CN105988720B - Data storage device and method - Google Patents
Data storage device and method Download PDFInfo
- Publication number
- CN105988720B CN105988720B CN201510067474.2A CN201510067474A CN105988720B CN 105988720 B CN105988720 B CN 105988720B CN 201510067474 A CN201510067474 A CN 201510067474A CN 105988720 B CN105988720 B CN 105988720B
- Authority
- CN
- China
- Prior art keywords
- data
- layer
- write
- memory
- data layer
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of date storage methods, comprising: after receiving write data requests, is write data into the first data Layer or the second data Layer according to data writing operation degree；After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, and return to query result.The invention also discloses a kind of data storage devices, comprising: processor, the first data Layer and the second data Layer.
Description
Technical field
The present invention relates to data storage technology more particularly to a kind of data storage device and methods.
Background technique
With the development and application of database technology, million (M) words of the data volume of database purchase from the 1980s
Billion (T) byte and peta- (P) byte of section and gigabit (G) byte till now.Meanwhile the query demand of user is also got over
Carry out more complicated, to be related to one or several record being not only in one relation table of inquiry or manipulation, and will be to multiple tables
In the data of ten million item record carry out data analysis and informix, relational database system, which cannot all meet this, to be wanted
It asks.Existing big data processing scheme has using massive parallel processing (MPP, Massively Parallel
Processing) the Database Systems Greenplum of framework, using Hadoop of distributed system infrastructure etc..Also, by
A large amount of X86PC add the framework of local hard drive (HDD, Hard Disk Drive) storage composition large-scale calculations cluster to have become
The main operation platform of most of big data solutions, a large amount of server local hard disk are also main at big data era
Storage mode.
But traditional data is all based on hard-disc storage, management and input and output (I/O), and the development speed of hard disk
Relatively very slowly.From IBM Corporation part of in September, 1956 release in the world first piece of hard disk so far, mainstream enterprise-level hard disk
Revolving speed has stopped many years in 15,000RPM (revolutions per minute), other than capacity can also be continuously increased, other various aspects of performance one
More effective raising can not directly be obtained, it may be said that the performance of hard disk has limited mentioning for systematic entirety energy to a certain extent
It rises.Although enterprise-level hard disk can promote whole data storage and access efficiency by hard disk array, due to hard disk
Data read/write performance can not greatly improve, data transfer latency is too long, cause system data read/write performance demand and
There are larger contradictions between supply.Currently employed hard disk array technology is also difficult to meet under the backgrounds such as big data, cloud computing
System application demand.
It fundamentally sees, the performance of storage system depends on its memory device relied on.HDD is still main at present
Large capacity, low cost storage device.The research and application of various new technology such as perpendicular recording make the capacity of hard disk in the past
30 years in increase 100,000 times, but the movement speed due to being limited by magnetic head, access delay only improve only 2 times.And
And it further increases the rotation speed of hard disk and can bring energy consumption and temperature problem, therefore the contradiction ten between the capacity and performance of hard disk
Divide protrusion.
With the development of big data era, mobile big data business will have very big development.Local hard drive storage
It will be applied on a large scale.However in most cases, based on server to be locally stored efficiency not high, especially greatly
The I/O processing capacity of capacity HDDs is more insufficient.Although some storage efficiencies can be solved to a certain extent using hard disk array
Problem, but the shortcomings that exist simultaneously cost, physical space management, energy consumption etc..
In recent years, the fast development of solid state hard disk (SSD, Solid State Disk/Drive) is to promote storage system
New major opportunity can be provided.SSD sampling solid-state electronic storage chip array, fast, low in energy consumption with random reading speed,
The advantages that shock resistance is good.Although current SSD, there is also on the high side, capacity is smaller, and exists and write preceding erasing, durability etc. and limit
System, the good complementarity of SSD and HDD remains as and designs large capacity, high-performance, the mixing storage system of low cost provide it is brand-new
Opportunity.
