CN106227598A - A kind of recovery method of cache resources - Google Patents

A kind of recovery method of cache resources Download PDF

Info

Publication number
CN106227598A
CN106227598A CN201610573894.2A CN201610573894A CN106227598A CN 106227598 A CN106227598 A CN 106227598A CN 201610573894 A CN201610573894 A CN 201610573894A CN 106227598 A CN106227598 A CN 106227598A
Authority
CN
China
Prior art keywords
block
data
cache
page
memory
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
Application number
CN201610573894.2A
Other languages
Chinese (zh)
Inventor
柳增运
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Inspur Electronic Information Industry Co Ltd
Original Assignee
Inspur Electronic Information Industry Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Inspur Electronic Information Industry Co Ltd filed Critical Inspur Electronic Information Industry Co Ltd
Priority to CN201610573894.2A priority Critical patent/CN106227598A/en
Publication of CN106227598A publication Critical patent/CN106227598A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5016Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals the resource being the memory
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5022Mechanisms to release resources

Abstract

The invention discloses the recovery method of a kind of cache resources, described method is for the purpose of expanding buffer memory capacity, internal memory is utilized to build a new caching level between internal memory and disk, and by cache resources reclaimer operation, it is to avoid identical data takies many parts of internal memories.The time overhead that the present invention is brought to reduce cache garbage collection to wipe, utilize caching internal data migration operation to reduce the valid data time overhead that brings of copy and resource is taken by caching, utilize many division operations to wipe while realizing multiple packet, the time efficiency expense repeatedly wiped with minimizing simultaneously.

