CN110162273A - A kind of attenuation type tiered storage system and method based on distributed memory system - Google Patents
A kind of attenuation type tiered storage system and method based on distributed memory system Download PDFInfo
- Publication number
- CN110162273A CN110162273A CN201910448847.9A CN201910448847A CN110162273A CN 110162273 A CN110162273 A CN 110162273A CN 201910448847 A CN201910448847 A CN 201910448847A CN 110162273 A CN110162273 A CN 110162273A
- Authority
- CN
- China
- Prior art keywords
- data
- migration
- object fragment
- layer
- measurement period
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0646—Horizontal data movement in storage systems, i.e. moving data in between storage devices or systems
- G06F3/0647—Migration mechanisms
- G06F3/0649—Lifecycle management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0683—Plurality of storage devices
- G06F3/0685—Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of attenuation type tiered storage system and method based on distributed memory system, are related to technical field of distributed memory.The present invention utilizes the timeliness of data in information system, the data for the system that flows into are identified with different temperatures, largely no longer accessed " cold " Data Migration is gone out into valuable high speed storing medium, then saved using the big low speed storage medium of cheap, capacity, so as to access frequent " heat " data by the quick interaction of high speed storing media implementation and user.With the migration of data, the data of the big quantitative change of high speed performance layer " cold " are mobile toward low speed capacity layer, and the data that low speed capacity layer becomes " heat " on a small quantity are mobile toward high speed performance layer, the high performance demands in large-scale storage systems to access rate had both been met in this way, also mass data is met to the large capacity demand of storage medium, while effectively reducing the total cost of ownership of system.
Description
Technical field
The present invention relates to technical field of distributed memory, and in particular to a kind of attenuation type based on distributed memory system point
Layer storage system and method.
Background technique
In information storage system, data have apparent life cycle characteristic, and timeliness is stronger.In large-scale data center
In, the total capacity of system is generally several PB, and when multichannel data continues to flow into storage system, most of data are in initial rank
Section is more active, and the access times of data just gradually decrease after active period, or even rarely have use, but still needs long-term protect
It deposits.
In existing persistent storage medium, the solid state hard disk performance using flash chip as memory bank is higher, whether
The connected reference characteristic or random access property of data, being compared to mechanical hard disk has apparent advantage.On the one hand, Gu
State hard disk performance advantage outstanding is that numerous low capacity equipment bring brilliant performance experience；On the other hand, high to deposit
Storage cost allows many large-scale storage systems to hang back.It is directly deposited using traditional approach building high-performance if not optimized
Large-scale storage systems for making in a distributed manner are brought the hardware spending of great number by reservoir.Therefore, existing Mass storage system
In system there are high-performance, large capacity storage demand and low storage overhead between contradiction.
Summary of the invention
(1) technical problems to be solved
The technical problem to be solved by the present invention is how to propose a kind of for distributed memory system while meet high property
The attenuation type tiered storage system and method based on distributed memory system of energy, large capacity demand.
(2) technical solution
In order to solve the above-mentioned technical problems, the present invention provides a kind of, and the attenuation type layering based on distributed memory system is deposited
The storage of hierarchically number of storage system, the distributed memory system includes at least 2 layers, and wherein peak performance layer is known as first performance
Layer, secondary high-performance layer are known as the second performance layer, and so on；It include based on distributed storage between every two performance layer
Two modules of the attenuation type tiered storage system of system: temperature identification module and migration control module；
The temperature identification module is used to the data of different access frequency carrying out decay statistic based on accessing time sequence,
To which user data to be carried out to the differentiation of different temperatures；The data that the migration control module is used to control each storage interlayer are moved
It moves, so that the data of different temperatures are stored into the accumulation layer of corresponding performance.
Preferably, the temperature identification module is specifically used for counting the object fragment of deposit distributed memory system
Label, statistical content include access times and the access time of object fragment, the access of the access times characterization object fragment
Frequency, access time characterize the recency of object fragment；The temperature of each object fragment is by being weighted system to access frequency
Meter, and decay calculation is carried out according to the sequence of access time and is obtained；For the hot statistics and weighted calculation of each object fragment
It is carried out in a measurement period, when exceeding measurement period, the weight of object fragment will be set to 0；In measurement period, when
Between be divided into multiple timeslices, i.e., sub- measurement period, the weight of every sub- measurement period is from current point in time to measurement period
End gradually decays to 0；The weight of the access times of fragment and current sub- measurement period, accumulation calculate summation according to each object
And determine the temperature of object fragment.
