CN103327116B - The dynamic copies storage means of a kind of network file - Google Patents
The dynamic copies storage means of a kind of network file Download PDFInfo
- Publication number
- CN103327116B CN103327116B CN201310282367.2A CN201310282367A CN103327116B CN 103327116 B CN103327116 B CN 103327116B CN 201310282367 A CN201310282367 A CN 201310282367A CN 103327116 B CN103327116 B CN 103327116B
- Authority
- CN
- China
- Prior art keywords
- user
- file
- class
- list
- storage
- 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.)
- Active
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The dynamic copies storage means of a kind of network file, belong to technical field of distributed memory, system comprises network disk, MySQL database server and distributed storage server cluster, and wherein, distributed storage server cluster comprises a Controlling vertex and multiple memory node. In the present invention, for distributed memory system arranges three duplicate ratings (1 to 3), according to logging in, how many classifying rationallies are 3 classes by the number of times that logs in of user to utilize K-means algorithm, correspondingly for the user of each class is according to logging in the different copy number of set of frequency. In the present invention, the method is applied in distributed memory system, it is possible to according to use frequency, the file of network disk user is carried out rational distributed storage, both ensure that the reliability of storage, turn improved the utilization ratio of storage resources.
Description
Technical field
The present invention relates to the dynamic copies storage means of a kind of network file, belong to technical field of distributed memory.
Background technology
Along with the fast development of network and information technology, the terminal of modernization such as thin-client, mobile phone etc. are just being full of routine work, the studying and living of people with a kind of irresistible strength. But, owing to the storage space of terminal is very limited, how to enable people oneself privately owned file be carried out effective store and management whenever and wherever possible and just become very meaningful. At present, network disk is as very popular in 115 net dishes, Kingsoft net dish etc., if and file data is centrally stored on a server by this kind of network storage, so the performance of storage server will become the bottleneck of system performance, also faces the problem such as reliability and security. So, along with the deep development of cloud computing, distributed document system is arisen at the historic moment, storing device dissimilar in a large number is gathered collaborative work by it, data scatter is stored on the equipment of multiple stage independence, forms the data-storage system of a safety, and access interface is externally provided, and locate storage information by management server, substantially increase the reliability of storage system, operability, extensibility and storage efficiency. Therefore distributed document system can be combined with network disk, it is achieved the distributed storage of network file, if the patent No. is 201110448040.9, name is called that the patent of " operation method that a kind of file distribution based on thin-client stores " just belongs to this row.
In view of the user of network disk uses frequency to have very big-difference, and existing distributed file storage system, if Hadoop, MooseFS etc. are the reliability improving storage, often each file is arranged identical copy number, so to a certain extent, this is the waste of a kind of storage resources. So, if can for using user frequently to arrange higher copy number, and use the user that frequency is very low that relatively low copy number is set just can rational utilization storage resources. The present invention is just based on this thought, propose the dynamic copies storage means of a kind of network file, for distributed memory system arranges three duplicate ratings (1 to 3), the number of times that logs in of user is carried out real-time statistical updating, how many classifying rationallies are 3 classes according to logging in utilize K-means algorithm that user logs in number of times, correspondingly for the user of each class arranges different copy number. Thus, the method is applied in distributed memory system, so that it may with the use frequency according to user, the file of network disk user is carried out rational distributed storage, both ensure that the reliability of storage, turn improve the utilization ratio of storage resources.
Summary of the invention
For the deficiency that the storage restriction of terminal and the copy of existing distributed file storage system of modernization are arranged, the present invention proposes the dynamic copies storage means of a kind of network file.
