A method of extending solid-state disk service life

文档序号:1755011 发布日期:2019-11-29 浏览:3次 中文

阅读说明:本技术 一种延长固态硬盘寿命的方法 (A method of extending solid-state disk service life ) 是由 赵琨 张涛 于 2019-06-27 设计创作,主要内容包括:本发明专利公开了一种延长固态硬盘寿命的方法,包括如下步骤:数据的预处理:对于主机写入的数据进行审核,随后对数据进行筛选,最后再对数据进行排序;数据的写入:将进行预处理后的数据写入Flash存储单元,且写入次数达到预设次数后,强制将数据仅在内存中更新,不写回至固态硬盘。本发明专利所采用的方法,可以通过降低磨损引起的误码率升高,延长固态硬盘的使用寿命,解决了目前固态硬盘的寿命往往由于Flash磨损程度导致误码率的提高,进而出现读取干扰、编程干扰和数据保持发生错误情况的问题,通过预处理的数据排序,可避免写入专门序列的繁琐步骤,借助自主的误码率分析,可对判决阈值及时调整,以降低误码率。(The invention patent discloses a kind of method for extending solid-state disk service life, includes the following steps: the pretreatment of data: the data of host write-in being audited, then screens to data, is finally ranked up again to data;After the write-in of data: will carry out pretreated data write-in Flash storage unit, and write-in number reaches preset times, pressure only updates data in memory, is not written back to solid state hard disk.Method used by the invention patent, caused bit error rate raising can be worn by reducing, extend the service life of solid state hard disk, the service life of current solid state hard disk is solved often due to the Flash degree of wear leads to the raising of the bit error rate, and then there are reading interference, programming interference and data and kept for the problem of error situation occurs, pass through pretreated data sorting, it can avoid the tedious steps of write-in dedicated sequences, by autonomous error rate analyzer, decision threshold can be adjusted in time, to reduce the bit error rate.)

1. a kind of method for extending solid-state disk service life, characterized by the following steps:

(1) pretreatment of data: the data of host write-in are audited, then data are screened, finally logarithm again According to being ranked up;

(2) write-in of data: will carry out pretreated data write-in Flash storage unit, and number is written and reaches default time After number, pressure only updates data in memory, is not written back to solid state hard disk, data are compressed or hashed by master controller Processing;

(3) processing of data: available free region is adopted in data to be written self-test and after carrying out the formatting of system data It is filled with invalid random number, then new data is written again after wiping the physical block in real time;

(4) reading of data: when carrying out reading data, data are handled and is returned using the algorithm for inversion of pre-treatment step Host actively reads data by master controller when user reads data or hard disk free time;

(5) Bit Error Ratio Measurement of data: the master controller of solid state hard disk reads Flash storage unit when reading data Error rate of original data is analyzed and is counted;

(6) reduction of error rates of data: the judgement after the analysis to the bit error rate, by master controller to Flash storage unit Threshold value is adjusted, and reduces the degree of wear of Flash.

2. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (1), To there are two types of the optional methods of data prediction, including the symmetric encipherment algorithms such as lossless compression algorithm and AES.

3. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (1), Initial data mainly should be gone to audit in terms of integrality and accuracy two, secondly further include applicability, timeliness and one The audit of secondary property;Data screening includes both sides content: first is that by certain undesirable data or with having apparent error Data are rejected, second is that the data screening for meeting certain specified conditions is come out, are given the data for not meeting specified conditions It rejects;The sides such as lifting sequence, initial, increasing or decreasing are used alternatingly according to different data types in sequence for data Method.

4. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (2), Flash storage unit belongs to multiple field storage flash memory, and the control unit of flash memory reads Flash storage unit Each entity memory block write-in/number of erasing ,/number of erasing is more than to be written/when upper limit value of erasing, entity is deposited when being written Storage area block is converted to single-layer type storage format by multiple field storage format.

5. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (3), When filling random number, corresponding physical block is labeled as executing garbage, and complete to remove the valid data of physical block to be recycled After shifting, erasing operation is not made to the physical block of recycling, is only executing garbage by the physics block mark.

6. a kind of method for extending solid-state disk service life according to claim 5, it is characterised in that: new data are being written When, the physics executing garbage of mark is recycled, the valid data above several Block are moved some Block, are then wiped Then original Block is again written Host data.

7. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (4), Three kinds of data reading modes: 1. preloading data also can be used, such mode is only suitable for small data, usually defines certain in a program Fixed value, such as cycle-index, and be seldom used to read training data;2. generating data, then data are fed for rear end, designed Placeholder is stayed when Graph, transmits data into placeholder when real Run, is fed for rear end training;3. from file It directly reads, the reading of data and calculating is individually placed in two threads, the data of reading save as a team in memory Column, the thread for being responsible for calculating can continuously read data from memory queue.

8. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (5), When analyzing and counting, common statistical distribution method and actual measurement method can be used, actual measurement technology is examined using cyclic redundancy Mode (CRC) is looked into determine the error condition occurred in a period of time.

9. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: in the step (5), In statistics, from the data called in step (1) after the completion of sorting, can be counted at any time, it is special without being written Sequence.

10. a kind of method for extending solid-state disk service life according to claim 1, it is characterised in that: the step (6) In, in the counting module that activates within the continuous period, calculating indicates that two groups of binary signal are respectively provided with the second binary system The difference of algebraic step between numerical value and the bit number of the first binary numeral, at the end in each period, according to the symbol of difference Number be respectively it is just or negative, adjustment module increases or reduces decision threshold by increment.

Technical field

The invention patent relates to solid state hard disk technical field, specially a kind of method for extending solid-state disk service life.

Background technique

Solid state drive, is commonly called as solid state hard disk, and solid state hard disk is manufactured hard with solid-state electronic storage chip array Solid capacitance is referred to as Solid and gained the name in English by disk, and SSD is by control unit and storage unit (FLASH chip, DRAM Chip) composition, solid state hard disk identical with common hard disc, In in the specification and definition, function and application method of interface It is also completely consistent with common hard disc in product shape and size, it is widely used in military, vehicle-mounted, industry control, video monitoring, net The numerous areas such as network monitoring, the network terminal, electric power, medical treatment, aviation, navigation equipment.

SSD solid state hard disk is that performance is strong, small in size, noiseless compared to the advantages of tradition HDD mechanical hard disk, solid state hard disk Structurally substantially there are many similar places, such as simulation sector, simulation magnetic using standard machinery hard disk with us in principle Road etc. is the core of entire solid state hard disk even if most crucial part controller, the inside includes very inside solid state hard disk More frameworks, such as read-write algorithm, interface definition etc., the main influence service life is exactly to read and write number, fixed in the algorithm of solid state hard disk In justice, modification is primary just to calculate primary really read-write, and the service life of solid state hard disk is often due to the Flash degree of wear causes to miss at present The raising of code rate, and then reading interference, programming interference and data occur and keep proposing that one kind can thus there is a situation where mistake To wear caused bit error rate raising by reducing, extend the method for solid state hard disk service life to solve the problems, such as this.

Patent of invention content

The invention patent is designed to provide a kind of method for extending solid-state disk service life, and reduction can be passed through by having The advantages of bit error rate caused by wearing increases, and extends solid state hard disk service life, solves the service life of current solid state hard disk often Since the Flash degree of wear leads to the raising of the bit error rate, and then there are reading interference, programming interference and data and keep that mistake occurs Accidentally the problem of situation.

To achieve the above object, the invention patent provides the following technical solutions: a kind of side extending solid-state disk service life Method includes the following steps:

(1) pretreatment of data: the data of host write-in are audited, then data are screened, finally again Data are ranked up;

(2) write-in of data: pretreated data write-in Flash storage unit will be carried out, and number is written and reaches pre- If after number, pressure only updates data in memory, it is not written back to solid state hard disk, data are compressed by master controller Or hashing;

(3) processing of data: data to be written self-test and after carrying out the formatting of system data, available free Region is filled using invalid random number, then new data is written again after wiping the physical block in real time;

(4) reading of data: when carrying out reading data, data are handled using the algorithm for inversion of pre-treatment step And host is returned, when user reads data or hard disk free time, data are actively read by master controller;

(5) Bit Error Ratio Measurement of data: the master controller of solid state hard disk reads Flash storage unit when reading data Error rate of original data out is analyzed and is counted;

(6) reduction of error rates of data: after the analysis to the bit error rate, by master controller to Flash storage unit Decision threshold is adjusted, and reduces the degree of wear of Flash.

Preferably, in the step (1), to there are two types of the optional methods of data prediction, including lossless compression algorithm and The symmetric encipherment algorithms such as AES.

Preferably, in the step (1), initial data mainly should be gone to examine in terms of integrality and accuracy two Secondly core further includes applicability, timeliness and disposable audit;Data screening includes both sides content: first is that by certain Undesirable data or data are rejected with having apparent error, second is that the data screening that certain specified conditions will be met Out, the data for not meeting specified conditions are rejected;Sequence for data, according to different data types, alternating makes With the methods of lifting sequence, initial, increasing or decreasing.

Preferably, in the step (2), Flash storage unit belongs to multiple field storage flash memory, and flash is deposited The control unit of reservoir reads write-in/number of erasing of each entity memory block of Flash storage unit, secondary when being written/erasing Number is more than that when being written/erase upper limit value, entity memory block is converted to single-layer type storage format by multiple field storage format.

Preferably, in the step (3), when filling random number, by corresponding physical block labeled as executing garbage, and After the valid data of physical block to be recycled are moved in completion, erasing operation is not made to the physical block of recycling, only by the physical block Mark is executing garbage.

Preferably, when new data are written, the physics executing garbage of mark is recycled, having above several Block Data-moving is imitated to some Block, then wipes original Block, then Host data are written again.

Preferably, in the step (4), three kinds of data reading modes: 1. preloading data, such mode also can be used It is only suitable for small data, usually defines certain fixed value, such as cycle-index in a program, and is seldom used to read training data;② Data are generated, then data are fed for rear end, placeholder are stayed when designing Graph, to placeholder when real Run Middle transmitting data are fed for rear end training;3. directly reading from file, the reading of data and calculating are individually placed to two lines Cheng Zhong, the data of reading save as a queue in memory, and the thread for being responsible for calculating can be continuously from memory queue Middle reading data.

Preferably, in the step (5), when analyzing and counting, common statistical distribution method and actual measurement can be used Method, actual measurement technology determine the error condition occurred in a period of time using Cyclical Redundancy Check mode (CRC).

Preferably, in the step (5), in statistics, from the data called in step (1) after the completion of sorting, Ke Yisui Shi Jinhang statistics, without special sequence is written.

Preferably, in the step (6), in the counting module that activates within the continuous period, calculating indicates binary system letter Number two groups of differences for being respectively provided with algebraic step between the second binary numeral and the bit number of the first binary numeral, In The end in each period, the symbol according to difference are just or to bear respectively, and adjustment module increases or reduces judgement by increment Threshold value.

Compared with prior art, the invention patent has the beneficial effect that:

1, method used by the invention patent can wear caused bit error rate raising by reducing, it is hard to extend solid-state The service life of disk, solves the service life of current solid state hard disk often due to the Flash degree of wear leads to the raising of the bit error rate, And then reading interference, programming interference and data occur and kept for the problem of error situation occurs, by pretreated data sorting, The tedious steps that can avoid write-in dedicated sequences can adjust in time decision threshold by autonomous error rate analyzer, with drop Low error rate.

2, the invention patent forces only to update data in memory by system, without writing back in solid state hard disk, can subtract Few erasable number to solid state hard disk, so that the service life of solid state hard disk is effectively improved, when new data is written, by preparatory Wipe the mode of rubbish physical block, it is ensured that most free physical block will not all be in smearing state for a long time, simultaneously When overcoming data write-in again the problem of ability erasing operation bring response delay, recycled by the physics executing garbage that will indicate, So that available Block can be found when data are written to be written, the waiting time is avoided, for as much HOST data, the data being written in FLASH are fewer, also mean that FLASH loss is smaller.

Detailed description of the invention

Fig. 1 is the invention patent flow chart.

Specific embodiment

Below in conjunction with the attached drawing in the invention patent embodiment, to the technical solution in the invention patent embodiment into Row clearly and completely describes, it is clear that and described embodiment is only the invention patent a part of the embodiment, rather than all Embodiment.Based on the embodiment in the invention patent, those of ordinary skill in the art are not making creative work premise Under every other embodiment obtained, belong to the invention patent protection range.

Referring to Fig. 1, a kind of method for extending solid-state disk service life, includes the following steps:

(1) pretreatment of data: the data of host write-in are audited, then data are screened, finally again Data are ranked up;

(2) write-in of data: pretreated data write-in Flash storage unit will be carried out, and number is written and reaches pre- If after number, pressure only updates data in memory, it is not written back to solid state hard disk, data are compressed by master controller Or hashing;

(3) processing of data: data to be written self-test and after carrying out the formatting of system data, available free Region is filled using invalid random number, then new data is written again after wiping the physical block in real time;

(4) reading of data: when carrying out reading data, data are handled using the algorithm for inversion of pre-treatment step And host is returned, when user reads data or hard disk free time, data are actively read by master controller;

(5) Bit Error Ratio Measurement of data: the master controller of solid state hard disk reads Flash storage unit when reading data Error rate of original data out is analyzed and is counted;

(6) reduction of error rates of data: after the analysis to the bit error rate, by master controller to Flash storage unit Decision threshold is adjusted, and reduces the degree of wear of Flash.

10页详细技术资料下载
上一篇:一种医用注射器针头装配设备
下一篇:一种动力锂电池SOC存储方法

网友询问留言

已有0条留言

还没有人留言评论。精彩留言会获得点赞!

精彩留言,会给你点赞!

技术分类