CN104639661A - Distributed storage system and storing and reading method for files - Google Patents
Distributed storage system and storing and reading method for files Download PDFInfo
- Publication number
- CN104639661A CN104639661A CN201510110507.7A CN201510110507A CN104639661A CN 104639661 A CN104639661 A CN 104639661A CN 201510110507 A CN201510110507 A CN 201510110507A CN 104639661 A CN104639661 A CN 104639661A
- Authority
- CN
- China
- Prior art keywords
- file
- stored
- memory
- files
- blocks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
Abstract
The invention discloses a distributed storage system which comprises a storage pool, a file blocking module and a plurality of storage nodes, wherein the storage pool establishes network connection with an external client and is used for receiving a stored file uploaded by the external client by a network; the file blocking module establishes communication connection with the storage pool and is used for performing stripping processing on the stored file to obtain a plurality of file blocks; the storage nodes respectively establish communication connection with the file blocking module and are used for respectively storing the file blocks generated by blocking the stored file. According to the distributed storage system disclosed by the invention, the stored file is divided into the file blocks which are respectively stored in a plurality of different storage nodes; the storage nodes commonly perform read-write operation, so that the pressure of one storage node is dispersed, and the read-write performance is improved.
Description
Technical field
The present invention relates to a kind of cloud memory technology, be specifically related to a kind of distributed memory system and file and store and read method.
Background technology
Cloud computing be one emerging in length and breadth to the set of technology.Its Rapid Variable Design, and much provides commercial city to think obtain one seat in field here and become leader.Cloud computing has two main place---privately owned cloud and public cloud.Generally speaking, on the network in a mechanism, storage and computer resource, cloud computing can be used as required.Privately owned cloud can be provided in the traditional forms of enterprises that is that more flexibly, less limit between each department and that more have capital leverage by IT mechanism and serve.Public cloud is also about the same, but it be provided by external company and conduct interviews from the Internet.In this commonality schemata, client does not need to buy any hardware, does not need the maintenance providing hardware/software performance yet.Ripe public cloud requires that client is for the various resource of concrete use (network, storage and calculating) and checking.
The data access of traditional approach depends on the ability of single disk, comprises performance and the capacity of disk.Disk is the equipment of physical structure, and the selection of the movement and disc that depend on magnetic head carrys out access data.When there is reading and the write of big data quantity, magnetic head and disc all can run into bottleneck, cannot respond too much request, and cause service application wait for without response or interrupt, these problems are all caused by the machine construction principle of disk.Particularly when data disk breaks down, the data existed on disk will face the risk of loss.
In addition, when implementing public or privately owned cloud, the factor of most critical makes cost and management minimize exactly.The whole theory of cloud computing is all based upon saves on cost, and this saving derives from shared resource (comprising virtual and low configuration), automation and standardization.No matter client is a public cloud provider or the client that will build a privately owned cloud, and they may be concerned about the cost of Gains resources, the utilance of resource and validity, data protection/disaster recovery capability very much, and the management cost continued to increase.In addition, client may need to implement virtualization solution and provide access to share.Citrix and Vmware, normally Citrix or Vmware and more and more welcome Microsoft HyperVShared technology all comprise many tenants of some forms usually; Can support different mechanism/companies on same memory device, this support simultaneously has again certain isolation.About performance, a lot of public cloud provider may be concerned about concurrency performance and extensibility; Privately owned cloud client more may be concerned about the demand of traditional forms of enterprises's performance, namely slightly low concurrency and more transactional I/O.
After having carefully analyzed TCO, in conjunction with present cloud computing technology, client needs a kind of single and storage solution of economy.This scheme needs the ability of the systematic ease of manageability of tool (pass in time and do not need extra resource), seamless expansion capacity (and/or performance) and utilizes the ability continuing commercial storage.Except the advantage of technical elements, " buying as required, progressive expansion " pattern is also needed to protect capital and operation cost.This ability allows cloud provider can balance their cost and actual storage utilance, and this is very crucial for a successful cloud business model.
Summary of the invention
The invention provides a kind of distributed memory system and file to store and read method, realize high aggregate bandwidth, high-speed background network, improve performance and fail safe that file stores.
For achieving the above object, the invention provides a kind of distributed memory system, be characterized in, this system comprises:
Storage pool, its network connects external client, is stored file by the external client upload of network reception;
File block module, it communicates to connect storage pool, will be stored File striping and be processed into several blocks of files;
Some memory nodes, it communicates to connect described file block module respectively, and the some blocks of files being stored file block generation are stored in each memory node respectively.
Each above-mentioned memory node is a physical storage device; ; Or some memory nodes are several memory modules in a physical storage device.
Above-mentioned memory node adopts InfiniBand framework or gigabit level Ethernet to carry out communication with synchronous each other.
Above-mentioned storage pool is formed for some physical storage devices provide part memory space to be polymerized respectively.
A file memory method for distributed memory system, is characterized in, this file memory method comprises:
External client will be stored files passe to distributed memory system by network;
Distributed memory system carries out striping process to being stored file, is divided into some blocks of files;
The some blocks of files being stored file are stored in some memory nodes of distributed memory system respectively.
Above-mentioned distributed memory system stores successively to some files that is stored;
Distributed memory system store each be stored file time, this is stored to file and carries out striping process, and completes after the blocks of files this being stored file is arranged in some memory nodes, then store the next one and be stored file.
Above-mentioned distributed memory system carries out image copying to each blocks of files be stored in its each memory node, original block copies and produces one or several image file blocks, one or several other the memory node outside the memory node that a or some parts of image file blocks are stored in original block place.
Above-mentioned distributed memory system completes the striping process being stored file and after each blocks of files is stored in each memory node, is stored file edit check code to this, is stored in by check code in each memory node respectively; When wherein arbitrary memory node replaces with new memory node, other memory nodes do not replaced according to check code by the date restoring on the memory node that is replaced to new memory node.
A file reading for distributed memory system, is characterized in, this file reading comprises:
External client proposes to distributed memory system the request that reading one is stored file;
Distributed memory system reads this respectively simultaneously and is stored some blocks of files that file is divided into from some memory nodes;
Some blocks of files combination reduction are stored file by distributed memory system, and export external client to by network.
During the memory node fault at the required blocks of files place of reading of above-mentioned distributed memory system, distributed memory system reads this blocks of files and backs up the image file block in other memory nodes.
Distributed memory system of the present invention and file store to be compared with the file storage technology of prior art with read method, its advantage is, the present invention is divided into some blocks of files by being stored file and being stored in some different memory nodes respectively, multiple memory node carries out read-write operation jointly, disperse the pressure of an original memory node, improve readwrite performance;
Distributed memory system of the present invention is cluster file system, and each node in cluster storage system enjoys par, and therefore any node can process request;
The present invention adopts the storage system of distributed structure/architecture to be automatically distributed on whole cluster by all data and error correction information, the all data of the error correction technique guarantee of its high efficient and reliable keep complete, even if also can be accessed easily when there is multiple component faults at the same time
The present invention carries out image copying to the blocks of files stored, be stored in other different memory nodes respectively, a or many parts of same blocks of files are retained at other memory nodes, this mechanism is when some memory nodes break down, blocks of files can be read by the memory node from other and recover, and improves the fail safe of system;
The present invention is by with the mode layout data on physical storage device successively of an a file then file, and cluster file system can carry out the level of redundancy of control store system in the level of volume, catalogue or even file.
Accompanying drawing explanation
Fig. 1 is the structural representation of distributed memory system of the present invention;
Fig. 2 is the method flow of the file memory method of distributed memory system of the present invention;
Fig. 3 is the method flow of the file reading of distributed memory system of the present invention.
Embodiment
Below in conjunction with accompanying drawing, further illustrate specific embodiments of the invention.
As shown in Figure 1, the invention discloses a kind of distributed memory system 106, this distributed memory system 106 communicates to connect some external clients 101 by network 102.External client 101 will be stored files passe by network 102 and store to distributed memory system 106.
Distributed memory system 106 comprises: storage pool 103, file block module 104 and some memory nodes 105.
Storage pool 103, for NameSpace, there is provided part memory space to be polymerized respectively by some physical storage devices to be formed, unified access interface and administration interface can be provided to application server, application server by disks all on the back-end physical memory device of this access interface very easy managed storage pond 103, can give full play to performance and the disk utilization of memory device.Storage pool 103 network connects external client 101, is stored file for what received by network 102 that external client 11 uploads.
File block module 104 communicates to connect storage pool 103, carries out striping process, be divided into several blocks of files for the file that is stored received by storage pool 103.Storage pool 103 and file block module 104 form cluster file system, and cluster file system realizes striping process and combination reduction treatment to being stored file.
Wherein, each in some memory nodes 105 can adopt a physical storage device, or some memory nodes 105 can be adopted as several memory modules in a physical storage device, the memory module of physical storage device is generally the disk stored here.Each memory node 105 communicates to connect file block module 104 respectively, and the some blocks of files being stored file block generation are stored in each memory node 105 by file block module 104 respectively.
Further, each memory node 105 in distributed memory system 106 enjoys par, and therefore any memory node 105 can process request.Some memory nodes 105 each other, adopt InfiniBand framework or gigabit level Ethernet to carry out communication and synchronously, each memory node 105 can be made to know the layout of whole distributed memory system 106, and each file and file part deposit position.
As shown in Figure 2, the invention also discloses a kind of file memory method of distributed memory system, the method comprises:
Step 1.1, external client 101 will be stored files passe to distributed memory system 106 by network 102, and storage pool 103 is stored file by network reception.
The file block module 104 of step 1.2, distributed memory system 106 carries out striping process to being stored file, will be stored file and be divided into some blocks of files.
Step 1.3, the some blocks of files being stored file are stored in some memory nodes 105 of distributed memory system 106 respectively.
The storage system of prior art transmits data by RAID and volume management layer, creates the poor efficiency of data layout.And distributed memory system 106 of the present invention by optimizing, by the distribution of file on cluster, can significantly improve the performance of disk subsystem.Concrete example is as follows: suppose that a memory node is a disk in a physical storage device, a physical storage device is had to contain 4 pieces of disks as 4 memory nodes, when external client to deposit the file of a 4GB size to distributed storage by network, this file can by cluster file system process, on average be cut into the file of 4 1GB sizes, then the file of each 1GB deposited on each disk one to one.Namely 4 pieces of disks respectively house the file of 1GB.Here optimization refer to the present invention by file is split realize walking abreast depositing and reading, suppose that the file of a disk access 1GB size needed for 10 seconds, when not carrying out file division, the file storing 4GB size needs 40 seconds; Become the file of 4 1GB to be accessed in respectively on 4 disks the file division of 4GB and then only need 10 seconds, the access efficiency of file improves 75%.
If need to store multiple file that is stored, then circulation step 1.1 to step 1.3 carries out storing step.Here, distributed memory system 106 stores successively to some files that is stored, distributed memory system 106 store each be stored file time, this is stored to file and carries out striping process, and complete after the blocks of files this being stored file is arranged in some memory nodes 105, then store the next one and be stored file.By with mode layout data on disk of an a file then file, distributed memory system 106, realizes the level of redundancy carrying out control store system in volume, the catalogue even level of file.
As Fig. 2 and shown in composition graphs 1, an embodiment of distributed memory system 106 storage file is described, before a file is stored in the physical storage medium on backstage, first be stored file by the file block module 104 of distributed memory system 106 to this to cut into slices, namely usually said " striping process ".There are four memory nodes 106 on such as backstage, a storage 10MB be stored file time, distributed memory system 106 can be cut into the blocks of files of four 2.5MB, then these 4 blocks of files is stored into respectively on each memory node 105.
And on cluster file system, this file is still shown as a complete file, client does not also know how file is preserved on backstage, has accomplished the transparence to client.
It is advantageous that the performance and fail safe that improve file storage simultaneously.Under traditional approach, the storage of a file depends on a memory node, under the mode of cluster, the access of each memory node 105 responsible partial documents, multiple memory node 105 carries out read-write operation jointly, has disperseed the pressure of an original memory node, has improve readwrite performance.
In addition in order to improve fail safe, blocks of files after section is except preserving at local memory node, distributed memory system carries out image copying to each blocks of files be stored in its each memory node, original block copies and produces one or several image file blocks, and one or several other the memory node outside the memory node at original block place retains a or many parts of same image file blocks.This mechanism, when some memory nodes break down, blocks of files can be read by the memory node from other and recover.According to adopt memory node quantity number, many parts of data backups can be accessed, greatly improve the fail safe of system.
Distributed memory system 106 can rebuild the disk broken down at short notice, and the reconstruction of distributed memory system of the present invention is in units of memory node 105.When after the divided storage of data, cluster file system can calculate a check code automatically, and check code is left in uniformly on all disks, when after certain disk failures, after cluster file system detects that new disk is changed, can according to the data in the check code recovery failed disk that other disks are deposited before on this new disk.Utilize the vacant memory space in whole cluster to evade the risk of loss of data, monitor on one's own initiative and try to be the first and data are gone out from risky assembly transfer.Equivalent reconstruction operation is confined to enter in the subset of storage system by traditional enterprise storage system, and this causes the bottleneck having occurred recovering, and adds again the time of protected data, considerably increases the risk of the loss of data when single component breaks down.If the value protections such as traditional RAID-5 if there is multiple component faults, will cause loss of data before reconstruction completes.By contrast, the present invention adopts the storage system of distributed structure/architecture to be automatically distributed on whole cluster by all data and error correction information, the all data of the error correction technique guarantee of its high efficient and reliable keep complete, even if also can be accessed easily when there is multiple component faults at the same time.
As shown in Figure 3, be a kind of file reading of distributed memory system, this file reading comprises:
Step 2.1, external client 101 propose to distributed memory system 106 request that reading one is stored file.
Step 2.2, distributed memory system 106 read this respectively simultaneously and are stored some blocks of files that file is divided into from some memory nodes 105;
Some blocks of files combinations are reduced into and are stored file by the cluster file system in step 2.3, distributed memory system 106, and export external client 101 to by network 102.
As shown in Figure 1, such as when client need to read some be stored file time, distributed memory system 106 can read four blocks of files from four memory nodes 105 simultaneously, is then combined into original file and submits to external client 101 again.
Further, the storage mode copied in conjunction with above-mentioned image is known, when memory node 105 fault at the required blocks of files place of reading of distributed memory system 106, this blocks of files that namely reads distributed memory system 106 backs up the image file block in other memory nodes, to ensure the fail safe being stored file.
Distributed memory system 106 of the present invention, compared with traditional storage architecture, has following advantage:
At aspect of performance, conventional store framework bandwidth is fixed and performance can not promote; Distributed memory system of the present invention have high aggregate bandwidth, high-speed background network, performance with nodes increase and improve characteristic;
In dilatation, conventional store frame finite capacity, cannot do horizontal dilatation, is up to the capacity of tens TB, and performance and capacity cannot linear growths; Distributed memory system how many memory nodes of the present invention are all Single document systems, can realize dilatation and carry out online, and do not revise client setting, dilatation in every 60 seconds node is to active volume.Namely dilatation increases memory node to distributed memory system, and the memory node of increase is by after the connection of Infiniband cable, and cluster file system can sense that new equipment adds, and is similar to the hot plug of USB device automatically.The memory node increase newly added can be entered in original file system by cluster file system, forms larger storage pool.And these operations are transparent concerning client.
In storage internal exchange of data, the mode that conventional store framework adopts backboard to exchange; And the present invention adopts the infiniband switch of high bandwidth low delay, substantially increase transfer of data and synchronous efficiency.
In fail safe, conventional store framework adopts Raid6 protection, allows two hard disk corruptions at most and need HotSpare disk to start process of reconstruction; And the present invention is by storing the storage means in conjunction with image copying after striping process, realizes allowing multiple node to damage not obliterated data simultaneously, allowing file-level to do different Preservation tactics simultaneously.The Preservation tactics of different stage refers to: the maximum disk failure tolerating how much quantity of system, the number of disks of fault tolerant is more, then the utilance of the whole system memory capacity about end, because most of power system capacity to be used as backup copies.So different Preservation tactics can be arranged according to the levels necessitate of user to data protection.
In self-regeneration, conventional store framework is without self-regeneration function, and just simple two redundancy, with regard to unparalleled redundancy protecting after system jam; And present system can self-regeneration, this storage system can arrange a check code counter for each memory node or disk, system can verify the data on disk on backstage, will flip-flop number when disk exists bad track or other problems, when being accumulated to some, even if disk still can use, also unsafe disk can be considered as by system, this be system automatically Data Migration to good memory node or disk, do not need manual intervention, system is two redundancy still.
In disk reconstruction, conventional store framework needs 10 hours, rebuilds by the naked capacity of hard disk; And the present invention is because there is the existence of check code, can calculate the actual amount of data in failed disk, only need to recalculate actual data volume in process of reconstruction, single magnetic rebuilds the capacity of actual use in 1 hour even shorter, reconstruction hard disk.
At management aspect, conventional store framework needs more professional stored knowledge; And the present invention adopts distributed structure/architecture, framework is simple, managing mass and low capacity, such as 10TB capacity and 1000TB capacity, difficulty the same, and realize to manage arbitrary node.
In addition, conventional store framework cannot intelligent distribution load, and multiple storage control can not the same item work of load simultaneously; And Supporting connectivity client dynamic load leveling of the present invention, support unlimited snapshot, be applicable to multi-user and many applied environments.Conventional store adopts controller architecture, is generally two controllers, and in service operation process, only have a controller in work, another one is in Status of Backups, and when master controller fault, backup controller is taken over, so be an active and standby mode of operation.And this distributed memory system disclosed in this invention is not exist with the form of controller, but exist with the form of memory node, the status of each memory node is impartial, so distributed memory system disclosed by the invention is main mode of operation, can be random after this just means client's side link to distributed storage device be connected on certain memory node, and can not load pressure be all carried on some or a small amount of node, serve the effect of dynamic load leveling.
From the report of survey institute, adopt distributed memory system of the present invention, the management resource of client reduces 97%, is specially from use six full-time resources to 15% of a use resource.And the space utilisation of these clients improves 22%, and required memory capacity decreases 37%.
Although content of the present invention has done detailed introduction by above preferred embodiment, will be appreciated that above-mentioned description should not be considered to limitation of the present invention.After those skilled in the art have read foregoing, for multiple amendment of the present invention and substitute will be all apparent.Therefore, protection scope of the present invention should be limited to the appended claims.
Claims (10)
1. a distributed memory system, is characterized in that, this system comprises:
Storage pool, its network connects external client, is stored file by the external client upload of network reception;
File block module, it communicates to connect storage pool, will be stored File striping and be processed into several blocks of files;
Some memory nodes, it communicates to connect described file block module respectively, and the some blocks of files being stored file block generation are stored in each memory node respectively.
2. distributed memory system as claimed in claim 1, it is characterized in that, each described memory node is a physical storage device; Or some memory nodes are several memory modules in a physical storage device.
3. distributed memory system as claimed in claim 1, is characterized in that, described memory node adopts InfiniBand framework or gigabit level Ethernet to carry out communication with synchronous each other.
4. distributed memory system as claimed in claim 1 or 2, is characterized in that, described storage pool is formed for some memory nodes provide part memory space to be polymerized respectively.
5. a file memory method for distributed memory system, is characterized in that, this file memory method comprises:
External client will be stored files passe to distributed memory system by network;
Distributed memory system carries out striping process to being stored file, is divided into some blocks of files;
The some blocks of files being stored file are stored in some memory nodes of distributed memory system respectively.
6. file memory method as claimed in claim 5, it is characterized in that, described distributed memory system stores successively to some files that is stored;
Distributed memory system store each be stored file time, this is stored to file and carries out striping process, and completes after the blocks of files this being stored file is arranged in some memory nodes, then store the next one and be stored file.
7. file memory method as claimed in claim 5, it is characterized in that, described distributed memory system carries out image copying to each blocks of files be stored in its each memory node, original block copies and produces one or several image file blocks, one or several other the memory node outside the memory node that a or some parts of image file blocks are stored in original block place.
8. file memory method as claimed in claim 5, it is characterized in that, described distributed memory system complete to be stored file striping process and after each blocks of files is stored in each memory node, file edit check code is stored to this, check code is stored in respectively in each memory node; When wherein arbitrary memory node replaces with new memory node, other memory nodes do not replaced according to check code by the date restoring on the memory node that is replaced to new memory node.
9. a file reading for distributed memory system, is characterized in that, this file reading comprises:
External client proposes to distributed memory system the request that reading one is stored file;
Distributed memory system reads this respectively simultaneously and is stored some blocks of files that file is divided into from some memory nodes;
Some blocks of files combination reduction are stored file by distributed memory system, and export external client to by network.
10. file reading as claimed in claim 9, is characterized in that, during the memory node fault at the required blocks of files place of reading of described distributed memory system, distributed memory system reads this blocks of files and backs up the image file block in other memory nodes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510110507.7A CN104639661A (en) | 2015-03-13 | 2015-03-13 | Distributed storage system and storing and reading method for files |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510110507.7A CN104639661A (en) | 2015-03-13 | 2015-03-13 | Distributed storage system and storing and reading method for files |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104639661A true CN104639661A (en) | 2015-05-20 |
Family
ID=53217962
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510110507.7A Pending CN104639661A (en) | 2015-03-13 | 2015-03-13 | Distributed storage system and storing and reading method for files |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104639661A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104932953A (en) * | 2015-06-04 | 2015-09-23 | 华为技术有限公司 | Data distribution method, data storage method, and relevant device and system |
| CN105007330A (en) * | 2015-08-04 | 2015-10-28 | 电子科技大学 | Modeling method for storage resource scheduling model of distributed flow data storage system |
| CN105138276A (en) * | 2015-07-14 | 2015-12-09 | 苏州科达科技股份有限公司 | Data storage method and data storage system |
| CN105187489A (en) * | 2015-08-05 | 2015-12-23 | 深圳联友科技有限公司 | File transfer method and system capable of clustering and supporting multiple users to upload simultaneously |
| CN105356997A (en) * | 2015-08-06 | 2016-02-24 | 华南农业大学 | Security distributed data management method based on public cloud |
| CN106354587A (en) * | 2016-09-06 | 2017-01-25 | 上海云轴信息科技有限公司 | Mirror image server and method for exporting mirror image files of virtual machine |
| CN106445643A (en) * | 2016-11-14 | 2017-02-22 | 上海云轴信息科技有限公司 | Method and device for cloning and updating virtual machine |
| CN106550024A (en) * | 2016-10-24 | 2017-03-29 | 上海斐讯数据通信技术有限公司 | The storage of media file and acquisition methods, Portal server and terminal in the Portal pages |
| CN107438096A (en) * | 2016-05-25 | 2017-12-05 | 思科技术公司 | For the congestion aware load balance of distributed storage |
| CN107613005A (en) * | 2017-09-20 | 2018-01-19 | 携程旅游信息技术（上海）有限公司 | Reverse proxy method and device, electronic equipment, storage medium |
| CN107786374A (en) * | 2017-10-19 | 2018-03-09 | 郑州云海信息技术有限公司 | A kind of Oracle cluster file systems and its method for realizing fence |
| CN108769137A (en) * | 2018-05-08 | 2018-11-06 | 北京初志科技有限公司 | Distributed structure/architecture data storing and reading method and device based on multigroup framework |
| CN109144787A (en) * | 2018-09-03 | 2019-01-04 | 郑州云海信息技术有限公司 | A kind of data reconstruction method, device, equipment and readable storage medium storing program for executing |
| CN110458567A (en) * | 2019-08-23 | 2019-11-15 | 广东工业大学 | A kind of object trading management system and method based on block chain |
| CN110602159A (en) * | 2019-07-30 | 2019-12-20 | 广州力挚网络科技有限公司 | Data processing method and system |
| CN110609810A (en) * | 2018-05-29 | 2019-12-24 | 深圳市云帆加速科技有限公司 | File storage method and system |
| CN111124740A (en) * | 2019-12-11 | 2020-05-08 | 新华三大数据技术有限公司 | Data reading method and device, storage equipment and machine-readable storage medium |
| CN111444043A (en) * | 2020-03-27 | 2020-07-24 | 中移雄安信息通信科技有限公司 | Data backup method, device, node, system and medium |
| CN112394876A (en) * | 2019-08-14 | 2021-02-23 | 深圳市特思威尔科技有限公司 | Large file storage/reading method, storage/reading device and computer equipment |
| CN113312009A (en) * | 2021-07-29 | 2021-08-27 | 南京鹏云网络科技有限公司 | Distributed storage system and read-write method thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1567237A (en) * | 2003-06-09 | 2005-01-19 | 浪潮电子信息产业股份有限公司 | Method for constructing high-available distributed memory system |
| EP2284735A1 (en) * | 2002-11-14 | 2011-02-16 | Isilon Systems, Inc. | Systems and methods for restriping files in a distributed file system |
| CN103034739A (en) * | 2012-12-29 | 2013-04-10 | 天津南大通用数据技术有限公司 | Distributed memory system and updating and querying method thereof |
| CN103049407A (en) * | 2012-12-27 | 2013-04-17 | 华为技术有限公司 | System, device and method for data storage |
| CN103116552A (en) * | 2013-03-18 | 2013-05-22 | 华为技术有限公司 | Method and device for distributing storage space in distributed type storage system |
| CN104272274A (en) * | 2013-12-31 | 2015-01-07 | 华为技术有限公司 | Data processing method and device in distributed file storage system |
-
2015
- 2015-03-13 CN CN201510110507.7A patent/CN104639661A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2284735A1 (en) * | 2002-11-14 | 2011-02-16 | Isilon Systems, Inc. | Systems and methods for restriping files in a distributed file system |
| CN1567237A (en) * | 2003-06-09 | 2005-01-19 | 浪潮电子信息产业股份有限公司 | Method for constructing high-available distributed memory system |
| CN103049407A (en) * | 2012-12-27 | 2013-04-17 | 华为技术有限公司 | System, device and method for data storage |
| CN103034739A (en) * | 2012-12-29 | 2013-04-10 | 天津南大通用数据技术有限公司 | Distributed memory system and updating and querying method thereof |
| CN103116552A (en) * | 2013-03-18 | 2013-05-22 | 华为技术有限公司 | Method and device for distributing storage space in distributed type storage system |
| CN104272274A (en) * | 2013-12-31 | 2015-01-07 | 华为技术有限公司 | Data processing method and device in distributed file storage system |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104932953A (en) * | 2015-06-04 | 2015-09-23 | 华为技术有限公司 | Data distribution method, data storage method, and relevant device and system |
| CN105138276B (en) * | 2015-07-14 | 2018-05-18 | 苏州科达科技股份有限公司 | Data storage method and data storage system |
| CN105138276A (en) * | 2015-07-14 | 2015-12-09 | 苏州科达科技股份有限公司 | Data storage method and data storage system |
| CN105007330A (en) * | 2015-08-04 | 2015-10-28 | 电子科技大学 | Modeling method for storage resource scheduling model of distributed flow data storage system |
| CN105007330B (en) * | 2015-08-04 | 2019-01-08 | 电子科技大学 | The modeling method of the storage resource scheduling model of distributed stream data-storage system |
| CN105187489A (en) * | 2015-08-05 | 2015-12-23 | 深圳联友科技有限公司 | File transfer method and system capable of clustering and supporting multiple users to upload simultaneously |
| CN105356997A (en) * | 2015-08-06 | 2016-02-24 | 华南农业大学 | Security distributed data management method based on public cloud |
| CN105356997B (en) * | 2015-08-06 | 2019-09-06 | 华南农业大学 | The distributed data management method of safety based on public cloud |
| CN107438096A (en) * | 2016-05-25 | 2017-12-05 | 思科技术公司 | For the congestion aware load balance of distributed storage |
| CN106354587A (en) * | 2016-09-06 | 2017-01-25 | 上海云轴信息科技有限公司 | Mirror image server and method for exporting mirror image files of virtual machine |
| CN106550024A (en) * | 2016-10-24 | 2017-03-29 | 上海斐讯数据通信技术有限公司 | The storage of media file and acquisition methods, Portal server and terminal in the Portal pages |
| CN106445643A (en) * | 2016-11-14 | 2017-02-22 | 上海云轴信息科技有限公司 | Method and device for cloning and updating virtual machine |
| CN106445643B (en) * | 2016-11-14 | 2019-10-22 | 上海云轴信息科技有限公司 | It clones, the method and apparatus of upgrading virtual machine |
| CN107613005A (en) * | 2017-09-20 | 2018-01-19 | 携程旅游信息技术（上海）有限公司 | Reverse proxy method and device, electronic equipment, storage medium |