But in the mixing storage system of SSD and HDD, writing preceding erasing and writing service life limitation for SSD writes behaviour to system
It is affected as performance.Also, due to the intrinsic attribute limitation of SSD, there is also cannot update (inplace on the spot by SSD
Updating), the problems such as preceding erasing, erasing service life need to be write.These problems lead in existing mixing storage system that there are write operations
It faces awkward predicament: for the write-in policy of data, if frequently writing into SSD, will affect the service life of SSD, but if
It frequently writes into HDD then and will affect the write performance of system.
Summary of the invention
To solve existing technical problem, an embodiment of the present invention is intended to provide a kind of data storage device and method,
It can be improved data writing operation and read the operating characteristics of data, and reduce carrying cost while guaranteeing data security.
The technical solution of the embodiment of the present invention is achieved in that
A kind of date storage method provided in an embodiment of the present invention, comprising:
After receiving write data requests, write data into the first data Layer or the second data Layer according to data writing operation degree；
After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, and return and look into
Ask result.
In above scheme, first data Layer includes the first nonvolatile memory, and second data Layer includes the
Two nonvolatile memories.
In above scheme, the method further includes being according to data usage frequency algorithm and prediction hit rate algorithm
When analyze, determine write data operational readiness or the reading data operational readiness.
In above scheme, described write data into the first data Layer or the second data Layer includes: using write-back mode
It is write data into described in Write-Back in the first nonvolatile memory or the second nonvolatile memory.
In above scheme, described write data into the first data Layer or the second data Layer includes: by the data with number
It is unit according to block, by directly mapping, data block is written described first and non-volatile deposits by the mapping that is connected of full associative mapping or group
In reservoir or second nonvolatile memory.
In above scheme, the method further includes: before inquiring the second data Layer or the first data Layer, according to reading
Data manipulation degree inquires volatile memory；The application memory space and read/write that the volatile memory is arranged accelerate storage
Space, and the read/write that the volatile memory is arranged accelerates the minimum and maximum available threshold of memory space.
A kind of data storage device provided in an embodiment of the present invention, comprising: processor, the first data Layer and the second data
Layer,
The processor is used for: after receiving write data requests, writing data into the first data Layer according to data writing operation degree
Or second in data Layer；After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, and
Return to query result；
First data Layer and the second data Layer are for storing the data.
In above scheme, first data Layer includes the first nonvolatile memory, and second data Layer includes the
Two nonvolatile memories.
In above scheme, the processor is further used for according to data usage frequency algorithm and predicts hit rate algorithm i.e.
When analyze and determine write data operational readiness or the reading data operational readiness.
In above scheme, described device further comprises: the physics mirror of volatile memory and the volatile memory
Picture；
The processor is further used for before inquiring the second data Layer or the first data Layer, according to reading data operational readiness
The read/write inquired the volatile memory, and the volatile memory is arranged accelerates memory space minimum and maximum available
Threshold value.
In above scheme, the processor is further used for by the data as unit of data block, by directly mapping,
Data block is written first nonvolatile memory or described second non-volatile deposited by the mapping that is connected of full associative mapping or group
In reservoir.
In above scheme, first nonvolatile memory includes hard disk；Second nonvolatile memory includes
Solid state hard disk and/or nonvolatile random access memory；The volatile memory includes random access memory.
Date storage method provided in an embodiment of the present invention, according to data writing operation degree write data into the first data Layer or
In second data Layer, and the first data Layer or the second data Layer are inquired according to data operational readiness is read, writes data behaviour to realize
The shunting made and read data manipulation not only improves data writing operation simultaneously and reads data manipulation performance, and enhances data
The reliability of storage.
Data storage device provided in an embodiment of the present invention, the first nonvolatile memory of setting, second non-volatile are deposited
Reservoir and volatile memory are respectively used to according to writing/reading operational readiness storing data, in this way, not only having comprehensively considered various
Storage media types, data payload, storage system framework, memory capacity, performance, reliability, cost and energy consumption etc. it is a variety of because
The influence of element, and data writing operation and reading are realized using SSD and/or NVRAM as the caching of HDD by cache hierarchy
The shunting of data manipulation can not only reduce carrying cost, additionally it is possible to while improving data writing operation and reading data manipulation performance.