Preferably, the migration control module, which is realized, carries out migration management to the object fragment of different temperatures；Data have
Two migratory directions, one is migrated from high-performance layer to low performance layer, the other is migrating from low performance layer to high-performance layer；
It is available that there are two types of the preferential storage locations of data, i.e., high priority data is stored to high-performance layer, and pending data migrates after turning cold
To low performance layer；Or high priority data is stored to low performance layer, is stored after pending data heating to high-performance layer；The migration of data
It can occur to may also occur between non-conterminous accumulation layer between any two adjacent accumulation layers.
Preferably, when realizing migration management, the migration of data can occur in any two phase the migration control module
Between adjacent accumulation layer, it may also occur between non-conterminous accumulation layer.
The present invention also provides a kind of attenuation type layerings based on distributed memory system realized using the system
Storage method, comprising the following steps: realize temperature identification step using the temperature identification module: by the number of different access frequency
Decay statistic is carried out based on accessing time sequence accordingly, so that user data to be carried out to the differentiation of different temperatures；Utilize the migration
Control module realizes migration rate-determining steps: the Data Migration of each storage interlayer of control, so that the data of different temperatures are stored to right
It answers in the accumulation layer of performance.
Preferably, in temperature identification step, the weighted cumulative and formula of object fragment hot statistics are as follows:
Wherein H indicates the hot statistics value of existing object fragment, and n indicates the sub- measurement period for including in a measurement period
Number, i indicates the serial number of sub- measurement period, and the sub- measurement period sequence number value i closer to current point in time is smaller, fiExpression pair
Answer the object fragment access count under sub- measurement period.
Preferably, the migration strategy in the migration rate-determining steps includes when the space accounting of object fragment in accumulation layer is small
When lowest capacity limitation, decaying aging only is carried out to object fragment, without migrating downward into it.
Preferably, the migration strategy in the migration rate-determining steps further includes the space accounting when object fragment in accumulation layer
When between lowest capacity limitation and peak capacity limitation, the object fragment that hot statistics value is 0 is only migrated.
Preferably, the migration strategy in the migration rate-determining steps further includes the space accounting when object fragment in accumulation layer
Greater than peak capacity limitation when, priority migration hot statistics value be 0 object fragment, if migration after object fragment space
Accounting is still greater than peak capacity limitation, then executes Data Migration with hot statistics value ascending order, until object fragment accounting is maintained at
Between lowest capacity limitation and peak capacity limitation.
Preferably, the migration strategy in the migration rate-determining steps further includes depositing in the migration control of data when each
When data in reservoir are less than lowest capacity limitation, stop Data Migration.
(3) beneficial effect
The present invention is marked the data for the system that flows into different temperatures using the timeliness of data in information system
Know, largely no longer accessed " cold " Data Migration is gone out into valuable high speed storing medium, then using cheap, capacity is big
Low speed storage medium saved, so as to access frequent " heat " data by the quick of high speed storing media implementation and user
Interaction.With the migration of data, the data of the big quantitative change of high speed performance layer " cold " are mobile toward low speed capacity layer, and low speed capacity layer is few
The data of quantitative change " heat " are mobile toward high speed performance layer, both met the high-performance in large-scale storage systems to access rate in this way
Demand also meets mass data to the large capacity demand of storage medium, while effectively reducing the total cost of ownership of system
(TCO)。
Detailed description of the invention
Fig. 1 is that data hierarchy of the invention stores schematic diagram；
Fig. 2 is the hot statistics schematic diagram in temperature identification step of the invention in a measurement period；
Fig. 3 is the weight decaying schematic diagram of object fragment in temperature identification step of the invention；
Fig. 4 is migrating data accounting schematic diagram in migration rate-determining steps of the invention.
Specific embodiment
To keep the purpose of the present invention, content and advantage clearer, with reference to the accompanying drawings and examples, to of the invention
Specific embodiment is described in further detail.
The present invention for distributed memory system propose it is a kind of and meanwhile meet high-performance, large capacity demand based on distribution
The attenuation type tiered storage system and method for formula storage system.The design is intended to will be different using a kind of statistical method of attenuation type
Access frequency, different access timing user data bedding storage into the medium of different performance, be finally reached high utilization rate
Data are stored to high-performance layer, and the data of low utilization rate are stored to the effect of low performance layer, to promote the access of user data
Speed, the space utilization rate for optimizing layered medium.
Traditional magnetic disk array stores in the AUTOMATIC ZONING technology of product, the behavioral statistics of data access and analysis granularity be with
Data block is unit, in the distributed memory system that the present invention describes, the behavioral statistics of data access, analysis granularity sum number
It is as unit of object fragment according to migration granularity, typical object data fragment size is 4MB, but is not limited to 4MB.In this hair
In the method for bright proposition, the accumulation layer of system is constructed based on the storage pool of system in a distributed manner, the storage of hierarchically of system
Number includes but is not limited to 2 layers.Wherein peak performance layer is known as primary ergosphere, and secondary high-performance layer is known as the second performance layer, with this
Analogize, may include multiple performance layers in system.It include the decaying based on distributed memory system between every two performance layer
Two modules of formula tiered storage system: temperature identification module and migration control module, as shown in Figure 1.
Temperature identification module is used to the data of different access frequency carrying out decay statistic based on accessing time sequence, thus
User data is carried out to the differentiation of different temperatures, which is specifically used for the object to deposit distributed memory system
Fragment carries out statistics label, and statistical content includes access times and the access time of object fragment, the access times characterization pair
As the access frequency (frequency) of fragment, access time characterizes the recency (recency) of object fragment；Each object point
The temperature of piece carries out decay calculation according to the sequence of access time and obtains by being weighted statistics to access frequency；System
Hot statistics and weighted calculation for each object fragment carry out in a measurement period, when exceeding measurement period, object
The weight of fragment will be set to 0；In measurement period, the time is divided into multiple timeslices, i.e., sub- measurement period, every height
The weight of measurement period gradually decays to 0 from current point in time to measurement period end；The access time of fragment according to each object
The weight of several and current sub- measurement period, accumulation calculate the temperature summed and determine object fragment.
Migration control module is used to control the Data Migration of each storage interlayer, so that the data of different temperatures are stored to correspondence
In the accumulation layer of performance, migration control module, which is realized, carries out migration management to the object fragment of different temperatures；There are two data
Migratory direction, one is migrated from high-performance layer to low performance layer, the other is migrating from low performance layer to high-performance layer；Data
Preferential storage location there are two types of available, i.e., high priority data is stored to high-performance layer, and pending data migrates after turning cold to low
Performance layer；Or high priority data is stored to low performance layer, is stored after pending data heating to high-performance layer；The migration of data can be sent out
It is raw to may also occur between non-conterminous accumulation layer between any two adjacent accumulation layers, for example, if high priority data stores
Into high-performance layer, then after the data heating in arbitrarily low performance layer can Direct Transfer to peak performance layer；In moving for data
It moves in control, when the data in each accumulation layer, which are less than lowest capacity, to be limited, stops Data Migration, it is complete to can avoid data in this way
It is concentrated to lowest performance layer entirely and Large-scale Migration occurs for the data to turn cold in short-term.
Correspondingly, the present invention also provides a kind of attenuation type bedding storage method of corresponding distributed memory system, packets
Include following steps:
Temperature identification step: carrying out decay statistic for the data of different access frequency based on accessing time sequence, thus will
User data carries out the differentiation of different temperatures, in the step, hot statistics using object fragment as basic unit, file store to point
After cloth system, with fixed size cutting for multiple object fragments, typical object fragment size is 4M, but is not limited to 4M,
The size of object fragment can configure on demand, such as from 512K to 256M etc..Each object fragment internal system with object ID into
Line identifier, meanwhile, internal system carries out the statistics of duration to each object fragment.It is united by the access times to object fragment
Meter can determine its access frequency, but this statistics behavior can not be in face of the request of data of burst, and does not have aging mechanism,
It is unable to the access feature of one object fragment of accurate description.
For the temperature of more accurate one object fragment of description, this step joined the attenuating mechanism of data access,
And the concept of measurement period is proposed based on this, when a measurement period can be described as the lasting statistics of one section of regular length
Between.In a measurement period, the time is divided into n parts, and every portion can be described as a sub- measurement period, as shown in Figure 2.According to
The interval of sub- measurement period and current time, every sub- measurement period have been assigned different size of weight, and the weight with
Time elapses and decays, wherein weight accounting is maximum near the sub- measurement period of current point in time.Weight accounting is in time shaft
On generally form a kind of attenuation trend.
Each object fragment may be accessed repeatedly in a sub- measurement period, it is also possible to and it is not visited, it is more accurate
One object fragment of expression hot statistics whithin a period of time, the present invention is by every sub- measurement period of each object fragment
Regular weighted cumulative is carried out, can both take into account the previous access frequency of a period of time interior object fragment in this way to current temperature
Contribution, and be unlikely to make influence of this access frequency to current temperature excessive.The weighted cumulative of object fragment hot statistics and
Formula is as follows:
Wherein H indicates the hot statistics value of existing object fragment, and n indicates the sub- measurement period for including in a measurement period
Number, i indicates the serial number of sub- measurement period, and the sub- measurement period sequence number value i closer to current point in time is smaller, fiExpression pair
Answer the object fragment access count under sub- measurement period.
Over time, weight accounting of the object fragment in every sub- measurement period will gradually decay, such as
Shown in Fig. 3, when certain sub- measurement period corresponding to the access times of object fragment exceeds the range of measurement period, weight
Value will decay to 0, no matter corresponding access times are how many, will no longer be included in the range of hot statistics.
By the setting to measurement period and sub- measurement period, system can effectively smoothed data burst flow,
Usual measurement period is longer, higher to the smoothness of data burst.To establish accurate hot statistics model, user needs root
According to service feature, the window times such as suitable measurement period, preheating cycle are selected.When storage pool is just created or system
When rigid electrifying startup, preheating cycle comes into force.System only carries out the hot statistics of object fragment in this period, without number
According to migration.The configuration of preheating cycle, for obtaining sufficient statistical sample, to establish relatively reliable model and to guarantee migration
Reliability.The duration of preheating cycle is usually less than measurement period.
Migrate rate-determining steps: the Data Migration of each storage interlayer of control, so that the data of different temperatures are stored to correspondence
Can accumulation layer in, realize and migration management carried out to the object fragments of different temperatures, in the step, there are two migration sides for data
To one is migrated from high-performance layer to low performance layer, the other is migrating from low performance layer to high-performance layer.Work as high-performance layer
In object fragment turn cold over time and gradually after, migrated to the accumulation layer of level-one lower than currently stored layer performance
In, after object fragment heating over time and gradually in low performance layer, migrated to higher than currently stored layer performance
In the accumulation layer of level-one, in contrast high-performance layer and low performance layer here is.
Before Data Migration, the preferential storage location of data is available, and respectively high-performance layer is preferential and low performance
Layer is preferential.If selecting high-performance layer preferential, first access data are stored to peak performance layer, will gradually be turned cold over time
Object fragment to the accumulation layer of the low level-one of performance migrate；If selecting low performance layer preferential, first access data are stored to most
Low performance layer over time migrates the object fragment of gradually heating to one or more levels high accumulation layer of performance.Finally, most
Cold data are located at lowest performance layer, and most hot data are located at peak performance layer.
The migration of data can occur between two adjacent accumulation layers, can also across non-conterminous accumulation layer it
Between.For example, the data positioned at lowest performance layer are accessed frequently, then lowest performance layer when selecting high-performance layer preferentially to store
Object fragment can reach peak performance layer across other performance layers.Certainly, the Data Migration proposed in this method can select
It selects across performance layer, can also strictly limit its migration circle, i.e., can only be migrated between adjacent performance layer.
The present invention is that each performance layer sets lowest capacity limitation and peak capacity limitation, which can be used default value,
It can also reset according to demand；After the hot statistics value for obtaining each object fragment, to each object fragment with hot statistics
Value carries out descending sort, is denoted as H1,H2,…,Hn, wherein H1>H2, according to the ranking results, can choose the object fragment of migration.
As shown in figure 4, the volume space of each performance layer is provided with lowest capacity limitation and peak capacity limitation, used in figure
L and H is indicated.In order to describe the migration control of data, is preferentially stored be illustrated with high-performance layer here.As time goes by,
The object fragment quantity of high-performance layer gradually increases, and partial objects fragment is final accumulative since the weight of sub- measurement period decays
Hot statistics value may be 0, statistical value be 0 object fragment will migrate first.
Migration strategy in migration rate-determining steps is described as follows:
When the space accounting of object fragment in accumulation layer, which is less than lowest capacity, to be limited, system only declines to object fragment
Subtract aging, without migrating downward into it；
When the space accounting of object fragment in accumulation layer limits between peak capacity limitation between lowest capacity, system
The only object fragment that migration hot statistics value is 0；
When the space accounting of object fragment in accumulation layer, which is greater than peak capacity, to be limited, System Priority migrates hot statistics value
For 0 object fragment, if the space accounting of object fragment is still greater than peak capacity limitation after migration, with hot statistics value
Ascending order executes Data Migration, until object fragment accounting is maintained between lowest capacity limitation and peak capacity limitation.
In the migration strategy of this step description, when the space accounting of accumulation layer object fragment is less than when lowest capacity limits not
It is migrated, can make to maintain a certain amount of data persistence in storage system in each layering, be unlikely to when system is a certain
Period does not have the data that system stores when user's access to concentrate entirely on lowest performance layer, can avoid becoming in the system short time yet
Large-scale migration occurs for cold data.
As can be seen that timeliness of the present invention using data in information system, by the data for the system that flows into different heat
Degree is identified, and largely no longer accessed " cold " Data Migration is gone out valuable high speed storing medium, then uses price low
The low speed storage medium honest and clean, capacity is big is saved so that access frequent " heat " data by high speed storing media implementation and
The quick interaction of user.With the migration of data, the data of the big quantitative change of high speed performance layer " cold " are mobile toward low speed capacity layer, and low
The data that fast capacity layer becomes " heat " on a small quantity are mobile toward high speed performance layer, both met in large-scale storage systems in this way to access speed
The high performance demands of rate also meet mass data to the large capacity demand of storage medium, while effectively reducing the total of system
Possess cost.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the technical principles of the invention, several improvement and deformations can also be made, these improvement and deformations
Also it should be regarded as protection scope of the present invention.
Claims (10)
1. a kind of attenuation type tiered storage system based on distributed memory system, which is characterized in that the distributed storage system
The storage of hierarchically number of system includes at least two layers, and wherein peak performance layer is known as primary ergosphere, and secondary high-performance layer is referred to as secondary
Ergosphere, and so on；It include the attenuation type tiered storage system based on distributed memory system between every two performance layer
Two modules: temperature identification module and migration control module；
The temperature identification module is used to the data of different access frequency carrying out decay statistic based on accessing time sequence, thus
User data is carried out to the differentiation of different temperatures；The migration control module is used to control the Data Migration of each storage interlayer, with
Store the data of different temperatures into the accumulation layer of corresponding performance.
2. the system as claimed in claim 1, which is characterized in that the temperature identification module is specifically used for depositing deposit distribution
The object fragment of storage system carries out statistics label, and statistical content includes access times and the access time of object fragment, the visit
Ask the access frequency of number characterization object fragment, access time characterizes the recency of object fragment；The temperature of each object fragment
By being weighted statistics to access frequency, and decay calculation is carried out according to the sequence of access time and is obtained；For each object
The hot statistics and weighted calculation of fragment carry out in a measurement period, and when exceeding measurement period, the weight of object fragment will
It is set to 0；In measurement period, the time is divided into multiple timeslices, i.e., sub- measurement period, the power of every sub- measurement period
Weight gradually decays to 0 from current point in time to measurement period end；The access times of fragment and current son system according to each object
The weight in period is counted, accumulation calculates the temperature summed and determine object fragment.
3. the system as claimed in claim 1, which is characterized in that the migration control module realizes the object to different temperatures
Fragment carries out migration management；Data there are two migratory direction, one be migrated from high-performance layer to low performance layer, the other is from
Low performance layer is migrated to high-performance layer；The preferential storage locations of data is corresponding there are two types of available, i.e., high priority data store to
High-performance layer, pending data migrate after turning cold to low performance layer；Or high priority data is stored to low performance layer, pending data heating it
After store to high-performance layer.
4. system as claimed in claim 3, which is characterized in that the migration control module is when realizing migration management, data
Migration can occur to may also occur between non-conterminous accumulation layer between any two adjacent accumulation layers.