Description

A kind of recovery method of cache resources
Technical field
The present invention relates to software system technical field, be specifically related to the recovery method of a kind of cache resources.
Background technology
In current cache, solid-state disk is gradually as main flow, due to the performance and price of solid-state disk and capacity all occupy DRAM and Between disk, disposing the mixing storage system of flash memory, disk, logical too small amount of flash memory improves systematic entirety and can give full play to The performance advantage of flash memory and the capacity of disk and cost advantage can realize storage performance and the direct efficient balance of capacity.
Owing to solid-state disk uses strange land update mechanism, physical block not only comprises effective data page and also comprises invalid data Page, so causes the waste of memory space, uses the most fewer and feweri, traditional cache garbage collection simultaneously in free block, is responsible for wiping Except physical block obtains new free block, so can cause and write scale-up problem, copy the most, the garbage reclamation time is the longest, so Ratio is relatively time-consuming, has had a strong impact on caching performance.
Summary of the invention
The technical problem to be solved in the present invention is: the present invention is directed to problem above, it is provided that the recovery side of a kind of cache resources Method.
The technical solution adopted in the present invention is:
The recovery method of a kind of cache resources, described method, for the purpose of expanding buffer memory capacity, utilizes internal memory to build one New caching level between internal memory and disk, the effect making full use of caching promotes the overall performance of storage system, and By cache resources reclaimer operation, it is to avoid identical data takies many parts of internal memories, do not have a large amount of repetition.
Data in caching are a copy rubbish of data in back end storage system mostly, directly delete relevant during recovery Block can't cause loss of data, such that it is able to be wiped free of data in block because reclaiming when solving garbage reclamation to a certain extent and The problems such as amplification of writing caused, reduce garbage reclamation expense.
Valid data in described caching have cold and hot dividing, and dsc data is easier to accessed than cold data, and described method is led to Cross the proportion improving dsc data, improve cache hit rate, when cache garbage collection, dsc data is moved, and by direct for cold data Delete.In tradition cache garbage collection, cold and hot data separation is not done in consideration, the valid data simply needed.Solid-state disk is as slow Deposit, owing to the valid data of storage have a copy in rear end, if therefore directly deleting the integrity having no effect on data And concordance.If but deleted valid data, particularly dsc data in block, subsequent access hit rate could be caused to decline.
Described caching comprises state in four: effective, invalid, clean and " containing dirty pages ", during cache garbage collection operation, containing dirty pages is first copied The block at its place, to clean block, could be wiped by shellfish,
One evaluation points t replaced of each piece of definition in solid state cache dish:
t = c + i ( h + d + m + 1 ) * N
Wherein c quantity of cold data page in being block, i is the quantity of invalid data page in block, and d is the number of block apoplexy involving ZANG-organ data page Amount, m is the quantity of clean page in page, and N is the quantity of data page in block;
Condition is met as follows when cache garbage collection:
1) in block, invalid page and cold page are tried one's best many;
2) in block, dsc data and containing dirty pages are the fewest;
3) in block, clean page is the fewest;
So t value is the biggest, and cache garbage collection cost is the least.
Described method is by using intervalometer, and access time and access times according to data block carry out data record, Specifically comprise the following steps that
S1, startup intervalometer;
S2, the access time obtaining described data block and access times, obtain wear factor ω, wherein:
E is maximum erasable number of times, and for known, e is the most erasable number of times;
S3, check whether described data block is non-hot spot data block, if described data block is not belonging to non-hot spot data block, Then check next data block;If described data block belongs to non-hot spot data block, then perform step S4;
S4, remove described data block;
Whether S5, first piece of memory block of inspection are idle, if the free time, discharge first piece of memory block;If not sky Spare time, then the second memory block data are put into free memory list;
After step S5, after multiple second memory blocks form the first memory block, then discharge the first memory block.
Described method provides replacement block selective factor B s according to evaluation points and wear factor, checks described number by s value According to the most non-hot spot data block of block, wherein:
S=ω 1*t+ ω 2* (1-ω),
ω 1+ ω 2=1
The probability that s is chosen the most greatly to reclaim is the biggest, is typically more than 80% to be then considered hot spot data block in s value.
The invention have the benefit that
The time overhead that the present invention is brought to reduce cache garbage collection to wipe, utilizes caching internal data migration operation Resource is taken by time overhead and caching that minimizing valid data copy brings, utilizes many division operations to realize many simultaneously Wipe while individual packet, the time efficiency expense repeatedly wiped with minimizing.
Present invention, avoiding identical data and take many parts of internal memories, do not have a large amount of repetition;Decrease file system to disk Access, improve the organic efficiency to cache resources, be greatly saved memory headroom, improve the overall profit of memory source By rate;According to solid state cache feature, optimize replacement block choosing strategy, data mode in overall consideration caching, entered by data type Row erasing, reduces excessive data copy, it is to avoid brings cache hit rate to reduce problem in erase process, and is reduced by algorithm slow Deposit abrasion.
Accompanying drawing explanation
Fig. 1 migrates schematic diagram for page status in caching;
Fig. 2 is the Memory recycle implementing procedure figure of caching.
Detailed description of the invention
Below according to accompanying drawing, in conjunction with detailed description of the invention, the present invention is further described:
Embodiment 1:
The recovery method of a kind of cache resources, described method, for the purpose of expanding buffer memory capacity, utilizes internal memory to build one New caching level between internal memory and disk, the effect making full use of caching promotes the overall performance of storage system, and By cache resources reclaimer operation, it is to avoid identical data takies many parts of internal memories, do not have a large amount of repetition.
Data in caching are a copy rubbish of data in back end storage system mostly, directly delete relevant during recovery Block can't cause loss of data, such that it is able to be wiped free of data in block because reclaiming when solving garbage reclamation to a certain extent and The problems such as amplification of writing caused, reduce garbage reclamation expense.
Embodiment 2
On the basis of embodiment 1, the valid data in caching described in the present embodiment have cold and hot dividing, and dsc data is than cold number Accessed according to being easier to, described method, by improving the proportion of dsc data, improves cache hit rate, when cache garbage collection by hot number According to moving, and cold data are directly deleted.In tradition cache garbage collection, cold and hot data separation is not done in consideration, simply needs Valid data.Solid-state disk is as caching, owing to the valid data of storage have a copy in rear end, if the most directly deleted Except the integrity and the concordance that have no effect on data.If but deleted valid data, particularly dsc data in block, could cause follow-up Access hit rate declines.
Embodiment 3
As it is shown in figure 1, on the basis of embodiment 2, caching described in the present embodiment comprises state in four: effective, invalid, Totally with " containing dirty pages ", during cache garbage collection operation, containing dirty pages must copy clean block to, could wipe the block at its place Remove,
One evaluation points t replaced of each piece of definition in solid state cache dish:
t = c + i ( h + d + m + 1 ) * N
Wherein c quantity of cold data page in being block, i is the quantity of invalid data page in block, and d is the number of block apoplexy involving ZANG-organ data page Amount, m is the quantity of clean page in page, and N is the quantity of data page in block;
Condition is met as follows when cache garbage collection:
1) in block, invalid page and cold page are tried one's best many
2) in block, dsc data and containing dirty pages are the fewest
3) in block, clean page is the fewest
So t value is the biggest, and cache garbage collection cost is the least.
Embodiment 4
As in figure 2 it is shown, on the basis of embodiment 3, method described in the present embodiment is by using intervalometer, according to data block The access time and access times carry out data record, specifically comprise the following steps that
S1, startup intervalometer;
S2, the access time obtaining described data block and access times, obtain wear factor ω, wherein:
E is maximum erasable number of times, and for known, e is the most erasable number of times;
S3, check whether described data block is non-hot spot data block, if described data block is not belonging to non-hot spot data block, Then check next data block;If described data block belongs to non-hot spot data block, then perform step S4;
S4, remove described data block;
Whether S5, first piece of memory block of inspection are idle, if the free time, discharge first piece of memory block;If not sky Spare time, then the second memory block data are put into free memory list;
After step S5, after multiple second memory blocks form the first memory block, then discharge the first memory block.
Embodiment 5
On the basis of embodiment 4, method described in the present embodiment provides replacement block choosing according to evaluation points and wear factor Select factor s, check the most non-hot spot data block of described data block by s value, wherein:
S=ω 1*t+ ω 2* (1-ω),
ω 1+ ω 2=1
The probability that s is chosen the most greatly to reclaim is the biggest, is typically more than 80% to be then considered hot spot data block in s value.
Embodiment is merely to illustrate the present invention, and not limitation of the present invention, about the ordinary skill of technical field Personnel, without departing from the spirit and scope of the present invention, it is also possible to make a variety of changes and modification, the most all equivalents Technical scheme fall within scope of the invention, the scope of patent protection of the present invention should be defined by the claims.

Claims (5)

1. the recovery method of a cache resources, it is characterised in that: described method, for the purpose of expanding buffer memory capacity, utilizes internal memory Build a new caching level between internal memory and disk, and by cache resources reclaimer operation, it is to avoid identical data Take many parts of internal memories.
The recovery method of a kind of cache resources the most according to claim 1, it is characterised in that: the significant figure in described caching According to there being cold and hot dividing, described method, by improving the proportion of dsc data, improves cache hit rate, when cache garbage collection by dsc data Move, and cold data are directly deleted.
The recovery method of a kind of cache resources the most according to claim 2, it is characterised in that: described caching comprises in four State: effective, invalid, clean and " containing dirty pages ", during cache garbage collection operation, containing dirty pages first copies clean block to, could be to its place Block wipe, and one evaluation points t replaced of in solid state cache dish each piece definition:
t = c + i ( h + d + m + 1 ) * N
Wherein c quantity of cold data page in being block, i is the quantity of invalid data page in block, and d is the quantity of block apoplexy involving ZANG-organ data page, m For the quantity of page clean in page, N is the quantity of data page in block;
And it is as follows to meet condition when cache garbage collection:
1) in block, invalid page and cold page are tried one's best many;
2) in block, dsc data and containing dirty pages are the fewest;
3) in block, clean page is the fewest;
So t value is the biggest, and cache garbage collection cost is the least.
The recovery method of a kind of cache resources the most according to claim 3, it is characterised in that: described method is fixed by using Time device, access time and access times according to data block carries out data record, specifically comprises the following steps that
S1, startup intervalometer;
S2, the access time obtaining described data block and access times, obtain wear factor ω, wherein:
E is maximum erasable number of times, and for known, e is the most erasable number of times;
S3, check whether described data block is non-hot spot data block, if described data block is not belonging to non-hot spot data block, then examine Look into next data block;If described data block belongs to non-hot spot data block, then perform step S4;
S4, remove described data block;
Whether S5, first piece of memory block of inspection are idle, if the free time, discharge first piece of memory block;If not the free time, then Second memory block data are put into free memory list;
After step S5, after multiple second memory blocks form the first memory block, then discharge the first memory block.
The recovery method of a kind of cache resources the most according to claim 4, it is characterised in that: described method according to evaluate because of Son and wear factor provide replacement block selective factor B s, check the most non-hot spot data block of described data block by s value, wherein:
S=ω 1*t+ ω 2* (1-ω),
ω 1+ ω 2=1
S value is more than 80% to be considered hot spot data block.
CN201610573894.2A 2016-07-20 2016-07-20 A kind of recovery method of cache resources Pending CN106227598A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610573894.2A CN106227598A (en) 2016-07-20 2016-07-20 A kind of recovery method of cache resources