The embodiment of the present invention is protected by the way that minimum and maximum available threshold is arranged for RAM disc (Ramdisk)
Minimum and maximum available quantity is demonstrate,proved, to ensure that application memory space is unaffected.
The embodiment of the present invention is avoided that by the physics mirror image of the setting volatile memory because volatile memory is damaged
Badly lead to not access data.
The embodiment of the present invention by write-back (Write-Back) mode during executing data writing operation to Ramdisk with
HDD is locked, to ensure that the safety of data.
Detailed description of the invention
Fig. 1 is the flow diagram of date storage method provided in an embodiment of the present invention；
Fig. 2 is cache hierarchy schematic diagram provided in an embodiment of the present invention；
Fig. 3 is the flow diagram of data writing operation provided in an embodiment of the present invention；
Fig. 4 is that data provided in an embodiment of the present invention map schematic diagram；
Fig. 5 is the composed structure schematic diagram of data storage device provided in an embodiment of the present invention.
Specific embodiment
In various embodiments disclosed by the invention, a kind of date storage method is described, including receive write data requests
Afterwards, it is write data into the first data Layer or the second data Layer according to data writing operation degree；After receiving read data request, according to reading
Data manipulation degree inquires the second data Layer or the first data Layer, and returns to query result；It reads data manipulation in this way, being able to achieve and writes
The shunting of data manipulation, while can improve and read data manipulation and data writing operation performance.
Below by specific embodiment combination attached drawing, the present invention is described in further detail.
In one embodiment of the invention as shown in Figure 1, disclosed date storage method the following steps are included:
Step S101: write data requests or read data request are received；For write data requests, step 102 is executed, for reading
Request of data executes step 103；
Step S102: it is write data into the first data Layer or the second data Layer according to data writing operation degree；
Wherein, above-mentioned first data Layer may include the first nonvolatile memory, and the second data Layer may include second
Nonvolatile memory；
Also, it writes data into the first data Layer or the second data Layer can be and writes data into first and non-volatile deposit
In reservoir or the second nonvolatile memory；
Above-mentioned first nonvolatile memory or the second nonvolatile memory can be according to the characteristics of various storage mediums
It is selected.Table 1 lists the Character Comparison of various storage mediums:
Table 1
Comprehensively consider various storage media types, data payload listed by table 1, storage system framework, memory capacity, property
The influence of many factors such as energy, reliability, cost and energy consumption, above-mentioned first nonvolatile memory may include HDD, institute
Stating the second nonvolatile memory may include SSD and/or nonvolatile random access memory (NVRAM, Non-Volatile
Random Access Memory)。
In step s 102, it can also be determined according to data usage frequency algorithm and prediction hit rate algorithm instant analysis
Write data operational readiness or the reading data operational readiness.
Wherein, data writing operation degree is the number HotW that data writing operation is executed in the unit time, and the unit time can be
Hour, day or week；Reading data operational readiness is that the number HotR for reading data manipulation is executed in unit memory space.
Wherein, the formula of frequency algorithm are as follows:
Parameter i is data block identifier, and value range is 1~k, and k is the quantity of the memory block of data storage device；Parameter
T indicates data collection cycle, and value can be hour, day or week；Parameter Wblock indicates that each data block is acquired in data
The number of data writing operation is executed in period, value range is 0~1000, and can be more than 1000 number by the value of Wblock
It is hot spot data according to block identification；
Wherein, the formula of hit rate algorithm is predicted are as follows: P=Sn/N；
P is prediction hit rate；Parameter Sn indicates that historical heat data acquire hot spot data set, and value range is preceding N
The most data block of a number for executing data writing operation, i.e. Sn=take top n RANK (HotWi)；
The meter of write data operational readiness is determined according to data usage frequency algorithm and prediction hit rate algorithm instant analysis
Calculate formula are as follows: M=α P+ β HotW；
Parameter M indicates the data for needing to cache, and parameter alpha, the value range of β are 0~1, and meet alpha+beta=1.
