CN102752381A - Multi-movable-duplicate mechanism applied to distributed storage and access method thereof - Google Patents
Multi-movable-duplicate mechanism applied to distributed storage and access method thereof Download PDFInfo
- Publication number
- CN102752381A CN102752381A CN2012102227754A CN201210222775A CN102752381A CN 102752381 A CN102752381 A CN 102752381A CN 2012102227754 A CN2012102227754 A CN 2012102227754A CN 201210222775 A CN201210222775 A CN 201210222775A CN 102752381 A CN102752381 A CN 102752381A
- Authority
- CN
- China
- Prior art keywords
- copy
- user
- data
- duplicate
- module
- 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
Images
Abstract
The invention provides a multi-movable-duplicate mechanism applied to distributed storage and an access method thereof, which are applicable to a distributed storage system with frequent reading operation. The dynamic duplicate access method is located on a metadata server of the distributed storage system, and the metadata server comprises a data node information maintenance module, a route analysis module, a user request processing module and a dynamic duplicate management module. A multi-movable-duplicate mechanism is provided, so that any duplicate in the system can be accessed by users at will, and the system reliability and input/output throughput can be improved; aiming at different application environments and user requirements, the quantity and the storage position of the duplicates can be dynamically adjusted, and the response speed of the system and the data flow transmission can be accelerated; and since the new duplicate has timeliness, the system can be enabled to rapidly respond to the access of the user, the space of a disc can also be saved, the resource waste can be avoided, and the system capacity can be increased.
Description
Technical field
The present invention relates to a kind of multi-activity copy mechanism and access method thereof that is applied to distributed storage, is a kind of new mechanism and the method that is applicable to distributed storage and cloud storage system, belongs to the Computer Science and Technology field.
Background technology
At big data age, distributed storage is the effective way that solves the mass data storage problem.The duplicate of the document strategy can improve the reliability and the access speed of distributed memory system, the equalizing system load as a kind of effective means that guarantees the data high availability.
Distributed memory system gathers collaborative work with a large amount of various dissimilar memory devices in the network through application software; And storage and Operational Visit function externally are provided jointly through certain application software or interface, realize that the distributed storage of mass data and parallel I/O visit.Its basic structure is as shown in Figure 1, mainly comprises meta data server (MS, Metadata servers), data server (DS, Data servers) and three parts of client (Client).Yet along with the expansion of system scale, hardware device or software architecture unavoidably can break down, thereby cause the damage of data and lose, and have influence on the normal access of data.
And the copy technology is duplicated data on the many parts of a plurality of nodes that are placed on distributed system; Data single or a plurality of nodes are damaged or lose; Even the inefficacy of whole node, can not cause can't visiting of copy, thereby can guarantee the integrality and the availability of data.And the dynamic copies strategy can be adjusted the number and the memory location of copy according to the change dynamics of system environments and application demand, thereby improves the response speed of system, and can realize load balancing.
Different application scenarios has also proposed different demands to replication policy.For the application that a read operation is concentrated, concern be the availability of data, need as much as possible copy is distributed on the wide node, improve the performance and the access speed of read operation; And the application of concentrating for a write operation, the consistency of data is even more important, need be with disposing in the copy set, and to make things convenient for writing and upgrading of data.
At present, the distributed memory system of main flow comprises GFS, HDFS, Dynamo, PVFS etc., and all there are two common problems in their replication policy.At first, when carrying out read operation, it is movable available having only a duplicate of the document usually, that is to say and have only a copy that data are provided that other copy just plays data backup and fault-tolerant effect.So just caused resource waste, and when the user capture amount was very big, the node at active copy place can produce I/O bottleneck or network congestion.Secondly, they do not deposit copy the open user of giving of realization of strategy, the user can't come the specified file copy to deposit according to the actual demand of oneself particular location.
In view of the situation, the present invention proposes a kind of multi-activity copy mechanism and access method thereof that is applied to distributed storage.This method guarantees at first that when reading of data each duplicate of the document all is movable available in the system, but user's arbitrary access any copy wherein overcomes the defective of single active copy; Secondly, system can dynamically adjust copy number and deposit position according to user's request, accelerates the visit of data.
Summary of the invention
Said to background technology; The present invention's " a kind of multi-activity copy mechanism and access method thereof that is applied to distributed storage " comprises two parts: multi-activity copy mechanism and based on this machine-processed dynamic copies access method; Be intended to solve the network congestion and the magnetic disc i/o bottleneck problem that are prone in the frequent distributed memory system of read operation, improve the reliability and the response speed of system.
1, multi-activity copy mechanism
So-called multi-activity copy mechanism is meant that each duplicate of the document all is movable available in the system, and the user is arbitrary access any copy wherein as required.
As shown in Figure 2, original document has three copies, leaves in respectively on the data server DS3 of data server DS1, DS2 and data center 2 of data center 1.At a time, when the user capture amount was very big, if having only copy 1 that data are provided, the data server DS1 that stores this copy so will reach the I/O bottleneck inevitably because of throughput is excessive, and the network at DS1 place also can produce congested.And multi-activity copy mechanism is through letting a plurality of copies that data are provided simultaneously, and user's visit is distributed on a plurality of data servers at different pieces of information center, increases the total magnetic disc i/o throughput of data server, reduces the offered load of individual server.At this moment, data server DS1, DS2 and DS3 externally provide access service jointly, and copy 1, copy 2 and copy 3 are read and write by the user simultaneously, thereby have avoided occurring above-mentioned situation.
2, dynamic copies access method
Dynamic copies access method based on multi-activity copy mechanism is positioned on the meta data server of distributed memory system, is a software module of distributed memory system.
Meta data server is mainly accomplished reading of multi-activity copy and dynamic adjustment function through four modules, comprises back end maintenance of information module, route analysis module, user's request processing module and dynamic copies administration module.Fig. 3 is the utilization relation of each module and information.Introduce each part below.
(1) data server maintenance of information module (DIM; Data server Information Module): the mapping relations of this module records whole system file to data block and data block to data server, according to the feedback information monitor data server of data server and the operating state of duplicate of the document.
(2) route analysis module (RAM, Routing Analysis Module): the routing relation between this module analysis data center and user and the data center, guarantee that data can transmit under best network condition, avoid network congestion.
(3) user's request processing module (URM; User Request Module): this module is carried out authentication to user's access rights; According to user's request corresponding metadata information is sent to the user, the user carries out data interaction according to these control informations and data server.
(4) dynamic copies administration module (DRM; Dynamic Replication Module). the reading and writing of this module management duplicate of the document; And can dynamically adjust copy number according to I/O, the network condition of user behavior and system, copy is left on the suitable data server.
A kind of dynamic copies access method based on multi-activity copy mechanism comprises 7 steps shown in Figure 4:
(1) data server maintenance of information: the operating state of data server in the DIM module monitors system, write down the metadata informations such as deposit position of I/O throughput, the network bandwidth and the duplicate of the document of each node.
(2) user writable Request Processing: URM module verification user's access rights, and the network site of recording user and routing condition.If user's request is write data, continue step (3); If user's request is read data, get into step (5).
(3) initialization of duplicate of the document storage: the URM module sends to the user with the control information of node, and the user transmits data to data server in view of the above, and the DRM module is carried out the copy storage.When not having the special applications demand, as shown in Figure 5, the copy number of DRM acquiescence is three.Wherein on copy 1 and copy 2 two different pieces of information server DS1 being placed on the local data center and the DS2, user and data center 1 carry out data interaction in same CDN territory, and network congestion is avoided in the transmission of quickening data flow; DRM adopts the Round-Robin method copy 3 to be placed on the data server DS3 of remote data center 2, guarantees the reliability of data.
(4) active of copy number and deposit position adjustment: under a lot of situation; The data that the user deposits at the local data center often are not to be read by the local user; But offer other users of far-end, when remote subscriber can cause network congestion from the local data center during a large amount of reading of data equally.Under the ideal situation, all deposit the fast access that a copy can guarantee the user in each data center, but this can cause the unnecessary wasting of resources, consume the bigger cost that administers and maintains.This just requires DRM to adjust according to the different application demand, the zone that DRM possibly used by a large number of users according to file, with copy push (PUSH) in this CDN territory in the data center of correspondence buffer memory get up.As shown in Figure 6, DRM initiatively is pushed to copy data on data center 3 and 4, and buffer memory, the conveniently visit of same CDN intra domain user on data server DS4 and DS5.
This step is the preprocessing process of copy, mainly accomplishes through human configuration.Because this step is that DRM initiatively accomplishes, therefore be referred to as initiatively adjustment process.
(5) selection of data center and data server: the RAM module finds the user to belong to data center and data server in the CDN territory according to user's network site and routing condition.
(6) reading of multi-activity copy: the URM module sends to the user according to the selection of RAM with corresponding metadata information.If the copy data that has the user to ask on the data server of this data center, then user and data server connect and carry out reading of multi-activity copy; Otherwise continue the adjustment second time that step (7) is carried out copy.
(7) the passive adjustment of copy number and deposit position: as shown in Figure 7; When data that the data center that is connected with the user (for example figure in data center 5 and 6) does not have to ask; Data center can start (PULL) method that pulls of DRM; The copy data of other data center (data center 1 among the figure) storage is cached on the local data server (data server DS6 and DS9), carries out data interaction with the user then; When the user capture amount is very big; The magnetic disc i/o throughput of storing the data cached server of this copy (for example data server DS6) may reach bottleneck very soon; Cause can't visiting of data; At this moment DRM can go up newly-built other buffer memory at other data server (DS7 and DS8) of this data center, and service outwards is provided simultaneously, to improve the I/O performance of system.Along with increasing of copy number, the memory space of system can be linear increase tendency, be to reduce the storage pressure of system, increases memory capacity, and copy data cached has actual effect usually, is defaulted as 24 hours.
This step is the passive completion of DRM, therefore is referred to as passive adjustment process.This method is the optimization to the copy adjustment first time, here copy number and with the adjustment of deposit position be demand assigned.
Can accomplish dynamic copies access procedure according to above step based on multi-activity copy mechanism.
Description of drawings
The basic structure of Fig. 1 distributed memory system
The basic principle of Fig. 2 multi-activity copy mechanism
Fig. 3 is based on each module and the utilization relation of the dynamic copies access method of multi-activity copy mechanism
Fig. 4 is based on the dynamic copies access step of multi-activity copy mechanism
The initialization storing process of Fig. 5 copy
The active adjustment process of Fig. 6 copy
The passive adjustment process of Fig. 7 copy
Fig. 8 is based on the dynamic copies access method use-case of multi-activity copy mechanism
Embodiment
To combine the accompanying drawing in the embodiment of the invention below, the technical scheme in the embodiment of the invention is carried out clear, intactly description, obviously, described embodiment also only is a part of embodiment of the present invention, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills are not making the every other embodiment that is obtained under the creative work prerequisite, all belong to the scope of the present invention's protection.
For a kind of dynamic copies access method based on multi-activity copy mechanism is described, provided an instance here.Instance provides Video service this basis that is applied as with certain portal website in the actual life, supposes that distributed memory system has Beijing, Shanghai, Guangzhou, Chengdu, five data centers, Wuhan, and there are hundred data servers in each data center.Relevant parameters is: connect with gigabit Ethernet between meta data server and the data center, network throughput is B
t: the maximum magnetic disc i/o of every data server of 1000M/S is U
Max: 300M/S; The data traffic at user capture local data center is V
T1: 1M/S, visit other data centers and then reduce to V
T2: 200KB/S.
Suppose that reporter A has taken the canto video from Guangzhou, video size is M:500M, need video data be offered user capture, adopts based on the dynamic copies access method of multi-activity copy mechanism and mainly accomplishes through four-stage as shown in Figure 8.
(1) default storage of copy: meta data server according to the choice of location at reporter place Guangzhou data center, and the metadata information of this data center sent to the client of reporter A, reporter A and Guangzhou data center connect and begin to upload data subsequently; Simultaneously meta data server adopts the mode of poll to select Wuhan data center, and when reporter A uploaded data, Guangzhou data center also can data center backs up to Wuhan with transfer of data; After accomplishing in the system copy number be 3, be respectively copy 1, copy 2, copy 3.Video at the uplink time of this locality is: T
Up=M/V
T1=500/1=500S, and the copy maximum delay that data center stores Wuhan into from Guangzhou is: T
Delay=M/B
t=500/1000=0.5S; If copy directly uploads to Wuhan data center from the user, required time is: T
Up2=M/V
T2=500/0.2=2500S, visible T
Up2Be far longer than T
UpAnd T
DelaySummation.Therefore the user should carry out with the data center in the same CDN territory alternately, and the transmission of copy data should be carried out between data center as far as possible.
(2) active of copy number and deposit position adjustment: the situation that video data maybe the user capture by is not where considered in the initialization storage of copy, and meta data server can tentatively be adjusted this according to video information.Suppose that the video that reporter A takes is the match of Beijing team and Shanghai team; For making things convenient for Beijing and the user capture of two places, Shanghai; The nearest copy (being respectively Wuhan and Guangzhou) of meta data server meeting chosen distance; Initiatively copy data is pushed (PUSH) data center to Beijing and Shanghai, the copy number in the system becomes 5 at this moment, has increased cached copies 4 and cached copies 5 newly.
(3) the passive adjustment of copy number and deposit position: the active adjustment of copy has certain subjectivity, can not satisfy actual conditions fully.Area, Chengdu user for example, several team members arranged from Chengdu in two teams in Beijing and Shanghai, so also can visit in a large number.Read when request when the user sends, this method is to send to the user to the metadata information that the user belongs to the data center in the CDN territory.And do not have copy data in the data center of Chengdu this moment; So its can pull (PULL) data row cache of going forward side by side from nearest data center (being Guangzhou) here; Send to the user with data cached simultaneously, this moment, copy number became 6, had increased cached copies 6 newly.The download time of video is like this: T
Download=M/V
T1+ T
Delay=500/1+0.5=500.5S; If directly data center visits then that the time is: T from Wuhan
Download2=M/V
T2=500/0.2=2500S.This shows that the efficiency far that first buffer memory reads again is higher than the efficient that directly reads.
(4) adjustment of buffer memory number: for the very high video of temperature, when user capture amount at a time was very big, the server of storing this video copy possibly reach the magnetic disc i/o bottleneck.In this example, when having 300 people to visit simultaneously in the data server per second at Beijing data center cached copies 5 places, total visit capacity of user: U
Sum=300*1=U
Max, server has reached transmission bottleneck.For avoiding this situation to take place, meta data server can be provided with threshold value a: U
Threshold=0.8*U
MaxWhen the user capture amount reaches threshold value, i.e. U
Sum=U
ThresholdThe time, data center can set up the buffer memory of copy on the another one data server, service externally is provided jointly, like cached copies newly-built among the figure 7 and cached copies 8.At this moment, the I/O throughput of data has enlarged 2 times, has reached 900M/S.Certainly, this is a cost to consume disk space, so the buffer memory of copy has actual effect usually.If in 24 hours, there is not user's visit, system can delete automatically that these are data cached, only keeps the copy under the initial situation, thereby avoids the waste of resource, improves power system capacity.
The major technique advantage
In order to improve the reliability of system, distributed memory system has all adopted the duplicate of the document strategy.Yet current distributed memory system mostly is based on the single-unit activity copy, can only operate a copy at any time, and the user is difficult to specify according to own needs the deposit position of copy.The present invention is based on multi-activity copy mechanism and proposed a kind of novel dynamic access copy method, copy can duplicate and store as required.
Advantage of the present invention is: (1) has proposed multi-activity copy mechanism; Be distributed to through visit on a plurality of data servers at different pieces of information center the user; Increase the magnetic disc i/o throughput of data server, reduced the offered load of individual server, avoided the waste of resource; (2) system also can arrive the data center in the different CDN territory with the copy data active push according to application demand on the basis of initialization storage, makes the user can obtain the data of request fast from the local data center; (3) meta data server is not to send to the user to the data server at copy place when the read request of process user; But send to the user to the data center in the same CDN territory; System carries out buffer memory according to user's request to duplicate of the document; Thereby realized distribution according to need, in the high efficiency of transmission that guarantees data, improved system memory size.This invention does not relate to data concrete storage mode in distributed memory system, and the application that read operation is concentrated has good effect.
Claims (3)
1. multi-activity copy mechanism has been proposed; The All Files copy of storing in the distributed memory system all is movable available at any time; The user is arbitrary access any copy wherein as required; Be distributed to through visit on a plurality of data servers at different pieces of information center, increase the magnetic disc i/o throughput, reduce the offered load of individual server the user.
2. proposed the dynamic adjusting method of copy number and deposit position, according to different application scenes and user's request, system is adjustment copy number and deposit position dynamically, improves the response speed and the reliability of system, mainly realizes through two steps:
The active of system adjustment when (1) user writes file, meta data server initiatively is pushed to copy data maybe be by the enterprising row cache of data server in the area of a large amount of visits.
The passive adjustment of system when (2) user reads file, the data center that is connected with user notice meta data server transmission data are also set up the copy buffer memory.
3. the copy buffer memory has ageingly, if in effective time, there is not user capture, these copies will be deleted by system, avoid this method in implementation process because of setting up the disk space outside a plurality of copy buffer memory occupying volumes, improve the memory capacity of system.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012102227754A CN102752381A (en) | 2012-06-28 | 2012-06-28 | Multi-movable-duplicate mechanism applied to distributed storage and access method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012102227754A CN102752381A (en) | 2012-06-28 | 2012-06-28 | Multi-movable-duplicate mechanism applied to distributed storage and access method thereof |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102752381A true CN102752381A (en) | 2012-10-24 |
Family
ID=47032279
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2012102227754A Pending CN102752381A (en) | 2012-06-28 | 2012-06-28 | Multi-movable-duplicate mechanism applied to distributed storage and access method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102752381A (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102932442A (en) * | 2012-10-26 | 2013-02-13 | 北京奇虎科技有限公司 | Equipment, system and method for distributed data storage system |
CN102932440A (en) * | 2012-10-26 | 2013-02-13 | 北京奇虎科技有限公司 | Data management method, equipment and system for distributed data storage system |
CN103095832A (en) * | 2013-01-15 | 2013-05-08 | 北京邮电大学 | Distributed storage method based on communication reliability |
CN103152396A (en) * | 2013-02-05 | 2013-06-12 | 华南师范大学 | Data placement method and device applied to content distribution network system |
CN103475709A (en) * | 2013-09-09 | 2013-12-25 | 国家电网公司 | Method and system for ectype dynamic control for cloud computing platform |
CN104754021A (en) * | 2013-12-31 | 2015-07-01 | 伊姆西公司 | Device and method for promoting access of data in distributed storage system |
CN106648456A (en) * | 2016-09-18 | 2017-05-10 | 重庆邮电大学 | Dynamic save file access method based on use page view and prediction mechanism |
CN106897440A (en) * | 2017-02-28 | 2017-06-27 | 郑州云海信息技术有限公司 | A kind of distributed file system file read-write processing method |
CN107247638A (en) * | 2012-12-25 | 2017-10-13 | 华为技术有限公司 | A kind of data back up method, device and system |
CN107707926A (en) * | 2017-03-09 | 2018-02-16 | 贵州白山云科技有限公司 | A kind of live TV stream transmission method, device and system |
CN107911710A (en) * | 2017-11-30 | 2018-04-13 | 上海佰贝科技发展有限公司 | A kind of live resource-sharing management platform and method |
CN108319618A (en) * | 2017-01-17 | 2018-07-24 | 阿里巴巴集团控股有限公司 | A kind of data distribution control method, system and the device of distributed memory system |
CN109471847A (en) * | 2018-09-18 | 2019-03-15 | 华中科技大学 | A kind of I/O jamming control method and control system |
CN110166542A (en) * | 2019-05-16 | 2019-08-23 | 广东电网有限责任公司信息中心 | Distributed memory system and data read-write method based on ICP/IP protocol |
CN110324429A (en) * | 2019-07-10 | 2019-10-11 | 中国工商银行股份有限公司 | Backup method and back-up device based on Distributed Storage |
CN111225027A (en) * | 2018-11-26 | 2020-06-02 | 丰田自动车株式会社 | Mobility-oriented data replication in vehicular micro-clouds |
CN111488324A (en) * | 2020-04-14 | 2020-08-04 | 浪潮商用机器有限公司 | Distributed network file system based on message middleware and working method thereof |
CN113010549A (en) * | 2021-01-29 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Data processing method based on remote multi-active system, related equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1881898A (en) * | 2006-05-12 | 2006-12-20 | 湖南大学 | Non-structured P2P copy consistency maintaining method based on node locus label |
CN101187931A (en) * | 2007-12-12 | 2008-05-28 | 浙江大学 | Distribution type file system multi-file copy management method |
CN101753625A (en) * | 2009-12-28 | 2010-06-23 | 北京理工大学 | Method for deployment of copy service and copy establishment in peer-to-peer network environment |
-
2012
- 2012-06-28 CN CN2012102227754A patent/CN102752381A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1881898A (en) * | 2006-05-12 | 2006-12-20 | 湖南大学 | Non-structured P2P copy consistency maintaining method based on node locus label |
CN101187931A (en) * | 2007-12-12 | 2008-05-28 | 浙江大学 | Distribution type file system multi-file copy management method |
CN101753625A (en) * | 2009-12-28 | 2010-06-23 | 北京理工大学 | Method for deployment of copy service and copy establishment in peer-to-peer network environment |
Non-Patent Citations (2)
Title |
---|
卢显良: "面向Internet的大规模分布式存储技术研究", 《万方数据知识服务平台》, 18 August 2005 (2005-08-18) * |
李挺屹,王劲林,王玲芳: "分布式媒体服务系统中的副本放置算法", 《计算机工程》, 31 January 2010 (2010-01-31) * |
Cited By (27)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102932440A (en) * | 2012-10-26 | 2013-02-13 | 北京奇虎科技有限公司 | Data management method, equipment and system for distributed data storage system |
CN102932442A (en) * | 2012-10-26 | 2013-02-13 | 北京奇虎科技有限公司 | Equipment, system and method for distributed data storage system |
CN107247638A (en) * | 2012-12-25 | 2017-10-13 | 华为技术有限公司 | A kind of data back up method, device and system |
CN107247638B (en) * | 2012-12-25 | 2020-12-15 | 华为技术有限公司 | Data backup method, device and system |
CN103095832A (en) * | 2013-01-15 | 2013-05-08 | 北京邮电大学 | Distributed storage method based on communication reliability |
CN103152396A (en) * | 2013-02-05 | 2013-06-12 | 华南师范大学 | Data placement method and device applied to content distribution network system |
CN103475709A (en) * | 2013-09-09 | 2013-12-25 | 国家电网公司 | Method and system for ectype dynamic control for cloud computing platform |
CN103475709B (en) * | 2013-09-09 | 2017-01-04 | 国家电网公司 | A kind of copy dynamic control method towards cloud computing platform and system thereof |
CN104754021B (en) * | 2013-12-31 | 2018-04-13 | 伊姆西公司 | Apparatus and method for promoting the access to the data in distributed memory system |
CN104754021A (en) * | 2013-12-31 | 2015-07-01 | 伊姆西公司 | Device and method for promoting access of data in distributed storage system |
CN106648456B (en) * | 2016-09-18 | 2019-05-31 | 重庆邮电大学 | Dynamic copies file access method based on user's amount of access and forecasting mechanism |
CN106648456A (en) * | 2016-09-18 | 2017-05-10 | 重庆邮电大学 | Dynamic save file access method based on use page view and prediction mechanism |
CN108319618A (en) * | 2017-01-17 | 2018-07-24 | 阿里巴巴集团控股有限公司 | A kind of data distribution control method, system and the device of distributed memory system |
CN106897440B (en) * | 2017-02-28 | 2020-06-26 | 郑州云海信息技术有限公司 | File read-write processing method for distributed file system |
CN106897440A (en) * | 2017-02-28 | 2017-06-27 | 郑州云海信息技术有限公司 | A kind of distributed file system file read-write processing method |