The technical scheme of the present invention is as follows:
A dynamic copies storage means for network file, is realized by following system, and this system comprises network disk, MySQL database server and distributed storage server cluster, and wherein network disk is the network file memory disk of sing on web; MySQL database server is the database server being provided with MySQL; Distributed storage server cluster is a distributed file storage system, it comprises a Controlling vertex (ControlNode) and multiple memory node (StorageNode), wherein, Controlling vertex is used for the file operation (create, heavily name, delete) of supervisory user storage directory, and safeguard user file copy list, file storage location list and clustered node list, memory node is the real deposit position of user file, is communicated by TCPSocket between Controlling vertex and memory node. Pass between this system three part is: the user profile table (users) that store network disk on MySQL database server, the Apache of distributing network disk and user's storage directory on the Controlling vertex of distributed storage server cluster, the network request of all users is all that maincenter processes taking Controlling vertex, the file of user is finally stored on memory node, and these storage means concrete steps are as follows:
1) on MySQL database server, user profile table is created, UserName, LoginCount and RepNum field is set, wherein UserName represents the user name of network disk, LoginCount represents that user's logs in number of times, RepNum represents the duplicate of the document number of user, and LoginCount can log in the increase of number of times along with user and constantly update;
2) Controlling vertex reads LoginCount field from user profile table, and is write and output in local LoginCount.txt file by the number of times that logs in of all users;
3) utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class;
4) class border and the class center of three classes is obtained respectively, i.e. minimum value, class center, maximum value, and left in an array, and then this array is sorted, thus by all class borders and class center according to depositing in array by little to big order, if A1 is the minimum value of class A, A2 is the class center of class A, A3 is the maximum value of class A, same B1, B2, B3 is the minimum value of class B respectively, class center and maximum value, C1, C2, C3 is the minimum value of class C respectively, class center and maximum value, then according to by little to numerical ordering in array being greatly: A1 < A2 < A3 < B1 < B2 < B3 < C1 < C2 < C3,
5) from user profile table, again read LoginCount field, determine the duplicate of the document number of all users according to the class border of A, B, C tri-classes:
If a) user logs in number of times and is less than or equal to A3, then RepNum=1;
If b) user logs in number of times and is more than or equal to B1 and is less than or equal to B3, then RepNum=2;
If c) user logs in number of times and is more than or equal to C1, then RepNum=3;
Again the value of the RepNum field of all users is upgraded;
6) for each user, using the value of UserName field as " key ", using the value of the RepNum field of its correspondence as " value ", deposit in user file copy list with the form of key value pair;
7) each memory node is as the server end of TCPSocket, opens monitoring, waits the connection of client terminal;
8) user's storage directory of network disk is carried out monitor in real time by Controlling vertex, once there be user to upload file, just proceeds to step 9) continue to perform;
9) obtaining the upper transmitting file path of this user, intercept user name and filename from this path, traversal user file copy list, until finding the position identical with this user name, reads the duplicate of the document number of its correspondence;
10) Controlling vertex is according to duplicate of the document number, Distribution Algorithm is utilized to obtain the memory node list (node number in this list with duplicate of the document number identical) corresponding with this filename from clustered node list, and using filename as " key ", using the memory node list that obtains as " value ", by this key value to joining in file storage location list;
11) Controlling vertex is as the client terminal of TCPSocket, sends Socket connection request with the pattern of multi-thread journey to each memory node in memory node list, and after setting up connection, transmits this file to the memory node of correspondence. Thus, the many copies completing this file store.
Above-mentioned steps 3) in utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class, it is as follows that it runs step:
<1>number by preparation classification is appointed as 3, and the path simultaneously arranging input file is above-mentioned steps 2) in the storage location of LoginCount.txt file;
<2>from input file, read each row data line by line, obtain all sampled datas and deposit in lists;
<3>from sampled data list, select the initial classes center of 3 sampled datas as 3 classes at random;
<4>travel through sampled data list, calculate the distance of each data to 3 Ge Lei centers, and these data are referred in the class nearest with 3 class width between centers;
<5>recalculate the mean value of each class, using this mean value as Xin Lei center, and upgrade the center of each class;
Whether the new and old center<6>comparing each class is identical, if identical, then algorithm terminates; Otherwise proceed to step<4>, continue circulation execution and go down.
Above-mentioned network disk is the network file storage application system being deployed under LAMP framework, and its source file is based on PHP language compilation, comprises JavaScript, HTML, CSS, PHP and database file. Wherein, JavaScript be a kind of based on object and event-driven and there is the client script language of relative safety, be also a kind of extensively for the script language of client terminal Web exploitation simultaneously, be commonly used to add dynamic function to html web page; CSS is the abbreviation of English CascadingStyleSheets (CSS list), and it is a kind of machine language being used for showing the file styles such as HTML or XML.
Above-mentioned LAMP is one group of open source software being used for building dynamic website or server, comprises Linux, Apache, MySQL and PHP, jointly constitutes a powerful Web application platform. Wherein, Linux is in lowermost layer, it is provided that operating system; Secondary low layer is Apache, and it is a Web server, and Apache provides and user can be allowed to obtain the mechanism of Web page; MySQL is a mini-relational type data base management system (DBMS), stores data; PHP is the abbreviation of English super Text Pretreatment language HypertextPreprocessor, meaning super Text Pretreatment, embedded language in its a kind of HTML is the script language of a kind of embedding html document performed at server end, the style of language is similar to C language, is used widely.
Above-mentioned Apache is a Web server, and it can provide the mechanism allowing user obtain Web page.
Above-mentioned MySQL is a mini-relational type data base management system (DBMS), for storing data.
Above-mentioned user file copy list is that the key value of the duplicate of the document number of user name and its correspondence is to set.