Second data are written it is possible to further which data writing operation degree to be met to the frequent hot spot data of M feature
Layer.
Above-mentioned write data into the first data Layer or the second data Layer may include: using described in Write-Back mode
It writes data into the first nonvolatile memory or the second nonvolatile memory.
Above-mentioned write data into can also include: to be in the first data Layer or the second data Layer with data block by the data
Unit, by directly mapping, full associative mapping or group be connected mapping by data block be written first nonvolatile memory or
In second nonvolatile memory.
Step S103: the second data Layer or the first data Layer are inquired according to data operational readiness is read；
Wherein it is possible to determine the reading data according to data usage frequency algorithm and prediction hit rate algorithm instant analysis
Operational readiness；It is possible to further meet the frequency statistics strategy of M feature according to reading data operational readiness setting, the second number is first inquired
Inquire the first data Layer again according to layer；It is further possible to first inquire volatile storage according to the height for reading data operational readiness
Device inquires the second nonvolatile memory if not having result return, if not having result return still, then to inquire first non-volatile
Property memory.
Application memory space can be set in above-mentioned volatile memory and read/write accelerates memory space, wherein volatibility is deposited
Reservoir can be random access memory (RAM, Random Access Memory), and read/write accelerates memory space can be into one
Step is set as Ramdisk, and it is minimum and maximum available to guarantee that minimum and maximum available threshold can be arranged for Ramdisk
Amount, so as to guarantee that application memory space is unaffected.
Also, the physics mirror image of the volatile memory can also be further set, so as to avoid because of volatibility
Memory damages and leads to not access data.
Step S104: query result is returned.
In one embodiment of the invention as shown in Figure 2, cache hierarchy is provided, by the data of Logical Data Layer
A-E is respectively stored in the first data Layer and the second data Layer, the logical address of Logical Data Layer and the first data Layer physical address
It corresponds.
Wherein, the first data Layer can be made of HDD equipment, and the second data Layer can be by SSD and/or NVRAM device structure
At.Whether caching of the SSD and/or NVRAM as HDD after receiving read data request, is first inquired in the second data Layer and has been delayed
The data are deposited, if buffered, access SSD and/or NVRAM device；Otherwise, HDD equipment is accessed.The embodiment of the present invention mentions
The cache hierarchy of confession can be cached most by introducing the second data Layer bigger than random access memory capacity
According to data manipulation performance is read in raising.
Also, it, can be first by the SSD and/or NVRAM device of the second data Layer of data pre-write after receiving write data requests
In, then HDD equipment is written in batch again.Cache hierarchy provided in an embodiment of the present invention, it is more random than HDD equipment by introducing
Higher second data Layer of write performance, can be improved data writing operation performance.
In one embodiment of the invention, cache hierarchy can be tied further using flash cache flashcache
Structure.
Cache hierarchy provided in an embodiment of the present invention has comprehensively considered various storage media types, data payload, has deposited
The influence of many factors such as storage system framework, memory capacity, performance, reliability, cost and energy consumption, by by SSD and/or
Caching of the NVRAM as HDD, realizes data writing operation and reads the shunting of data manipulation, not only improves simultaneously and writes data behaviour
Make and read data manipulation performance, and enhances the reliability of data storage.
In one embodiment of the invention as shown in Figure 3, data writing operation uses Write-Back mode, including following
Step:
Step S301: if the data to be written have been stored in volatile memory and the first nonvolatile memory；
Here, it is described stored can be have stored in Ramdisk and HDD, then, just Ramdisk and HDD are added
It locks (Lock)；
Step S302: it writes data into the second nonvolatile memory；
Here, second nonvolatile memory can be in SSD and/or NVRAM；
Step S303: it by the corresponding data write-in Ramdisk in the second nonvolatile memory, releases on RAMDISK
Lock；
Step S304: by the corresponding data write-in HDD in the second nonvolatile memory, the Lock on HDD is released.