WO2018161933A1 (en) * | 2017-03-09 | 2018-09-13 | 贵州白山云科技有限公司 | Live broadcast stream transmission method, apparatus and system, and computer readable storage medium, and computer device |
CN107707926A (en) * | 2017-03-09 | 2018-02-16 | 贵州白山云科技有限公司 | A kind of live TV stream transmission method, device and system |
CN107911710B (en) * | 2017-11-30 | 2021-04-23 | 上海佰贝科技发展股份有限公司 | Live broadcast resource sharing management platform and method |
CN107911710A (en) * | 2017-11-30 | 2018-04-13 | 上海佰贝科技发展有限公司 | A kind of live resource-sharing management platform and method |
CN109471847B (en) * | 2018-09-18 | 2020-06-09 | 华中科技大学 | I/O congestion control method and control system |
CN109471847A (en) * | 2018-09-18 | 2019-03-15 | 华中科技大学 | A kind of I/O jamming control method and control system |
CN111225027A (en) * | 2018-11-26 | 2020-06-02 | 丰田自动车株式会社 | Mobility-oriented data replication in vehicular micro-clouds |
CN110166542A (en) * | 2019-05-16 | 2019-08-23 | 广东电网有限责任公司信息中心 | Distributed memory system and data read-write method based on ICP/IP protocol |
CN110324429A (en) * | 2019-07-10 | 2019-10-11 | 中国工商银行股份有限公司 | Backup method and back-up device based on Distributed Storage |
CN111488324A (en) * | 2020-04-14 | 2020-08-04 | 浪潮商用机器有限公司 | Distributed network file system based on message middleware and working method thereof |
CN111488324B (en) * | 2020-04-14 | 2023-12-29 | 浪潮商用机器有限公司 | Distributed network file system based on message middleware and working method thereof |
CN113010549A (en) * | 2021-01-29 | 2021-06-22 | 腾讯科技(深圳)有限公司 | Data processing method based on remote multi-active system, related equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102752381A (en) | Multi-movable-duplicate mechanism applied to distributed storage and access method thereof | |
US10956246B1 (en) | Isolated read channel management interfaces at streaming data service | |
US7047287B2 (en) | Method and apparatus for automatically adapting a node in a network | |
US20150215405A1 (en) | Methods of managing and storing distributed files based on information-centric network | |
US20190163371A1 (en) | Next generation storage controller in hybrid environments | |
US11075984B1 (en) | Workload management at streaming data service supporting persistent connections for reads | |
US20020133491A1 (en) | Method and system for managing distributed content and related metadata | |
CN105025053A (en) | Distributed file upload method based on cloud storage technology and system | |
WO2011044829A1 (en) | Resource buffering method, resource acquisition method, apparatus and system thereof | |
EP1368948A2 (en) | Method and apparatus for large payload distribution in a network | |
Chen et al. | E2FS: an elastic storage system for cloud computing | |
JP5259835B2 (en) | Data node device, peer information acquisition method and system | |
US20110302242A1 (en) | File system and method for delivering contents in file system | |
US11675501B2 (en) | Streaming data service with isolated read channels | |
US11509700B2 (en) | Stream data record reads using push-mode persistent connections | |
Honnutagi | The Hadoop distributed file system | |
WO2021082401A1 (en) | Data uploading method, system and apparatus, and electronic device | |
CN104702702A (en) | System and method for downloading data | |
US20210084102A1 (en) | Isolated read channel categories at streaming data service | |
US11070600B1 (en) | Optimization techniques to support lagging readers at streaming data service | |
Huang et al. | BufferBank: a distributed cache infrastructure for peer-to-peer application | |
Sajithabanu et al. | Cloud based content delivery network using genetic optimization algorithm for storage cost | |
Chen et al. | BufferBank storage: an economic, scalable and universally usable in-network storage model for streaming data applications. | |
Chen et al. | Distributing very-large content from cloud to smart home hubs: Measurement and implications | |
Haridas et al. | CAPS: A cloud-assisted approach to handle spikes in peer-to-peer web search |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20121024 |