5. a kind of attenuation type based on distributed memory system realized using system described in any one of any one of claims 1 to 44
Bedding storage method, which comprises the following steps: realize temperature identification step using the temperature identification module: will
The data of different access frequency carry out decay statistic based on accessing time sequence, so that user data to be carried out to the area of different temperatures
Point；Migration rate-determining steps: the Data Migration of each storage interlayer of control are realized using the migration control module, so that different temperatures
Data store into the accumulation layer of corresponding performance.
6. method as claimed in claim 5, which is characterized in that in temperature identification step, object fragment hot statistics add
Power is accumulative and formula is as follows:
Wherein H indicates the hot statistics value of existing object fragment, and n indicates for the sub- measurement period for including in a measurement period
Number, i indicate the serial number of sub- measurement period, and the sub- measurement period sequence number value i closer to current point in time is smaller, fiIndicate corresponding son
Object fragment access count under measurement period.
7. method as claimed in claim 6, which is characterized in that the migration strategy in the migration rate-determining steps includes when storage
Layer in object fragment space accounting be less than lowest capacity limit when, only decaying aging is carried out to object fragment, without to its into
Row migrates downward into.
8. the method for claim 7, which is characterized in that the migration strategy in the migration rate-determining steps further includes working as to deposit
When the space accounting of object fragment is limited between peak capacity limitation between lowest capacity in reservoir, hot statistics value is only migrated
For 0 object fragment.
9. method according to claim 8, which is characterized in that the migration strategy in the migration rate-determining steps further includes working as to deposit
When the space accounting of object fragment is greater than peak capacity limitation in reservoir, the object fragment that priority migration hot statistics value is 0, if
The space accounting of object fragment is still greater than peak capacity limitation after migration, then executes data with hot statistics value ascending order and move
It moves, until object fragment accounting is maintained between lowest capacity limitation and peak capacity limitation.
10. method as claimed in claim 9, which is characterized in that it is described migration rate-determining steps in migration strategy further include
In the migration control of data, when the data in each accumulation layer, which are less than lowest capacity, to be limited, stop Data Migration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910448847.9A CN110162273A (en) | 2019-05-28 | 2019-05-28 | A kind of attenuation type tiered storage system and method based on distributed memory system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910448847.9A CN110162273A (en) | 2019-05-28 | 2019-05-28 | A kind of attenuation type tiered storage system and method based on distributed memory system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110162273A true CN110162273A (en) | 2019-08-23 |
Family
ID=67629228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910448847.9A Pending CN110162273A (en) | 2019-05-28 | 2019-05-28 | A kind of attenuation type tiered storage system and method based on distributed memory system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110162273A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113377287A (en) * | 2021-04-26 | 2021-09-10 | 深圳云宣科技有限公司 | Big data multidimensional intelligent storage system and method |
| CN113946274A (en) * | 2020-07-15 | 2022-01-18 | 浙江宇视科技有限公司 | Data processing method, device, equipment and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101697526A (en) * | 2009-10-10 | 2010-04-21 | 中国科学技术大学 | Method and system for load balancing of metadata management in distributed file system |
| CN103914516A (en) * | 2014-02-25 | 2014-07-09 | 深圳市中博科创信息技术有限公司 | Method and system for layer-management of storage system |
| CN104102454A (en) * | 2013-04-07 | 2014-10-15 | 杭州信核数据科技有限公司 | Method for automatically realizing hierarchical storage and system for managing hierarchical storage |
| CN104978362A (en) * | 2014-04-11 | 2015-10-14 | 中兴通讯股份有限公司 | Data migration method of distributive file system, data migration device of distributive file system and metadata server |
| CN105653642A (en) * | 2015-12-25 | 2016-06-08 | 曙光信息产业股份有限公司 | Data management method and device of distributed file system |
| CN106502576A (en) * | 2015-09-06 | 2017-03-15 | 中兴通讯股份有限公司 | Migration strategy method of adjustment, capacity change suggesting method and device |