In one embodiment of the invention as shown in Figure 4, data are that unit is stored in storage dress with data block (block)
In setting.Below with reference to Fig. 4 to the mapping of the logical address and physical address of data storage device provided in an embodiment of the present invention do into
The detailed description of one step.
For writing/reading operation requests, require consideration for how to be mapped to the logical address of data accordingly physically
Location, i.e. mapping ruler.Wherein, mapping ruler includes mapping granule, i.e. the most basic list of data block that can identify of storage device
Position, may include one or more data blocks.
Block level mapping is the most thin mapping of mapping granule, can be using the data block of each 4KB as basic mapping
Unit.All writing/reading operation requests are according to start physical address, such as sector number dbn, and transport according to modulo operation, Hash
Scheduling algorithm is calculated, the logical address of data is mapped to the physical address of data storage device.
Block level mapping can use cache hierarchy, and by directly mapping, full associative mapping or group be connected mapping
Data block is mapped in the corresponding physical address of data storage device.
Embodiment shown in Fig. 4 is connected mapping using group, wherein first passes through following formula calculating target group:
Target group=(start sector number/(data block size * group size)) mod (quantity of group)
For example, traversal is searched idle or is labeled as invalid (invalid) in target group 3 after being determined as target group 3
Sector, such as: in target group 3 first free time or labeled as invalid sector as Fig. 4 bend filling shown in.In Fig. 4
The sector of dark color filling is expressed as busy or is labeled as effective (valid), if not finding corresponding sky in target group
Sector not busy or labeled as invalid, then according in the recent period at least using algorithm (Least Recently Used, LRU) replacement mesh
Respective sectors in mark group.
In one embodiment of the invention as shown in Figure 5, the data storage device includes: processor, the first data
Layer and the second data Layer.
The processor is used for: after receiving write data requests, writing data into the first data Layer according to data writing operation degree
Or second in data Layer；After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, and
Return to query result；
First data Layer and the second data Layer are for storing the data.
Above-mentioned first data Layer may include the first nonvolatile memory, and the second data Layer may include second non-volatile
Property memory.
Above-mentioned processor can be further used for according to data usage frequency algorithm and prediction hit rate algorithm instant analysis,
Determine write data operational readiness or the reading data operational readiness.
Above-mentioned apparatus may further include volatile memory, for storing the data；Above-mentioned processor can be into
One step is used for before inquiring the second data Layer or the first data Layer, inquires the volatile storage according to data operational readiness is read
Device.
Above-mentioned data storage device may further include the physics mirror image of the volatile memory；The processor into
One step is for being arranged the minimum and maximum available threshold of the volatile memory.
Above-mentioned processor can be further used for by the data as unit of data block, by directly mapping, complete association
Mapping or group are connected to map and data block are written in first nonvolatile memory or second nonvolatile memory.
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the scope of the present invention.
Claims (10)
1. a kind of date storage method, which is characterized in that this method comprises:
After receiving write data requests, write data into the first data Layer or the second data Layer according to data writing operation degree, specifically
, the second data Layer, then batch the first data Layer of write-in first is written into the data that data writing operation degree reaches frequent hot spot；
After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, specifically, first the
Inquired in two data Layers whether the buffered data, if buffered, access the second data Layer, otherwise, access first number
According to layer, and return to query result；
The method further includes according to data usage frequency algorithm and predicting hit rate algorithm instant analysis, described in determination
Data writing operation degree or the reading data operational readiness M=α P+ β HotWi；
Wherein, parameter alpha, the value range of β are 0~1, and meet alpha+beta=1；
P is to predict hit rate: P=Sn/N, N are the most data block of the number of execution data writing operation, and Sn=takes top n RANK
(HotWi)；HotWiFor frequency algorithm:Wherein parameter i is data block identifier, value range is 1~
K, k are the quantity of the memory block of data storage device；Parameter t indicates that data collection cycle, parameter Wblock indicate each data
Block executes the number of data writing operation in data collection cycle, and value range is 0~1000.
