CN105574104B - A kind of LogStructure storage system and its method for writing data based on ObjectStore - Google Patents
A kind of LogStructure storage system and its method for writing data based on ObjectStore Download PDFInfo
- Publication number
- CN105574104B CN105574104B CN201510924037.8A CN201510924037A CN105574104B CN 105574104 B CN105574104 B CN 105574104B CN 201510924037 A CN201510924037 A CN 201510924037A CN 105574104 B CN105574104 B CN 105574104B
- Authority
- CN
- China
- Prior art keywords
- data
- logstructure
- file
- write
- storage system
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0638—Organizing or formatting or addressing of data
- G06F3/0643—Management of files
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/061—Improving I/O performance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0614—Improving the reliability of storage systems
- G06F3/0616—Improving the reliability of storage systems in relation to life time, e.g. increasing Mean Time Between Failures [MTBF]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0646—Horizontal data movement in storage systems, i.e. moving data in between storage devices or systems
- G06F3/0647—Migration mechanisms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0673—Single storage device
- G06F3/068—Hybrid storage device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0683—Plurality of storage devices
- G06F3/0685—Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays
Abstract
The present invention provides a kind of LogStructure storage system based on ObjectStore, including meta data manager and data management system, wherein is decoupling between the meta data manager and the data management system;The data management system is LogStructure engine;The LogStructure engine includes multiple memory concordance lists, multiple journal files and multiple archive log index files and obj ect file;The multiple memory concordance list, the multiple journal file and multiple archive log index files and the obj ect file are stored respectively in memory, SSD and HDD, and the multiple memory concordance list in the memory corresponds the multiple journal file in the SSD respectively.The present invention also provides its method for writing data.The present invention optimizes the readwrite performance of data by LogStructure engine, to promote the readwrite performance of Ceph.
Description
Technical field
The present invention relates to computer memory technical fields, more specifically to a kind of based on ObjectStore's
LogStructure storage system and its method for writing data.
Background technique
Ceph is a set of open source univesral distribution formula storage system, has excellent performance, reliability, scalability, each
There is extensive use in a field, and obtains the support of a large amount of open source tissue, mainly by OSD, MON and MDS three parts structure
At, wherein the data storage cell based on OSD, as shown in Figure 1, being the overall architecture of OSD, as shown, in bottom
Currently mainly there are FileStore, KeyValueStore and NewStore in storage, the different storages of the bottom realize face
To the relevant interface of objected database system ObjectStore.
However, there are the following problems by current Object-Oriented Database System ObjectStore:
(1) in FileStore, data write-in is first stored on journal, is conducive to the performance for promoting small IO, but deposit
In data double the problem of writing amplification, the double scale-up problem of writing of the data is especially serious when to be sequentially written in, so that whole
The speed of a Ceph system being sequentially written in is not high, and when data are repaired, and inner most is sequence read-write, so that
There are IO obstructions for system, influence the normal use of system.Meanwhile current SSD accelerates that all embodies to write in acceleration, simultaneously to reading
Do not improve;
(2) total data is all deposited into database by KeyValueStore, but since general database is not
It is designed for storage blob, therefore there is a problem of that write performance is not high, it is especially serious when being read and write for sequence;
(3) NewStore is also known as KeyFileStore, is deposited by using database purchase metadata and file system
Real data is stored up, solves the above problem in FileStore and KeyValueStore, but in data write-in, caching is straight
It connects and is stored using database, can equally there are problems that database is not suitable for storage long data block, so that actually using
The performance of Shi Shunxu read-write is promoted, but the promotion of the performance of random read-write is not obvious.
Summary of the invention
The technical problem to be solved in the present invention is that being applied to for Object-Oriented Database System ObjectStore
There are the above problem when Ceph, a kind of LogStructure storage system based on ObjectStore and its data write-in are provided
Method.
In order to achieve the above objects and other related objects, the present invention provides a kind of based on ObjectStore's
LogStructure storage system, the LogStructure storage system include meta data manager and data management system,
In, it is decoupling between the meta data manager and the data management system;The data management system is LogStructure
Engine;The LogStructure engine includes multiple memory concordance lists, multiple journal files and multiple archive logs index text
Part and obj ect file;The multiple memory concordance list, the multiple journal file and multiple archive log index files and
The obj ect file is stored respectively in memory, SSD and HDD, and the multiple memory concordance list difference in the memory
Correspond the multiple journal file in the SSD.