| CN107613005B (en) * | 2017-09-20 | 2021-04-13 | 携程旅游信息技术（上海）有限公司 | Reverse proxy method and device, electronic device and storage medium |
| CN107786374A (en) * | 2017-10-19 | 2018-03-09 | 郑州云海信息技术有限公司 | A kind of Oracle cluster file systems and its method for realizing fence |
| CN108769137A (en) * | 2018-05-08 | 2018-11-06 | 北京初志科技有限公司 | Distributed structure/architecture data storing and reading method and device based on multigroup framework |
| CN110609810A (en) * | 2018-05-29 | 2019-12-24 | 深圳市云帆加速科技有限公司 | File storage method and system |
| CN109144787A (en) * | 2018-09-03 | 2019-01-04 | 郑州云海信息技术有限公司 | A kind of data reconstruction method, device, equipment and readable storage medium storing program for executing |
| CN110602159A (en) * | 2019-07-30 | 2019-12-20 | 广州力挚网络科技有限公司 | Data processing method and system |
| CN112394876A (en) * | 2019-08-14 | 2021-02-23 | 深圳市特思威尔科技有限公司 | Large file storage/reading method, storage/reading device and computer equipment |
| CN112394876B (en) * | 2019-08-14 | 2024-02-23 | 深圳市特思威尔科技有限公司 | Large file storage/reading method, storage/reading device and computer equipment |
| CN110458567A (en) * | 2019-08-23 | 2019-11-15 | 广东工业大学 | A kind of object trading management system and method based on block chain |
| CN111124740A (en) * | 2019-12-11 | 2020-05-08 | 新华三大数据技术有限公司 | Data reading method and device, storage equipment and machine-readable storage medium |
| CN111444043A (en) * | 2020-03-27 | 2020-07-24 | 中移雄安信息通信科技有限公司 | Data backup method, device, node, system and medium |
| CN113312009A (en) * | 2021-07-29 | 2021-08-27 | 南京鹏云网络科技有限公司 | Distributed storage system and read-write method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104639661A (en) | Distributed storage system and storing and reading method for files | |
| US11645173B2 (en) | Resilient mediation between storage systems replicating a dataset | |
| US11113161B2 (en) | Local storage clustering for redundancy coded data storage system | |
| US20180024964A1 (en) | Disaggregated compute resources and storage resources in a storage system | |
| CN100452046C (en) | Storage method and system for mass file | |
| CN103929500A (en) | Method for data fragmentation of distributed storage system | |
| CN111158587B (en) | Distributed storage system based on storage pool virtualization management and data read-write method | |
| US20140380093A1 (en) | Resilient distributed replicated data storage system | |
| CN103763383A (en) | Integrated cloud storage system and storage method thereof | |
| CN103455577A (en) | Multi-backup nearby storage and reading method and system of cloud host mirror image file | |
| US7702757B2 (en) | Method, apparatus and program storage device for providing control to a networked storage architecture | |
| US20180307439A1 (en) | Lock-free raid implementation in multi-queue architecture | |
| CN103942112A (en) | Magnetic disk fault-tolerance method, device and system | |
| CN105426427A (en) | MPP database cluster replica realization method based on RAID 0 storage | |
| CN106066890A (en) | A kind of distributed high-performance data storehouse integrated machine system | |
| JP2023099186A (en) | Storage system spanning multiple failure domains | |
| US20100082793A1 (en) | Server-Embedded Distributed Storage System | |
| CN111897486A (en) | Intelligent unified storage system based on software definition | |
| US8117493B1 (en) | Fast recovery in data mirroring techniques | |
| Yu et al. | Design and architecture of dell acceleration appliances for database (DAAD): A practical approach with high availability guaranteed | |
| US10366062B1 (en) | Cycled clustering for redundancy coded data storage systems | |
| US10678664B1 (en) | Hybridized storage operation for redundancy coded data storage systems | |
| US20220394091A1 (en) | Storing a Data Object as Data Regions in a Storage Network | |
| Dell | ||
| CN111831230A (en) | Intelligent storage system for high-performance calculation and big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150520 |
|
| RJ01 | Rejection of invention patent application after publication |