2. the method according to claim 1, wherein first data Layer includes the first non-volatile memories
Device, second data Layer include the second nonvolatile memory.
3. according to the method described in claim 2, it is characterized in that, described write data into the first data Layer or the second data Layer
In include: that the first nonvolatile memory or the second nonvolatile memory are write data into using write-back mode Write-Back
In.
4. according to the method described in claim 2, it is characterized in that, described write data into the first data Layer or the second data Layer
In include: by the data as unit of data block, by directly mapping, the mapping that is connected of full associative mapping or group writes data block
Enter in first nonvolatile memory or second nonvolatile memory.
5. method according to any one of claims 1 to 4, which is characterized in that the method further includes: in inquiry the
Before two data Layers or the first data Layer, volatile memory is inquired according to data operational readiness is read；The volatile storage is set
The application memory space and read/write of device accelerate memory space, and the read/write that the volatile memory is arranged accelerates memory space
Minimum and maximum available threshold.
6. a kind of data storage device, comprising: processor, the first data Layer and the second data Layer, which is characterized in that
The processor is used for: after receiving write data requests, writing data into the first data Layer or the according to data writing operation degree
In two data Layers, specifically, the second data Layer first is written in the data that data writing operation degree reaches frequent hot spot, then write in batches
Enter the first data Layer；
After receiving read data request, the second data Layer or the first data Layer are inquired according to data operational readiness is read, specifically, first the
Inquired in two data Layers whether the buffered data, if buffered, access the second data Layer, otherwise, access first number
According to layer, and return to query result；
The processor is further used for according to data usage frequency algorithm and prediction hit rate algorithm instant analysis determination
Data writing operation degree or the reading data operational readiness M=α P+ β HotWi；
Wherein, parameter alpha, the value range of β are 0~1, and meet alpha+beta=1；
P is to predict hit rate: P=Sn/N, N are the most data block of the number of execution data writing operation, and Sn=takes top n RANK
(HotWi)；HotWiFor frequency algorithm:Wherein parameter i is data block identifier, value range is 1~
K, k are the quantity of the memory block of data storage device；Parameter t indicates that data collection cycle, parameter Wblock indicate each data
Block executes the number of data writing operation in data collection cycle, and value range is 0~1000；
First data Layer and the second data Layer are for storing the data.
7. data storage device according to claim 6, which is characterized in that first data Layer includes first non-volatile
Property memory, second data Layer include the second nonvolatile memory.
8. data storage device according to claim 7, which is characterized in that described device further comprises: volatibility is deposited
The physics mirror image of reservoir and the volatile memory；
The processor is further used for before inquiring the second data Layer or the first data Layer, inquires according to data operational readiness is read
The volatile memory, and the read/write that the volatile memory is arranged accelerates the minimum and maximum available threshold of memory space.
9. data storage device according to claim 7, which is characterized in that the processor is further used for the number
Data block is unit accordingly, by directly mapping, full associative mapping or group be connected mapping by data block be written described first it is non-easily
In the property lost memory or second nonvolatile memory.