In one embodiment of the invention, the index information in the memory concordance list by skip list structure come tissue, and
The query time complexity of the concordance list is O (logN).
In one embodiment of the invention, the LogStructure storage system further includes processor, the processor
For being written to data in journal file in such a way that tail of file is additional when data are written, while establishing in memory should
The index information of relevant position in journal file.
In one embodiment of the invention, the processor is also used to exist when the data entry in a journal file
When reaching preset threshold, the corresponding memory concordance list of the journal file is freezed, and generate a new journal file
Continue to take over the operation of write-in.
In one embodiment of the invention, the processor is also used to reach default threshold in the space of SSD when journal file
When value, by the Data Migration of the journal file in SSD into HDD.
The present invention also provides a kind of method for writing data, applied to as it is above-mentioned based on ObjectStore
In LogStructure storage system, the LogStructure storage system includes meta data manager and data management system,
It wherein, is decoupling between the meta data manager and the data management system;The data management system is
LogStructure engine;The LogStructure engine includes that multiple memory concordance lists, multiple journal files are returned with multiple
Shelves log index file and obj ect file;The multiple memory concordance list, the multiple journal file and multiple archive logs
Index file and the obj ect file are stored respectively in memory, SSD and HDD, and the multiple interior in the memory
It deposits concordance list and corresponds the multiple journal file in the SSD respectively;It the described method comprises the following steps: S1, creation
The obj ect file simultaneously writes data into the obj ect file of creation;S2, tail portion can be carried out to the data in judgement
When addition, the data are written in the obj ect file of bottom;It S3, in the length for judging the data is more than described
When the threshold limit of the write-in buffer area of SSD, the data are written in the obj ect file of the HDD, and will be described
Partial data present in write-in buffer area is set as deletion state;S4, the data page for being written to write-in buffer area
To be divided into the small object of multiple page alignments, and the multiple small object is written in the record of the memory concordance list.
In one embodiment of the invention, the index information in the memory concordance list by skip list structure come tissue, and
The query time complexity of the concordance list is O (logN).
In one embodiment of the invention, the LogStructure storage system further includes processor, the processor
For being written to data in journal file in such a way that tail of file is additional when data are written, while establishing in memory should
The index information of relevant position in journal file.
In one embodiment of the invention, the processor is also used to reaching when the data entry in a journal file
The corresponding memory concordance list of the journal file is freezed when preset threshold, and generates a new journal file after continued access
For the operation of write-in.
In one embodiment of the invention, the processor is also used to reach default threshold in the space of SSD when journal file
When value, by the Data Migration of the journal file in SSD into HDD.
As described above, the LogStructure storage system and its data write-in side of the invention based on ObjectStore
Method has the advantages that LogStructure engine is stored in memory, SSD and HDD, has fully considered different storages
The characteristic of medium, so that the LogStructure storage system carries out the storage performance that data storage has played data.In SSD
Data storage is carried out using LogStructure structure, converts sequential write for random write well, improves writing for system
Enter performance, to promote the whole readwrite performance of Ceph.Meanwhile being stored in HDD using obj ect file, weigh data
Read-write requires, so that the LogStructure storage system is applicable in different application scene.In addition, LogStructure is deposited
What storage system solved original ObjectStore writes scale-up problem, improves the effective rate of utilization of equipment.The LogStructure
Storage system realizes bedding storage, dsc data is cached in SSD, and realize readable write buffer.For writing for alphabetic data
Enter, directly write in disk, without SSD, avoids a large amount of sequence IO made when data reparation or migration to cause system
Performance shake occur " Slow IO " phenomenon.
Detailed description of the invention
Fig. 1 is shown as the basic framework figure of the OSD of current Ceph.
Fig. 2 is shown as the structure chart of the LogStructure storage system based on ObjectStore of the embodiment of the present invention.
Fig. 3 is shown as the structure chart of the LogStructure engine in Fig. 2.
Fig. 4 is shown as the structure chart of the memory concordance list in Fig. 3.
Fig. 5 is shown as the structure chart of the journal file in Fig. 3.
Fig. 6 is shown as the schematic diagram of the additional write-in of log
Fig. 7 is shown as the basic I/O data flow of the LogStructure storage system of the invention based on ObjectStore
Figure.
The data that Fig. 8 is shown as the LogStructure storage system based on ObjectStore of the embodiment of the present invention are write
Enter the flow chart of method.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that illustrating the basic structure that only the invention is illustrated in a schematic way provided in following embodiment
Think, only shown in schema then with related component in the present invention rather than component count, shape and size when according to actual implementation
Draw, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its assembly layout kenel
It is likely more complexity.
2 are please referred to Fig. 8.It is that the LogStructure based on ObjectStore of the embodiment of the present invention is deposited shown in Fig. 2
The structure chart of storage system.With reference to Fig. 2, which includes meta data manager and data management system,
In, it is decoupling between the meta data manager and the data management system, which is a K/V database (K/V
Database), which is LogStructure engine (LogStructure Engine).
The K/V database includes Omap and ObjectMeta, and Omap is used to match the phase of corresponding Ceph ObjectSore
Interface, and the correlation attribute information of storage object are closed, ObjectMeta is used for data management, and is mapped to LogStructure and draws
In the Object Management group held up.In the LogStructure engine include WAL, ReadCache and ObjectFileStore, WAL and
Accelerator of the ReadCache as read-write, ObjectFileStore store specific data pair using independent blocked file
As.
As shown in figure 3, LogStructure engine includes multiple memory concordance lists (MemIndexTable), multiple logs
File (Log File) and multiple archive logs index (ArchiveLogIndex) file and multiple obj ect file (Object
File), wherein multiple memory concordance lists are stored in memory, multiple journal files and the storage of multiple archive log index files
In SSD (Solid State Drive, solid state hard disk) and multiple obj ect files are stored in HDD (Mechanical hard
Disk, mechanical hard disk) in, and the memory concordance list of memory corresponds to the journal file of SSD, realizes LogStructure engine
It is stored in the hybird environment of SSD and HDD.Further, the size of obj ect file is 1M~16M.
Wherein, by skip list data structure come the index information in tissue memory concordance list, when the inquiry of memory concordance list
Between complexity be O (logN), as shown in figure 4, the basic structure of index information are as follows: key, log ID (logno), log offset
(offset) and seq, wherein log ID and log are deviated for orienting specific data record;Seq indicates data storage
Mode does not need modification data content in this way, only with additional mode, therefore the LogStructure storage system be can be with
Realize that system concurrence is " writing mutual exclusion to read shared and write not mutual exclusion reading ".
Journal file includes a plurality of log recording, and every log recording is the total data that a complete transaction is submitted, by
The submission of Yu Yici complete transaction may include multiple data modification operations, as shown in figure 5, every data record includes multiple K/V
Data cell.As can be seen that the operation of complete transaction is recorded as due to every, it is when the LogStructure is stored
When system is restarted, storage index structure can be re-established by playback journal file.
Further, all data are updated by additional form, i.e., do not cover modification.For example, such as
Shown in Fig. 6, the data of first time key1 write-in are value1, and later, key1 is modified to value3 again, at this point, the number of key1
According to write-in can't be capped, but the corresponding data modified of write-in in the new position, it can read seq when system is read
Effective Numerical of the relatively large value of numerical value as corresponding key, the i.e. Effective Numerical of key1 are to read when seq numerical value is 3
value3。
In an embodiment of the present invention, which further includes processor (not shown).?
In LogStructure engine, only one memory concordance list and a journal file are revisable, as shown in figure 3, being most left
One memory concordance list and a journal file on side.What processor was used to add data by tail of file when data write-in
Mode is written in journal file, while establishing the index information of the relevant position in the journal file in memory.And then WAL
(Write Ahead Log, ahead log) is conducive to the accurately recovery processing in system crash.
Further, processor is also used to when the data entry in a journal file is when reaching preset threshold, by this
The corresponding memory concordance list of journal file is freezed, and is generated a new journal file and continued to take over the operation of write-in.
At this point, processor, which triggers corresponding thread, carries out filing processing to the memory concordance list freezed, when filing processing, according to memory
In memory concordance list generate an archive log index file, and the archive log index file is stored in SSD.In this way
It just can be avoided system and need to directly read the case where journal file is to restore the index information in memory in restarting, from
And directly read the fast quick-recovery that relevant archive log index file realizes system.
Since the query time complexity of memory concordance list is O (logN), need to carry out data at horizontal meaders processing, place
Reason device is also used to merge the memory concordance list freezed, and multiple memory concordance lists are merged into a memory concordance list, this
When the corresponding archive log index file of one memory concordance list, and corresponding multiple journal files.
Further, longitudinal merger processing is carried out to data, processor is used to reach pre- in the space of SSD when journal file
If when threshold value, by the Data Migration of the journal file in SSD into HDD.At this point, when carrying out the refresh operation of data, traversal
Oldest memory concordance list obtains the data block of an approximate sequence, and the operation of multiple small data blocks is merged into a big IO
Operation, writes with a brush dipped in Chinese ink in specific obj ect file.
In an embodiment of the present invention, LogStructure engine is stored in memory, SSD and HDD, has fully considered not
With the characteristic of storage medium, so that the LogStructure storage system carries out the storage performance that data storage has played data.
Data storage is carried out using LogStructure structure in SSD, converts sequential write for random write well, improves and is
The write performance of system, to promote the whole readwrite performance of Ceph.In addition, being stored in HDD using obj ect file, weigh
The read-write requirement of data, so that the LogStructure storage system is applicable in different application scene.
As shown in fig. 7, being the basic I/O data of the LogStructure storage system of the invention based on ObjectStore
Flow graph is written IO points as sequence (Sequentail) IO and random (Random) IO with reference to Fig. 7, wherein sequence IO without
SSD directly writes in disk, avoids a large amount of sequence IO made when data reparation or migration to cause the performance of system to be shaken
There is " Slow IO " phenomenon.It is illustrated below with reference to method for writing data of the Fig. 8 to the LogStructure storage system:
As shown in figure 8, being the flow chart of the method for writing data of the embodiment of the present invention, this method is applied to above-mentioned
LogStructure storage system.With reference to Fig. 8, method includes the following steps:
The write-in data are simultaneously written in the obj ect file of creation by S1, creation obj ect file;
In this step, in judgement there is no creation obj ect file when the obj ect file of write-in data and by the write-in data
It is written in the obj ect file of creation
S2, judgement can to the data carry out tail portion addition when, write the data in the obj ect file of bottom;
S3, when the length for judging the data is more than the threshold limit of write-in buffer area (WriteBuffer) of SSD, will
The data are written in the obj ect file of HDD, and are set as partial data present in write-in buffer area to delete (delete)
State;
S4, the data page of write-in buffer area is written to be divided into the small object of multiple page alignments, and will be multiple small
Object is written in the record of memory concordance list.
Accordingly, for the reading of data, add due to having in the memory and SSD in the LogStructure storage system
Fast mechanism, therefore, the reading of data need different levels to be accessed one by one, i.e., first read the write-in buffer area in SSD
Data, then the data in two-stage reading cache (Cache) are read, last read the data in the obj ect file in HDD.
In conclusion the present invention is based on the LogStructure storage system and its method for writing data of ObjectStore,
LogStructure engine is stored by the hybird environment that hardware device is SSD and HDD, the high IOPS for giving full play to SSD is special
Property, the IOPS of LogStructure storage system is improved, solves the scale-up problem of writing of original ObjectStore, is improved
The effective rate of utilization of equipment.Meanwhile the LogStructure storage system realizes bedding storage, and dsc data is cached in SSD,
And realize readable write buffer.In addition, the write-in for alphabetic data, directly writes in disk, without SSD, avoid counting
There is " Slow IO " phenomenon to cause the performance of system to be shaken in a large amount of sequence IO made when according to reparation or migration.So this hair
It is bright effectively to overcome various shortcoming in the prior art and have high industrial utilization value.
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.It is any ripe
The personage for knowing this technology all without departing from the spirit and scope of the present invention, carries out modifications and changes to above-described embodiment.Cause
This, institute is complete without departing from the spirit and technical ideas disclosed in the present invention by those of ordinary skill in the art such as
At all equivalent modifications or change, should be covered by the claims of the present invention.
Claims (5)
1. a kind of method for writing data of the LogStructure storage system based on ObjectStore, described
LogStructure storage system includes meta data manager and data management system, wherein the meta data manager with it is described
It is decoupling between data management system;The data management system is LogStructure engine;The LogStructure engine
Including multiple memory concordance lists, multiple journal files and multiple archive log index files and obj ect file;In the multiple
Deposit concordance list, in the multiple journal file and multiple archive log index files and the obj ect file are stored respectively in
It deposits, in SSD and HDD, and the multiple memory concordance list in the memory is corresponded respectively in the described more of the SSD
A journal file;It is characterized by: the described method comprises the following steps:
S1, the creation obj ect file simultaneously write data into the obj ect file of creation;
S2, judgement can to the data carry out tail portion addition when, the data are written in the obj ect file of bottom;
S3, the length for judging the data be more than the SSD write-in buffer area threshold limit when, the data are written
Deletion state is set as into the obj ect file of the HDD, and by partial data present in said write buffer area;
S4, the data page of write-in buffer area is written to be divided into the small object of multiple page alignments, and will be the multiple
Small object is written in the record of the memory concordance list.
2. the data write-in side of the LogStructure storage system according to claim 1 based on ObjectStore
Method, which is characterized in that index information in the memory concordance list by skip list structure come tissue, and the inquiry of the concordance list
Time complexity is O (logN).
3. the data write-in side of the LogStructure storage system according to claim 2 based on ObjectStore
Method, which is characterized in that the LogStructure storage system further includes processor, and the processor is used for when data write-in
Data are written in journal file in such a way that tail of file is additional, while establishing the correlation in the journal file in memory
The index information of position.
4. the data write-in side of the LogStructure storage system according to claim 3 based on ObjectStore
Method, which is characterized in that the processor is also used to when the data entry in a journal file is when reaching preset threshold, by this
The corresponding memory concordance list of journal file is freezed, and is generated a new journal file and continued to take over the operation of write-in.
5. the data write-in side of the LogStructure storage system according to claim 4 based on ObjectStore
Method, which is characterized in that the processor is also used to when journal file is when the space of SSD reaches preset threshold, by the day in SSD
The Data Migration of will file is into HDD.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510924037.8A CN105574104B (en) | 2015-12-11 | 2015-12-11 | A kind of LogStructure storage system and its method for writing data based on ObjectStore |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510924037.8A CN105574104B (en) | 2015-12-11 | 2015-12-11 | A kind of LogStructure storage system and its method for writing data based on ObjectStore |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105574104A CN105574104A (en) | 2016-05-11 |
CN105574104B true CN105574104B (en) | 2019-04-05 |
Family
ID=55884235
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510924037.8A Expired - Fee Related CN105574104B (en) | 2015-12-11 | 2015-12-11 | A kind of LogStructure storage system and its method for writing data based on ObjectStore |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105574104B (en) |
Families Citing this family (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106371770B (en) * | 2016-09-05 | 2019-07-12 | 杭州华为数字技术有限公司 | Method for writing data and device |
CN106506668B (en) * | 2016-11-23 | 2019-07-16 | 浪潮云信息技术有限公司 | A method of object storage is realized based on distributed storage |
US11113159B2 (en) | 2017-06-30 | 2021-09-07 | Intel Corporation | Log structure with compressed keys |
CN107515827B (en) * | 2017-08-21 | 2021-07-27 | 湖南国科微电子股份有限公司 | PCIE SSD custom log storage method and device and SSD |
CN107943858A (en) * | 2017-11-07 | 2018-04-20 | 广州探迹科技有限公司 | A kind of crawler system IO optimization method and devices |
CN107979640B (en) * | 2017-11-21 | 2020-08-14 | 广东电网有限责任公司电力科学研究院 | Data transmission method and device |
CN109828862B (en) * | 2017-11-23 | 2023-08-22 | 成都华为技术有限公司 | Method and device for replaying log |
CN110096220B (en) | 2018-01-31 | 2020-06-26 | 华为技术有限公司 | Distributed storage system, data processing method and storage node |