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610573894.2A CN106227598A (en) 2016-07-20 2016-07-20 A kind of recovery method of cache resources

Publications (1)

Publication Number Publication Date
CN106227598A true CN106227598A (en) 2016-12-14

Family

ID=57531523

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610573894.2A Pending CN106227598A (en) 2016-07-20 2016-07-20 A kind of recovery method of cache resources

Country Status (1)

Country Link
CN (1) CN106227598A (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106959898A (en) * 2017-03-27 2017-07-18 广州优视网络科技有限公司 Reduce method, device and the mobile terminal of EMS memory occupation
CN109376094A (en) * 2018-09-27 2019-02-22 郑州云海信息技术有限公司 A kind of method of the cold and hot Data Migration of garbage reclamation in storage system
CN109634915A (en) * 2018-11-28 2019-04-16 深圳市网心科技有限公司 File dispositions method, Cloud Server, system and storage medium
CN110908923A (en) * 2018-09-14 2020-03-24 阿里巴巴集团控股有限公司 Memory recovery method and device
CN111274039A (en) * 2020-02-14 2020-06-12 Oppo广东移动通信有限公司 Memory recovery method and device, storage medium and electronic equipment
WO2021031408A1 (en) * 2019-08-22 2021-02-25 平安科技(深圳)有限公司 Cached data clearing method, device, equipment, and computer-readable storage medium
CN112905129A (en) * 2021-05-06 2021-06-04 蚂蚁金服(杭州)网络技术有限公司 Method and device for eliminating cache memory block and electronic equipment
CN113672166A (en) * 2021-07-08 2021-11-19 锐捷网络股份有限公司 Data processing method and device, electronic equipment and storage medium
WO2023051715A1 (en) * 2021-09-30 2023-04-06 华为技术有限公司 Data processing method and apparatus, processor, and hybrid memory system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102081576A (en) * 2011-03-01 2011-06-01 华中科技大学 Flash memory wear balance method
CN102156753A (en) * 2011-04-29 2011-08-17 中国人民解放军国防科学技术大学 Data page caching method for file system of solid-state hard disc
CN104317735A (en) * 2014-09-24 2015-01-28 北京云巢动脉科技有限公司 High-capacity cache and method for data storage and readout as well as memory allocation and recovery

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102081576A (en) * 2011-03-01 2011-06-01 华中科技大学 Flash memory wear balance method
CN102156753A (en) * 2011-04-29 2011-08-17 中国人民解放军国防科学技术大学 Data page caching method for file system of solid-state hard disc
CN104317735A (en) * 2014-09-24 2015-01-28 北京云巢动脉科技有限公司 High-capacity cache and method for data storage and readout as well as memory allocation and recovery

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
梁俊杰 等: ""以固态硬盘为缓存的存储技术研究"", 《微电子学与计算机》 *
魏登萍 等: ""面向缓存的固态盘垃圾回收策略"", 《第十九届计算机工程与工艺年会暨第五届微处理器技术论坛论文集》 *

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106959898A (en) * 2017-03-27 2017-07-18 广州优视网络科技有限公司 Reduce method, device and the mobile terminal of EMS memory occupation
CN110908923B (en) * 2018-09-14 2023-05-02 阿里巴巴集团控股有限公司 Memory recycling method and device
CN110908923A (en) * 2018-09-14 2020-03-24 阿里巴巴集团控股有限公司 Memory recovery method and device
CN109376094A (en) * 2018-09-27 2019-02-22 郑州云海信息技术有限公司 A kind of method of the cold and hot Data Migration of garbage reclamation in storage system
CN109634915A (en) * 2018-11-28 2019-04-16 深圳市网心科技有限公司 File dispositions method, Cloud Server, system and storage medium
WO2021031408A1 (en) * 2019-08-22 2021-02-25 平安科技(深圳)有限公司 Cached data clearing method, device, equipment, and computer-readable storage medium
CN111274039A (en) * 2020-02-14 2020-06-12 Oppo广东移动通信有限公司 Memory recovery method and device, storage medium and electronic equipment
CN111274039B (en) * 2020-02-14 2023-12-08 Oppo广东移动通信有限公司 Memory recycling method and device, storage medium and electronic equipment
CN112905129A (en) * 2021-05-06 2021-06-04 蚂蚁金服(杭州)网络技术有限公司 Method and device for eliminating cache memory block and electronic equipment
WO2022233272A1 (en) * 2021-05-06 2022-11-10 北京奥星贝斯科技有限公司 Method and apparatus for eliminating cache memory block, and electronic device
CN112905129B (en) * 2021-05-06 2021-08-13 蚂蚁金服(杭州)网络技术有限公司 Method and device for eliminating cache memory block and electronic equipment
CN113672166A (en) * 2021-07-08 2021-11-19 锐捷网络股份有限公司 Data processing method and device, electronic equipment and storage medium
WO2023051715A1 (en) * 2021-09-30 2023-04-06 华为技术有限公司 Data processing method and apparatus, processor, and hybrid memory system

Similar Documents

Publication Publication Date Title
CN106227598A (en) A kind of recovery method of cache resources
CN103136121B (en) Cache management method for solid-state disc
CN102508788B (en) SSD (solid state drive) and SSD garbage collection method and device
CN106484323A (en) A kind of loss equalizing method of solid-state storage and system
CN101169751B (en) System with flash memory device and data recovery method thereof
CN104298610B (en) Data storage system and its management method
CN102981963B (en) A kind of implementation method of flash translation layer (FTL) of solid-state disk
CN103365788B (en) The adaptive local rubbish recovering method that real-time flash memory conversion layer uses
TWI399644B (en) Block management method for a non-volatile memory
CN103455435A (en) Data writing method and device
CN106598878A (en) Method for separating cold data and hot data of solid state disk
CN100501868C (en) Method for realizing memory file system based on NAND Flash
CN104484283B (en) A kind of method for reducing solid state disk write amplification
CN105528301A (en) NAND Flash memory garbage collection method
CN101923448A (en) Method for reading and writing conversion layer of NAND flash memory
CN107092563B (en) Garbage recovery method and device
CN101067800A (en) Memory deterioration balance method based on local sampling
CN106528438A (en) Segmented junk recovery method for solid-state storage device
CN102163175A (en) Hybrid address mapping method based on locality analysis
CN105204783B (en) A kind of solid-state disk rubbish recovering method based on data lifetime
CN102819494B (en) Optimization method when a kind of flash memory is sequentially written in
CN103440207A (en) Caching method and caching device
CN102768645A (en) Solid state disk (SSD) prefetching method for mixed caching and SSD
CN107797772A (en) A kind of garbage retrieving system and method based on flash media
CN107515728A (en) Play the data managing method and device of concurrent characteristic inside flash memory device

Legal Events

Date Code Title Description
C06 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

Application publication date: 20161214

RJ01 Rejection of invention patent application after publication