10. data storage device according to claim 8 or claim 9, which is characterized in that the first nonvolatile memory packet
Include hard disk；Second nonvolatile memory includes solid state hard disk and/or nonvolatile random access memory；It is described volatile
Property memory includes random access memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510067474.2A CN105988720B (en) | 2015-02-09 | 2015-02-09 | Data storage device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510067474.2A CN105988720B (en) | 2015-02-09 | 2015-02-09 | Data storage device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105988720A CN105988720A (en) | 2016-10-05 |
| CN105988720B true CN105988720B (en) | 2019-07-02 |
Family
ID=57040959
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510067474.2A Active CN105988720B (en) | 2015-02-09 | 2015-02-09 | Data storage device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105988720B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106873903B (en) * | 2016-12-30 | 2020-02-18 | 深圳忆联信息系统有限公司 | Data storage method and device |
| CN106844052A (en) * | 2017-01-22 | 2017-06-13 | 郑州云海信息技术有限公司 | A kind of method and device that fusion cluster is built based on Windows Server |
| CN109491616B (en) * | 2018-11-14 | 2022-05-24 | 三星（中国）半导体有限公司 | Data storage method and device |
| CN113656204B (en) * | 2020-05-12 | 2023-11-21 | 中国移动通信集团浙江有限公司 | Solid state disk management method and device and computing equipment |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101552032B (en) * | 2008-12-12 | 2012-01-18 | 深圳市晶凯电子技术有限公司 | Method and device for constructing a high-speed solid state memory disc by using higher-capacity DRAM to join in flash memory medium management |
| CN101533335B (en) * | 2009-04-09 | 2010-12-08 | 杭州华三通信技术有限公司 | Disk storage system and method for controlling disk storage and device thereof |
| CN102521260B (en) * | 2011-11-18 | 2014-04-02 | 华为技术有限公司 | Data preheating method and device |
| CN102609360B (en) * | 2012-01-12 | 2015-03-25 | 华为技术有限公司 | Data processing method, data processing device and data processing system |
| US20130238851A1 (en) * | 2012-03-07 | 2013-09-12 | Netapp, Inc. | Hybrid storage aggregate block tracking |
| CN102662459A (en) * | 2012-04-22 | 2012-09-12 | 复旦大学 | Method for reducing energy consumption of server by using mixed storage of solid-state drive and mechanical hard disk |
| EP2877928A4 (en) * | 2012-07-24 | 2016-03-23 | Intel Corp | System and method for implementing ssd-based i/o caches |
| CN104268099B (en) * | 2014-08-29 | 2017-06-13 | 浪潮(北京)电子信息产业有限公司 | A kind of method and device for managing reading and writing data |
-
2015
- 2015-02-09 CN CN201510067474.2A patent/CN105988720B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105988720A (en) | 2016-10-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10430084B2 (en) | Multi-tiered memory with different metadata levels | |
| CN107193646B (en) | High-efficiency dynamic page scheduling method based on mixed main memory architecture | |
| US9767032B2 (en) | Systems and methods for cache endurance | |
| US9940261B2 (en) | Zoning of logical to physical data address translation tables with parallelized log list replay | |
| Jiang et al. | S-FTL: An efficient address translation for flash memory by exploiting spatial locality | |
| CN104115134B (en) | For managing the method and system to be conducted interviews to complex data storage device | |
| US9146688B2 (en) | Advanced groomer for storage array | |
| US9195396B2 (en) | Estimating access frequency statistics for storage device | |
| US10019352B2 (en) | Systems and methods for adaptive reserve storage | |
| US10997080B1 (en) | Method and system for address table cache management based on correlation metric of first logical address and second logical address, wherein the correlation metric is incremented and decremented based on receive order of the first logical address and the second logical address | |
| US9959054B1 (en) | Log cleaning and tiering in a log-based data storage system | |
| CN110888600B (en) | Buffer area management method for NAND flash memory | |
| WO2020174428A2 (en) | Cache management of logical-physical translation metadata | |
| JP6711121B2 (en) | Information processing apparatus, cache memory control method, and cache memory control program | |
| CN105988720B (en) | Data storage device and method | |
| Chang et al. | Marching-based wear-leveling for PCM-based storage systems | |
| Kargar et al. | Challenges and future directions for energy, latency, and lifetime improvements in NVMs | |
| Yoon et al. | Self-adaptive filtering algorithm with PCM-based memory storage system | |
| Liu et al. | FLAP: Flash-aware prefetching for improving SSD-based disk cache | |
| JP2020013318A (en) | Database management system, memory management device, database management method, and program | |
| US11663136B2 (en) | Storage capacity recovery source selection | |
| CN114746848B (en) | Cache architecture for storage devices | |
| Koltsidas et al. | Spatial data management over flash memory | |
| Kwon et al. | Fast responsive flash translation layer for smart devices | |
| Han et al. | Hybrid write buffer algorithm for improving performance and endurance of NAND flash storages |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |