CN104536908B - A kind of magnanimity small records efficient storage management method towards unit - Google Patents
A kind of magnanimity small records efficient storage management method towards unit Download PDFInfo
- Publication number
- CN104536908B CN104536908B CN201410617251.4A CN201410617251A CN104536908B CN 104536908 B CN104536908 B CN 104536908B CN 201410617251 A CN201410617251 A CN 201410617251A CN 104536908 B CN104536908 B CN 104536908B
- Authority
- CN
- China
- Prior art keywords
- file
- tree
- data
- cpath
- record
- 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.)
- Expired - Fee Related
Links
- 238000007726 management method Methods 0.000 title claims abstract description 25
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000000638 solvent extraction Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims description 14
- 230000008520 organization Effects 0.000 claims description 3
- 241000238876 Acari Species 0.000 claims 1
- 230000003139 buffering effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 238000013500 data storage Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- RTZKZFJDLAIYFH-UHFFFAOYSA-N Diethyl ether Chemical compound CCOCC RTZKZFJDLAIYFH-UHFFFAOYSA-N 0.000 description 2
- 238000012550 audit Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
Abstract
The present invention relates to a kind of magnanimity small records efficient storage management method towards unit, belong to technical field of information storage.Including：(1) data receiver and caching：Small records data are received, and are stored in core buffer；(2) automatic Partitioning：When core buffer is full, the data in buffering area are written in file, to prevent the too big influence performance of single file, each file has size limitation, when single file reaches piecemeal size, automatic new files；(3) tree-like directory generates：Performance is influenceed too much to prevent from working as file under a catalogue, limits the total number of files below each file, the tree-shaped tissue of file and blocked file is realized with simplified tree management algorithm；(4) record identification：The physical address of record, the access of convenient record are directly recorded using special identifier.Prior art is contrasted, the inventive method can obtain the ability of the management magnanimity small records data of higher performance under stand-alone environment with lower hardware performance requirements.
Description
Technical field
The present invention relates to a kind of memory management method of the magnanimity small records of efficient storage on a single computer, belong to information
Store correlative technology field.
Background technology
With the arriving of the development of Internet technology, especially big data epoch, the numerous areas of informationized society produces
The data of magnanimity, and be in explosive growth.The demand of mass memory faces a severe test, and this causes data storage management skill
Art residing status in information system is also more and more important.
In many application systems, it is often necessary to large-scale log recording is managed under stand-alone environment.Than more typical
Scene is security audit record.This scene has following features：
1) the wall scroll record of this record is generally smaller.Such as the audit log record in information security, it is average every
Bar record only has hundreds of bytes；
2) enormous amount.The daily record quantity that usual unit needs to manage reach tens it is even more more, memory capacity reaches
It is even more big to TB levels；
3) efficiency requirements are high.Very high is required to warehouse-in efficiency, reaches the even more records of the rules up to ten thousand each second.And
Computer process ability under this environment is generally also weaker.
4) data manipulation is simple.Generally data order storage is, it is necessary to support to support random read take.Centre need not be supported
Insertion operation, record renewal frequency are relatively low, it is not necessary to support affairs.
In such a case, using traditional Relational DataBase system either in data storage size, or storage effect
Rate etc. can not all meet to require.And the generation of a large amount of disk fragmentses will be caused by being directly stored in file system.So have
Necessity is developed under a kind of stand-alone environment towards the efficient storage administrative skill of magnanimity small records.
In existing NoSQL research, the storage for mass data uses the management by district technology of multiple machine distributing,
Be not suitable for the targeted stand-alone environment of the present invention.
The content of the invention
It is a primary object of the present invention to can not meet magnanimity short biography address book stored in stand-alone environment for existing storage method
The problem of demand of management, propose a kind of magnanimity small records efficient storage management method towards unit, enabling single
Efficiently received on computer and store large-scale small records.
To reach above-mentioned purpose, the present invention is achieved by the following technical solutions：
A kind of magnanimity small records efficient storage management method towards unit, this method comprise at least following several contents：
A. data receiver and caching：Small records data are received, to avoid frequent I/O from asking and causing disk fragmentses, first
By record storage in core buffer；
B. automatic Partitioning：When the core buffer can not accommodate new record, the data in the core buffer are write
Enter into magnetic disk storage, and empty the core buffer to write the new record of the new reception.To prevent single text
The too big influence I/O performances of part, limit the size threshold value of each data file.If current blocked file size exceedes the threshold
Value, automatic newly-built blocked file, and new data is written in new blocked file；
C. tree-like directory generates：I/O performances are influenceed too much to prevent from working as file under a catalogue, are limited below each file
Total number of files, the tree-shaped tissue of file and blocked file is realized with simplified tree management algorithm；
D. record identification：The physical address each recorded, the access of convenient record are directly recorded using special identifier.
Preferably：
The size of the core buffer can configure.
The block file dynamic generative process comprises the following steps：
If step 1. obtains work at present catalogue without current blocked file cFile by tree-like directory generating algorithm
CPath, blocked file is generated in cPath, using the block file as current block file cFile, terminated；
If step 2. cFile remaining spaces are enough the data for writing the core buffer, terminate；
It is raw in current directory if block file sum is not above predetermined threshold in step 3. current directory cPath
Composition block file, and using the block file as work at present block file cFile, terminate；
Step 4. generates new storage catalogue by the tree-like directory generating algorithm set, using new storage catalogue as current mesh
CPath is recorded, block file is generated in cPath, and using the block file as current block file cFile.
The Directory tree structure can use B- trees, B+ trees or Adelson-Velskii-Landis tree structure；The exponent number of Directory tree structure, it is single
The quantity of file under file, the quantity of block file and size are configurable.
Described tree-like directory generating algorithm can use simplified B- trees, B+ trees or Adelson-Velskii-Landis tree generating algorithm.Simplified calculation
Method makes full use of the feature that data order is inserted, and avoids the rotation process of tree.
The each record is added a mark, and the mark is made up of blocked file numbering and block bias internal.
Blocked file path can be navigated to according to described blocked file mark, can be navigated to and be recorded according to described piece of bias internal
Original position inside block file.So as to can directly position the physical address of record by the record identification, convenient record
Random access.
Beneficial effect
Method provided by the invention has following advantage：
1) higher performance.The present invention proposes a kind of memory management method that magnanimity small records are realized based on unit.Profit
The tree construction of automatic growth is realized with the document directory structure of operating system, realized on this tree construction dynamic data storage and
Safeguard, constant high writing speed can be reached, while reduce requirement of the storage to hardware performance.
2) it is easier management mass data.By auto plate separation and automatic directory management, number can be very easily realized
According to backup, the management such as deletion.
3) ability of complex data type is preferably handled.The present invention can be according to the difference of user's application field, with reality
Border is needed to be oriented to the field of every record of customization, so as to enhance the ability of present invention processing complex data type.
Brief description of the drawings
Fig. 1 is the tree-shaped storage organization schematic diagram of the embodiment of the present invention；
Fig. 2 is the workflow schematic diagram of the embodiment of the present invention；
Fig. 3 is the data buffer area schematic diagram of the embodiment of the present invention；
Fig. 4 is the data buffer storage flow chart of the embodiment of the present invention；
Fig. 5 is the block file product process figure of the embodiment of the present invention；
Fig. 6 is the B- trees tree-like directory generation schematic diagram of the embodiment of the present invention；
Fig. 7 is the B+ trees tree-like directory generation schematic diagram of the embodiment of the present invention.
Embodiment
Many details are given in being described below in detail by way of example, to ensure to the technology of the present invention method
It is thorough to understand.But for knowing the people of the basic general knowledge in this area, it is to be understood that without these details, reality of the invention
Applying example can also realize.In addition, well-known method, process, part and circuit are not described in, to avoid the reality of the present invention
Now become unclear.
The present invention basic thought be：Small records are write into core buffer first, will buffering when core buffer is full
Area's content is written in block file；Block file is dispersed in the tree file bibliographic structure automatically generated.
As shown in Fig. 2 specific workflow of the present invention is as follows：
1st, the new record that user submits is received；
2nd, record identification is generated, the physical address each recorded, the access of convenient record are directly recorded using special identifier；
If the 3, core buffer is full, 4 are gone to；Otherwise new record is stored in core buffer, goes to 1；
If the 4, current blocked file is full, turn 5；Otherwise core buffer is appended to current blocked file, turns 1；
If the 5, current directory is full, new catalogue is created, and using new directory as work at present catalogue, in work at present
New blocked file is created in catalogue and using this document as current blocked file, and core buffer is appended to current piecemeal text
Part, turn 1.
Embodiment one：
The present embodiment is the relevant contract explanations such as data structure, parameter in the present invention.
Without loss of generality, Fig. 1 is B- tree storage organization schematic diagrames.Day is created under equipment disk catalogue/Root/...
Will storage catalogue/data, for/data files equivalent to the root node of B- trees, cFile is the block file of current data to be written,
Catalogue where cFile is work at present catalogue cPath.
The present invention uses the thought of automatic Partitioning, and the form that multiple small records are connected to blocked file is stored into disk.
Block file of the present invention refers to being saved in the file in magnetic disk storage.The rank that the present invention defines B- trees is M, that is, is saved
Maximum file number on point under single file folder is M, and maximum number of files is M.The largest amount for defining single blocked file is
fSize.It is sequential storage to be recorded in block file, what block file created in file directory and sequentially, and is each block
File distributes a globally unique ID number BID.
Can be that each record assigns a record number RID to realize the fast positioning of record.The RID by BID and
Data are recorded in the skew composition of block file.It is as shown in table 1 data record location structure information, content is as follows：
The data record location structure information table of table 1
| BID | Put each first place of the data file on disk |
| OFFSET | The deviation post of data atom hereof |
BID：The ID of blocks of files where recording, can calculate its file path PATH in disk according to the ID；
OFFSET：The skew being recorded in block file.
During storage, block file size will meet agreed terms, and the block file number under single file folder will also meet about
Fixed condition.B- trees are created by order incremented by successively, and block file is also to be established successively according to incremental order, so block file
BID be also incremented by.So the true path of block file can be quickly calculated according to the BID of block file.Can according to the RID of record
To calculate the physical address being recorded in disk, quick record accessing is realized.
Embodiment two：
The present embodiment is the implementation of data buffer storage mechanism, as shown in Figure 3.Due to the data recording speed that receives very
It hurry up, and record smaller.Writing disk manipulation is done at once if receiving a record every time, it will causes magnetic disc i/o to read and write
In frequent, tracking overlong time, and then cause storage efficiency low.In order to avoid frequently writing disk manipulation, invention introduces
The mechanism of data buffer storage.The new record received is not put in storage directly, but is first stored in buffer zone, is only expired in spatial cache
In the case of, just perform writing disk manipulation.This mode can reduce disk write indegree as far as possible, improve warehouse-in efficiency.
In data buffer storage mechanism, one piece of memory field of distribution as data buffer area, delays in advance in internal memory for present invention selection
Area size is deposited to can configure.Internal memory cache region is illustrated in figure 3, including two pointers pBuffer, pBuffered, difference
The initial address of buffer area data with existing record and last address are pointed to.
If buffer zone clear area size leaves new record enough, i.e. 302 signified regions in Fig. 3 then directly will record
301 region afterbodys are inserted into, and then are moved after tail pointer；Otherwise, soar buffer area, then be inserted into new record.
Embodiment three：
The present embodiment is data buffer storage phase flow, as shown in figure 4, comprising the following steps：
Step 401：Receive new record；
Step 402：Record identification is generated to new record；
Step 403：Judge whether memory buffer region is full, i.e., whether it exceedes institute with size plus new record size
Core buffer maximum is set, if less than going to step 405, otherwise goes to step 404；
Step 404：Current block file cFile is obtained, the data in memory cache region are write in current block file, emptied
Memory cache region, the value for setting pointer buffered is null；
Step 405：Record is kept in memory buffer region, adjustment pointer buffered makes it point to core buffer
The end address of data record in domain.
Example IV：
The tissue of document directory structure and block file is reasonable, and the quantity of the lower file of single file folder can not be too big, single
The size of individual file also can not ether it is big, otherwise recording the performance of storage can decline.The present embodiment is block file product process, such as
Shown in Fig. 5.Comprise the following steps：
Step 501：Read the data in memory buffer region；
Step 502：If without work at present block file cFile, work at present mesh is obtained by the catalogue generating algorithm set
CPath is recorded, block file is generated in cPath, using the block file as work at present block file cFile；
Step 503：Judge whether work at present block file is full, i.e., its size is plus the data length of core buffer
It is no to exceed threshold value fSize.If completely going to step 504,507 are otherwise gone to step；
Step 504：Judge whether the block file number under working node cPath reaches threshold value M, if having reached M, go to step
505, otherwise go to step 506；
Step 505：Defined according to tree directory structure and generate new catalogue, and using the catalogue as new working directory
cPath；
Step 506：New block file is created under working directory cPath, and as work at present block file
cFile；
Step 507：Afterbody by the data supplementing in core buffer to work at present block file cFile；
Embodiment five：
The present embodiment is tree-like directory product process, according to the present embodiment, realizes that the data structure of file directory generation can
To be B- trees or B+ trees and the other well known tree construction with ranking function, such as AVL.
The generating process of bibliographic structure is introduced by taking B- and B+ trees as an example below, other data structures are by that analogy.
The generating process of B- trees is top-down, all has block file under all B- tree nodes, intermediate node is deposited simultaneously
File and block file are stored up.Catalogue product process based on B- trees is as shown in Figure 6.Comprise the following steps：
Step 601：It is space-time in tree, daily record storage catalogue/data is created under assigned catalogue, and with/data files
As the root node of B- trees, root node is designated as current directory cPath, i.e. current directory, terminated；
Step 602：If present tree is expired, turn 603.Otherwise in the leaf node layer foundation successively from left to right of present tree
New leaf node.And new node is labeled as current directory cPath, terminate；
Step 603：New leaf node is established below in the leftmost leaf node of present tree, and by new node mark
Current directory cPath is designated as, is terminated.The number of plies increase now set.
So generation storage catalogue structure, is filled up until by disk space.
The generating process of B+ trees is that from bottom to top, and block file is merely stored on the leaf node of B+ trees.Based on B+ trees
Catalogue product process, as shown in Figure 7.Comprise the following steps：
Step 701：Be space-time in tree, establishment daily record storage catalogue/data under assigned catalogue, and under/data catalogues
Root node R is established, root node R is designated as working node cPath, i.e. current directory, is terminated；
Step 702：If present tree is expired, turn 703.Otherwise in the leaf node layer foundation successively from left to right of present tree
New leaf node.And new node is labeled as current directory cPath, terminate；
Step 703：A new root node R is established for tree, and is attached to elite tree as left subtree on R, is gone to step
702.The number of plies increase now set.
So generation storage catalogue structure, is filled up until by disk space.
Embodiment six：
The present embodiment is that the fast positioning of record how is realized in the present invention.
A mark is all generated in the present invention to each new record, employs the thing of special mark directly record record
Manage address.
As described in embodiment one, each record has a numbering RID, record number RID location structures as shown in Table 1
Middle block number BID and block bias internal OFFSET two parts composition, and the RID has global uniqueness.
Assuming that we name by file and block file of integer, that is, it is named as 0,1,2,3 mode.With in embodiment five
B- trees exemplified by, it is known that BID bid, then coordinates measurement algorithm be：
Algorithm GenPath：The path of block file is generated according to the bid of input.
According to algorithm above, it is assumed that M=3, bid=7, then the relative path of blocked file is 1/1.If bid=12, divide
The relative path of block file is 0/0/0.
Embodiment seven：
The present embodiment is the embodiment and its performance that the technical program is realized.
It is a realization of the inventive method on industrial computer below, industrial computer major parameter is：
CPU：I3,2.7GHZ；
Internal memory：4G；
Hard disk：Enterprise-level 1TB, 7200 turns；
Operating system：LINUX system.
Foundation/root of the data catalogues as storage system under disk root, and carried out using one process mode
In-stockroom operation.
Actual test storage performance is about 22.8 ten thousand/s, flow about 100MB/s.And with storage record quantity not
Disconnected to increase, storage will not be slack-off with the increase of recorded amounts.
A kind of magnanimity small records towards stand-alone environment that the embodiment of the present invention is provided are described in detail above efficiently to deposit
Store up management method.It assists in the core concept and its ins and outs for understanding the present invention.In scope of the appended claims
Interior, the technical staff with the field general knowledge can have found that this specification content is not to be construed as limiting the protection model of the present invention
Enclose, within the spirit and principles of the invention, any modification, equivalent substitution and improvements done etc., should be included in the present invention
Within protection domain.
Claims (6)
1. a kind of magnanimity small records efficient storage management method towards unit, it is characterised in that including herein below：
A. data receiver and caching：Small records data are received, first will note to avoid frequent I/O from asking and causing disk fragmentses
Address book stored is in core buffer；
B. automatic Partitioning：When the core buffer can not accommodate new record, the data in the core buffer are written to
In magnetic disk storage, and the core buffer is emptied to write the new record of the new reception；To prevent single file too
It is big to influence I/O performances, limit the size threshold value of each data file：If current blocked file size exceedes the threshold value, from
Newly-built blocked file is moved, and is write new data into new blocked file；
C. tree-like directory generates：To prevent that file influences I/O performances too much under single catalogue, limits the text below each file
Part sum, the tree-shaped tissue of file and blocked file is realized with simplified B-tree management algorithm；
D. record identification：The physical address each recorded, the access of convenient record are directly recorded using special identifier.
A kind of 2. magnanimity small records efficient storage management method towards unit according to claim 1, it is characterised in that：
The automatic newly-built blocked file, i.e. blocked file dynamic generative process comprise the following steps：
If step 1. obtains current directory cPath, in cPath without current block file cFile by tree-like directory generating algorithm
Middle generation blocked file, using the block file as current block file cFile, terminate；
If step 2. cFile remaining spaces are enough the data for writing the core buffer, terminate；
If block file sum is not above predetermined threshold in step 3. current directory cPath, the generation point in current directory
Block file, and using the block file as current block file cFile, terminate；
Step 4. generates new storage catalogue by the tree-like directory generating algorithm set, using new storage catalogue as current directory
CPath, block file is generated in cPath, and using the block file as current block file cFile.
3. a kind of magnanimity small records efficient storage management method towards unit according to claim 1 or 2, its feature
It is：The tree-shaped tissue that file and blocked file are realized with simplified B-tree management algorithm, Directory tree structure herein
Using B- trees, B+ trees or Adelson-Velskii-Landis tree data structure organization blocked file；The exponent number of Directory tree structure, single file folder under text
The size of quantity, the quantity of block file and blocked file that part presss from both sides is configurable.
A kind of 4. magnanimity small records efficient storage management method towards unit according to claim 3, it is characterised in that：
The catalogue generating algorithm of B- trees comprises the following steps：
Step 1：Be space-time in tree, establishment daily record storage catalogue/data under assigned catalogue, and B- is used as using/data files
The root node of tree, root node is designated as current directory cPath, terminated；Otherwise, if magnetic disk storage is full, terminate；
Step 2：In the leaf node layer where cPath, the next leaf section that can establish cPath is checked whether from left to right
Point, i.e., it is less than threshold value m node with the presence or absence of number of child nodes, m is the rank of B- trees herein：It is if it is present low at first
New node is created under threshold value m node, and new node is labeled as current directory cPath, is terminated；
Step 3：New leaf node is established below in the leftmost leaf node of present tree, and is to work as by new vertex ticks
Preceding catalogue cPath, the number of plies increase now set.
A kind of 5. magnanimity small records efficient storage management method towards unit according to claim 3, it is characterised in that：
The catalogue generating algorithm of B+ trees comprises the following steps：
Step 1：Be space-time in tree, establishment daily record storage catalogue/data under assigned catalogue, and establish B+ under/data catalogues
The root node R of tree, root node R is designated as current directory cPath, terminated；Otherwise, if magnetic disk storage is full, terminate；
Step 2：In the leaf node layer where cPath, the next leaf section that can establish cPath is checked whether from left to right
Point, i.e., it is less than threshold value m node with the presence or absence of number of child nodes, m is the rank of B+ trees herein：It is if it is present low at first
New node is created under threshold value m node, and new node is labeled as current directory cPath, is terminated；
Step 3：A new root node R is established for tree, and is attached to elite tree as left subtree on R, the number of plies now set increases
Add, go to step 2.
6. a kind of magnanimity small records efficient storage management method towards unit according to claim 1 or 2, its feature exist
In：The record identification is numbered by blocked file and block bias internal forms；Blocked file is navigated to according to blocked file numbering
Path, the original position being recorded in inside block file is navigated to according to block bias internal, so that can be direct by the record identification
Position the physical address of record, the access of convenient record.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410617251.4A CN104536908B (en) | 2014-11-05 | 2014-11-05 | A kind of magnanimity small records efficient storage management method towards unit |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410617251.4A CN104536908B (en) | 2014-11-05 | 2014-11-05 | A kind of magnanimity small records efficient storage management method towards unit |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104536908A CN104536908A (en) | 2015-04-22 |
| CN104536908B true CN104536908B (en) | 2017-12-29 |
Family
ID=52852438
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410617251.4A Expired - Fee Related CN104536908B (en) | 2014-11-05 | 2014-11-05 | A kind of magnanimity small records efficient storage management method towards unit |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104536908B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107229429B (en) * | 2017-06-27 | 2020-06-16 | 苏州浪潮智能科技有限公司 | Storage space management method and device |
| CN109491962B (en) * | 2018-11-06 | 2022-02-22 | 深圳市迅雷网络技术有限公司 | File directory tree management method and related device |
| CN109918341B (en) * | 2019-02-26 | 2021-11-30 | 厦门美图之家科技有限公司 | Log processing method and device |
| CN111858399B (en) * | 2019-04-30 | 2023-11-24 | 贵州白山云科技股份有限公司 | Data transmission method, device, computer readable storage medium and computer equipment |
| CN110727608B (en) * | 2019-09-03 | 2022-07-12 | 深圳市道通智能航空技术股份有限公司 | Log data storage method, electronic equipment and storage medium |
| CN112948343A (en) * | 2021-03-25 | 2021-06-11 | 兴业数字金融服务（上海）股份有限公司 | Mass content storage system and method based on distributed big data block |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101854388A (en) * | 2010-05-17 | 2010-10-06 | 浪潮(北京)电子信息产业有限公司 | Method and system concurrently accessing a large amount of small documents in cluster storage |
| CN102012851A (en) * | 2010-12-20 | 2011-04-13 | 浪潮(北京)电子信息产业有限公司 | Continuous data protection method and server |
| CN102332029A (en) * | 2011-10-15 | 2012-01-25 | 西安交通大学 | Hadoop-based mass classifiable small file association storage method |
| CN103020315A (en) * | 2013-01-10 | 2013-04-03 | 中国人民解放军国防科学技术大学 | Method for storing mass of small files on basis of master-slave distributed file system |
-
2014
- 2014-11-05 CN CN201410617251.4A patent/CN104536908B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101854388A (en) * | 2010-05-17 | 2010-10-06 | 浪潮(北京)电子信息产业有限公司 | Method and system concurrently accessing a large amount of small documents in cluster storage |
| CN102012851A (en) * | 2010-12-20 | 2011-04-13 | 浪潮(北京)电子信息产业有限公司 | Continuous data protection method and server |
| CN102332029A (en) * | 2011-10-15 | 2012-01-25 | 西安交通大学 | Hadoop-based mass classifiable small file association storage method |
| CN103020315A (en) * | 2013-01-10 | 2013-04-03 | 中国人民解放军国防科学技术大学 | Method for storing mass of small files on basis of master-slave distributed file system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104536908A (en) | 2015-04-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104536908B (en) | A kind of magnanimity small records efficient storage management method towards unit | |
| CN102662992B (en) | Method and device for storing and accessing massive small files | |
| CN102821138B (en) | Metadata distributed storage method applicable to cloud storage system | |
| US20160335304A1 (en) | Data partitioning and ordering | |
| CN102985909B (en) | Object for good lattice provides the method and apparatus of the high scalability network storage | |
| CN104504105B (en) | A kind of storage method of real-time data base | |
| CN106446001B (en) | A kind of method and system of the storage file in computer storage medium | |
| US20070162506A1 (en) | Method and system for performing a redistribute transparently in a multi-node system | |
| CN111427847B (en) | Indexing and querying method and system for user-defined metadata | |
| CN104657459A (en) | Massive data storage method based on file granularity | |
| CN104881466B (en) | The processing of data fragmentation and the delet method of garbage files and device | |
| CN103812939A (en) | Big data storage system | |
| CN102541985A (en) | Organization method of client directory cache in distributed file system | |
| CN102915278A (en) | Data deduplication method | |
| CN104054071A (en) | Method for accessing storage device and storage device | |
| CN103425785A (en) | Data storage system and user data storage and reading method thereof | |
| CN107506477A (en) | A kind of archive management system | |
| CN103559229A (en) | Small file management service (SFMS) system based on MapFile and use method thereof | |
| CN102915340A (en) | Expanded B+ tree-based object file system | |
| CN109542861A (en) | File management method, device and system | |
| CN110427364A (en) | A kind of data processing method, device, electronic equipment and storage medium | |
| CN102779138A (en) | Hard disk access method of real time data | |
| CN105677252B (en) | Read method, data processing method and the associated storage device of data | |
| CN103942301A (en) | Distributed file system oriented to access and application of multiple data types | |
| CN107408132A (en) | The effective performance of insertion and point inquiry operation in row store |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100085 Jihua building, No. 9, Third Street, Haidian District, Beijing, Beijing 1006 Applicant after: VISUALSEC (BEIJING) TECHNOLOGY Co.,Ltd. Address before: 100085 Jihua building, No. 9, Third Street, Haidian District, Beijing, Beijing 1006 Applicant before: BEIJING CSBIT TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP02 | Change in the address of a patent holder |
Address after: 100081, No. 3, No. 34, Zhongguancun South Street, Haidian District, Beijing, 16 floor 1908 Patentee after: VISUALSEC (BEIJING) TECHNOLOGY Co.,Ltd. Address before: 100085 Jihua building, No. 9, Third Street, Haidian District, Beijing, Beijing 1006 Patentee before: VISUALSEC (BEIJING) TECHNOLOGY Co.,Ltd. |
|
| CP02 | Change in the address of a patent holder | ||
| DD01 | Delivery of document by public notice | ||
| DD01 | Delivery of document by public notice |
Addressee: Zhang Hanqiu Document name: payment instructions |
|
| DD01 | Delivery of document by public notice | ||
| DD01 | Delivery of document by public notice |
Addressee: Zhang Hanqiu Document name: Notification of Termination of Patent Right |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171229 Termination date: 20211105 |