means, it is characterised in that also include:
Address record unit, for described migration units is split into many sub-migration units after, described
Before many sub-migration units are merged into described migration units, record described migration units, the plurality of
Corresponding relation between sub-migration units, and record the physical store that the plurality of sub-migration units is current
Address.
Data hierarchy the most according to claim 7 storage processing means, it is characterised in that described in move
Shifting control unit includes:
Statistical counting subelement, for obtaining first of described sub-migration units in described high-level disk
Quantity, or, obtain the second quantity of described sub-migration units in described low-level disk;
State-detection subelement, being used in described first quantity is zero, or, described second quantity is equal to
During the number of sub-migration units of described migration units division, it is determined that multiple described sub-migration units are whole
Move in described low-level disk.
10. a data hierarchy storage processing means, it is characterised in that including:
Migrate control unit, for the migration units of low-level disk is moved to high-level disk;
Resource adjustment unit, for moving to high-level disk by described migration units from low-level disk
After described migration units is split into many sub-migration units, and, for by the plurality of sub-migration list
Unit merges into described migration units;
Temperature recognition unit, for detecting the data access frequency of each described sub-migration units respectively;
Migrate control unit, be additionally operable to when described temperature recognition unit detects the number of described sub-migration units
According to visiting frequency less than when setting threshold value, described sub-migration units is migrated to described low-level disk;With
And, it is used for detecting the plurality of sub-migration units and all moves in described low-level disk.
11. data hierarchy according to claim 10 storage processing meanss, it is characterised in that also wrap
Include:
Address record unit, for described migration units is split into many sub-migration units after, described
Before many sub-migration units are merged into described migration units, record described migration units, the plurality of
Corresponding relation between sub-migration units, and record the physical store that the plurality of sub-migration units is current
Address.
12. data hierarchy according to claim 10 storage processing meanss, it is characterised in that described
Migration control unit includes:
Statistical counting subelement, for obtaining first of described sub-migration units in described high-level disk
Quantity, or, obtain the second quantity of described sub-migration units in described low-level disk;
State-detection subelement, being used in described first quantity is zero, or, described second quantity is equal to
During the number of sub-migration units of described migration units division, it is determined that multiple described sub-migration units are whole
Move in described low-level disk.
Priority Applications (7)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610831100.8A CN106469029B (en) | 2011-12-31 | 2011-12-31 | Data hierarchy storage processing method, device and storage equipment |
CN201110460211.XA CN103558991B (en) | 2011-12-31 | Data hierarchy storage processing method, device and storage device | |
PCT/CN2012/086670 WO2013097621A1 (en) | 2011-12-31 | 2012-12-14 | Data-tiered storage processing method and device and storage device |
EP16177688.5A EP3223132B1 (en) | 2011-12-31 | 2012-12-14 | Method and apparatus for tiered storage processing of data, and storage device |
EP12861614.1A EP2746927B1 (en) | 2011-12-31 | 2012-12-14 | Data-tiered storage processing method and device and storage device |
US14/190,264 US9665291B2 (en) | 2011-12-31 | 2014-02-26 | Migration of hot and cold data between high-performance storage and low-performance storage at block and sub-block granularities |
US14/706,616 US9665293B2 (en) | 2011-12-31 | 2015-05-07 | Method for a storage device processing data and storage device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110460211.XA CN103558991B (en) | 2011-12-31 | Data hierarchy storage processing method, device and storage device |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610831100.8A Division CN106469029B (en) | 2011-12-31 | 2011-12-31 | Data hierarchy storage processing method, device and storage equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103558991A CN103558991A (en) | 2014-02-05 |
CN103558991B true CN103558991B (en) | 2016-12-14 |
Family
ID=
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106469029B (en) | Data hierarchy storage processing method, device and storage equipment | |
US9423970B2 (en) | Method and system for predicting block failure in a non-volatile memory | |
US9329797B2 (en) | Method and system for adjusting block erase or program parameters based on a predicted erase life | |
EP2940691B1 (en) | Storage controller, storage device, storage system and method of operating the storage controller | |
US9239781B2 (en) | Storage control system with erase block mechanism and method of operation thereof | |
EP3588259B1 (en) | Garbage collection method for storage media, storage medium, and program product | |
CN104216665B (en) | The memory management method of multilevel-cell solid state hard disc | |
US8478939B2 (en) | Relative heat index based hot data determination for block based storage tiering | |
US20100030948A1 (en) | Solid state storage system with data attribute wear leveling and method of controlling the solid state storage system | |
CN112639979A (en) | Write time based interference management | |
CN109599143A (en) | With the storage system and its operating method for reading threshold mechanism | |
CN109947353A (en) | Memory management method, solid state hard disk and readable storage medium storing program for executing | |
CN106371762A (en) | Optimization method and system of storage data | |
CN112650443A (en) | Method of operating memory controller for controlling nonvolatile memory device | |
CN107203341A (en) | Date storage method, device and flash chip based on flash memory | |
US10049037B2 (en) | Data management in a storage system | |
KR101070511B1 (en) | Solid state drive controller and method for operating of the solid state drive controller | |
CN103558991B (en) | Data hierarchy storage processing method, device and storage device | |
KR20240073166A (en) | Controlled system management based on storage device thermal load | |
CN108986863A (en) | Data storage device with rewritable original place memory | |
KR101811444B1 (en) | Method and apparatus of migrating data distribution for decreasing read interference error in flash memory | |
KR20220138269A (en) | Data storage device dynamically determinning whether data is hot or cold and operation method of the same | |
CN108984111A (en) | Data storage device with rewritable original place memory | |
TWI763050B (en) | Self-adaptive wear leveling method and algorithm and related memory device and apparatus | |
CN113986140B (en) | High-transmission-rate high-capacity cloud storage solid state disk |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
TR01 | Transfer of patent right |
Effective date of registration: 20221012 Address after: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee after: HUAWEI TECHNOLOGIES Co.,Ltd. Address before: 611731 Qingshui River District, Chengdu hi tech Zone, Sichuan, China Patentee before: HUAWEI DIGITAL TECHNOLOGIES (CHENG DU) Co.,Ltd. |