CN109086133B (en) * | 2018-07-06 | 2019-08-30 | 第四范式(北京)技术有限公司 | The method and system of data is safeguarded in memory |
CN110703981A (en) * | 2018-07-10 | 2020-01-17 | 中兴通讯股份有限公司 | Data reading and writing method, terminal and computer readable storage medium |
CN109376100A (en) * | 2018-11-05 | 2019-02-22 | 浪潮电子信息产业股份有限公司 | A kind of caching wiring method, device, equipment and readable storage medium storing program for executing |
CN109857339A (en) * | 2019-01-10 | 2019-06-07 | 平安科技(深圳)有限公司 | File memory method, device, computer equipment and storage medium |
CN111949601B (en) * | 2019-05-16 | 2022-12-13 | 中移(苏州)软件技术有限公司 | Data storage method and device and computer storage medium |
CN112181973B (en) * | 2019-07-01 | 2023-05-30 | 北京涛思数据科技有限公司 | Time sequence data storage method |
CN110609813B (en) * | 2019-08-14 | 2023-01-31 | 北京华电天仁电力控制技术有限公司 | Data storage system and method |
WO2019228575A2 (en) * | 2019-09-12 | 2019-12-05 | Alibaba Group Holding Limited | Log-structured storage systems |
CN112434002A (en) * | 2020-12-25 | 2021-03-02 | 冯凌云 | HBase and Phoenix based low-cost massive structured data rapid retrieval method |
CN113031857B (en) * | 2021-01-30 | 2022-05-10 | 柏科数据技术(深圳)股份有限公司 | Data writing method, device, server and storage medium |
CN113568883A (en) * | 2021-07-29 | 2021-10-29 | 上海哔哩哔哩科技有限公司 | Data writing method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102129435A (en) * | 2010-01-13 | 2011-07-20 | 中国移动通信集团公司 | Data storage service control method and system |
CN102521269A (en) * | 2011-11-22 | 2012-06-27 | 清华大学 | Index-based computer continuous data protection method |
CN103473258A (en) * | 2013-06-01 | 2013-12-25 | 西安邮电大学 | Cloud storage file system |
-
2015
- 2015-12-11 CN CN201510924037.8A patent/CN105574104B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102129435A (en) * | 2010-01-13 | 2011-07-20 | 中国移动通信集团公司 | Data storage service control method and system |
CN102521269A (en) * | 2011-11-22 | 2012-06-27 | 清华大学 | Index-based computer continuous data protection method |
CN103473258A (en) * | 2013-06-01 | 2013-12-25 | 西安邮电大学 | Cloud storage file system |
Non-Patent Citations (2)
Title |
---|
The design and implementation of a log-structured file system;飞鸟Mu;《https://blog.csdn.net/flymu0808/article/details/40892675》;20141107;第1-3页 |
日志结构文件系统研究综述;贺勤;《计算机工程应用技术》;20090715;第128-130页 |
Also Published As
Publication number | Publication date |
---|---|
CN105574104A (en) | 2016-05-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105574104B (en) | A kind of LogStructure storage system and its method for writing data based on ObjectStore | |
CN108459826B (en) | Method and device for processing IO (input/output) request | |
CN110825748B (en) | High-performance and easily-expandable key value storage method by utilizing differentiated indexing mechanism | |
CN108268219B (en) | Method and device for processing IO (input/output) request | |
JP6713934B2 (en) | Storage device and its operating method and system | |
CN107491523A (en) | The method and device of data storage object | |
KR20150035560A (en) | Optimized context drop for a solid state drive(ssd) | |
CN103399823B (en) | The storage means of business datum, equipment and system | |
CN105955664B (en) | A kind of reading/writing method of watt record conversion layer based on segment structure | |
CN104267912A (en) | NAS (Network Attached Storage) accelerating method and system | |
CN102999428A (en) | Four-stage addressing method for tile recording disk | |
CN109739696B (en) | Double-control storage array solid state disk caching acceleration method | |
US20240020240A1 (en) | Method for storing l2p table, system, device, and medium | |
US9189408B1 (en) | System and method of offline annotation of future accesses for improving performance of backup storage system | |
CN105260139B (en) | A kind of disk management method and system | |
CN106527987A (en) | Non-DRAM SSD master control reliability improving system and method | |
CN102929935B (en) | A kind of Large Volume Data reading/writing method based on affairs | |
CN110989924B (en) | Metadata storage performance optimization method and storage server | |
CN103942301A (en) | Distributed file system oriented to access and application of multiple data types | |
CN106897311B (en) | Database batch updating method, data reduction log generating method and storage device | |
CN110134551B (en) | Continuous data protection method and device | |
CN111443874B (en) | Solid-state disk memory cache management method and device based on content awareness and solid-state disk | |
US11010091B2 (en) | Multi-tier storage | |
Lv et al. | Zonedstore: A concurrent zns-aware cache system for cloud data storage | |
CN102508790A (en) | Content-based cache method applied to content analysis storage |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190405 Termination date: 20191211 |
|
CF01 | Termination of patent right due to non-payment of annual fee |