| CN108846114A (en) * | 2018-06-26 | 2018-11-20 | 郑州云海信息技术有限公司 | Distributed system control method, device, equipment and readable storage medium storing program for executing |
| CN108874311A (en) * | 2018-05-29 | 2018-11-23 | 北京盛和大地数据科技有限公司 | Merge the data migration method and device in storage system |
| CN109144895A (en) * | 2017-06-15 | 2019-01-04 | 中兴通讯股份有限公司 | A kind of date storage method and device |
-
2019
- 2019-05-28 CN CN201910448847.9A patent/CN110162273A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101697526A (en) * | 2009-10-10 | 2010-04-21 | 中国科学技术大学 | Method and system for load balancing of metadata management in distributed file system |
| CN104102454A (en) * | 2013-04-07 | 2014-10-15 | 杭州信核数据科技有限公司 | Method for automatically realizing hierarchical storage and system for managing hierarchical storage |
| CN103914516A (en) * | 2014-02-25 | 2014-07-09 | 深圳市中博科创信息技术有限公司 | Method and system for layer-management of storage system |
| CN104978362A (en) * | 2014-04-11 | 2015-10-14 | 中兴通讯股份有限公司 | Data migration method of distributive file system, data migration device of distributive file system and metadata server |
| CN106502576A (en) * | 2015-09-06 | 2017-03-15 | 中兴通讯股份有限公司 | Migration strategy method of adjustment, capacity change suggesting method and device |
| CN105653642A (en) * | 2015-12-25 | 2016-06-08 | 曙光信息产业股份有限公司 | Data management method and device of distributed file system |
| CN109144895A (en) * | 2017-06-15 | 2019-01-04 | 中兴通讯股份有限公司 | A kind of date storage method and device |
| CN108874311A (en) * | 2018-05-29 | 2018-11-23 | 北京盛和大地数据科技有限公司 | Merge the data migration method and device in storage system |
| CN108846114A (en) * | 2018-06-26 | 2018-11-20 | 郑州云海信息技术有限公司 | Distributed system control method, device, equipment and readable storage medium storing program for executing |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113946274A (en) * | 2020-07-15 | 2022-01-18 | 浙江宇视科技有限公司 | Data processing method, device, equipment and medium |
| CN113946274B (en) * | 2020-07-15 | 2024-02-02 | 浙江宇视科技有限公司 | Data processing method, device, equipment and medium |
| CN113377287A (en) * | 2021-04-26 | 2021-09-10 | 深圳云宣科技有限公司 | Big data multidimensional intelligent storage system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105653591B (en) | A kind of industrial real-time data classification storage and moving method | |
| CN110134514B (en) | Extensible memory object storage system based on heterogeneous memory | |
| US9940024B1 (en) | Techniques for determining workload skew | |
| CN104702625B (en) | Access request dispatching method and device in a kind of CDN network | |
| US9753987B1 (en) | Identifying groups of similar data portions | |
| US9052830B1 (en) | Techniques for automated evaluation and movement of data between storage tiers for thin devices | |
| CN103425756B (en) | The replication strategy of data block in a kind of HDFS | |
| CN104699424B (en) | A kind of isomery EMS memory management process based on page temperature | |
| CN111367469B (en) | Method and system for migrating layered storage data | |
| CN105068940B (en) | A kind of adaptive page strategy based on Bank divisions determines method | |
| CN103761051B (en) | One flows concurrent write performance optimization method based on duration data Multiinputoutput | |
| CN106709068B (en) | Hot spot data identification method and device | |
| CN110162273A (en) | A kind of attenuation type tiered storage system and method based on distributed memory system | |
| US20170285973A1 (en) | Managing ssd wear rate in hybrid storage arrays | |
| CN109358821A (en) | A kind of cold and hot data store optimization method of cloud computing of cost driving | |
| US9619169B1 (en) | Managing data activity information for data migration in data storage systems | |
| CN109976917A (en) | A kind of load dispatching method, device, load dispatcher, storage medium and system | |
| CN106156283B (en) | Heterogeneous Hadoop storage method based on data temperature and node performance | |
| CN104102454A (en) | Method for automatically realizing hierarchical storage and system for managing hierarchical storage | |
| CN104035542B (en) | A kind of virtual machine migration method and system calculated with energy consumption for cooling balance | |
| CN103150245B (en) | Determine method and the storage controller of the access characteristics of data entity | |
| CN107122126A (en) | The moving method of data, device and system | |
| CN102129472A (en) | Construction method for high-efficiency hybrid storage structure of semantic-orient search engine | |
| CN107315540A (en) | A kind of AUTOMATIC ZONING storage method and system | |
| CN103136120A (en) | Method and device for determining line buffering management strategies and bank classification method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190823 |