Above-mentioned file storage location list is that the key value of the memory node IP address of filename and its correspondence is to set.
Above-mentioned clustered node list is the set of all memory node IP addresses in distributed storage server cluster.
Above-mentioned TCP be a kind of SCCP connection-oriented, reliably, based on the transportation level communication protocol of byte stream.
Above-mentioned IP address is a kind of mode to main frame addressing on internet, also referred to as internet protocol address. Common IP address, is divided into the big class of IPv4 and IPv6 two.
Above-mentioned K-means algorithm is hard cluster algorithm, its basic thought be random given K sample value as K initial classes center, using Europe formula distance as similarity measure, according to most proximity principle, sample value to be sorted is assigned to each class. Then the center of each class is recalculated by statistical method, so that it is determined that new center. Iteration always, until meeting given standard to estimate function.
Above-mentioned Euclidean distance also claims Euclidean distance, and it is a distance definition usually adopted, and is true distance between two points in m dimension space.
Above-mentioned multi-thread journey is to make multiple thread parallel work to complete the multi-task, raise the efficiency.
Above-mentioned thread has needed at one time to be implemented when the multi-task. Each program just run in system is a process, and each process comprises one or more threads. Thread is the set of one group of instruction, or the particular segment of program, and it can independent in program perform.
Above-mentioned Socket, also referred to as " socket ", for describing IP address and port, is the handle of a communication chain, and application program sends request or response network request usually by " socket " to network, and it comprises server end and client terminal. According to the target that the mode and local socket that connect startup to be connected, the connection procedure between socket can be divided into three steps: server is monitored, client-requested, connects and confirms. Wherein, server is monitored and is server end socket and the concrete client socket of no fix, but the state to be connected such as is in, monitor in real time network state; Client-requested refers to and proposes connection request by the socket of client terminal, and the target that connect is the socket of server end; Connect and confirm to refer to when server end socket listens to or say the connection request receiving client socket, it is with regard to the request of customer in response end socket, set up a new thread, client terminal is issued in the description of server end socket, describe once client terminal confirms this, connect and just established. And server end socket keeps monitoring state, continue to receive the connection request of other client socket.
Why the present invention uses K-means algorithm that network disk user being logged in, how many number of times is divided into three classes according to logging in, instead of use numerical evaluation Direct Classification, it is because the number of times that logs in of network disk user exists very big individual difference, namely number of times distribution is logged in very uneven, if according to formula:It is divided into three segments, then user's copy number can be caused to arrange very unreasonable. And using K-means algorithm effectively how much can rationally be divided into three classes by logging in number of times according to logging in so that resolution height between the class of three classes, and compactness height in class, arrange duplicate of the document number according to this and just more rationalize.
The present invention logs in number of times according to network disk user's, utilize K-means algorithm that difference uses the user of frequency be rationally divided into three classes, the user file copy number using frequency lower is set to 1, the user file copy number using frequency higher is set to 3, and remaining user file copy number is set to 2. The advantage of operation method of the present invention is to apply in distributed memory system by dynamic copies storage means, just according to the use frequency of user, the file of net dish user can be carried out rational copy setting, both ensure that the reliability of storage, turn improve the utilization ratio of storage resources.
Accompanying drawing explanation
Embodiment
Below in conjunction with embodiment, the present invention will be further described, but is not limited to this.
Embodiment:
A dynamic copies storage means for network file, is realized by following system, and this system comprises network disk, MySQL database server and distributed storage server cluster, and wherein network disk is the network file memory disk of sing on web; MySQL database server is the database server being provided with MySQL; Distributed storage server cluster is a distributed file storage system, it comprises a Controlling vertex (ControlNode) and multiple memory node (StorageNode), wherein, Controlling vertex is used for the file operation (create, heavily name, delete) of supervisory user storage directory, and safeguard user file copy list, file storage location list and clustered node list, memory node is the real deposit position of user file, is communicated by TCPSocket between Controlling vertex and memory node. Pass between this system three part is: the user profile table (users) that store network disk on MySQL database server, the Apache of distributing network disk and user's storage directory on the Controlling vertex of distributed storage server cluster, the network request of all users is all that maincenter processes taking Controlling vertex, the file of user is finally stored on memory node, and these storage means concrete steps are as follows:
1) on MySQL database server, user profile table is created, UserName, LoginCount and RepNum field is set, wherein UserName represents the user name of network disk, LoginCount represents that user's logs in number of times, RepNum represents the duplicate of the document number of user, and LoginCount can log in the increase of number of times along with user and constantly update;
2) Controlling vertex reads LoginCount field from user profile table, and is write and output in local LoginCount.txt file by the number of times that logs in of all users;
3) utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class;
4) class border and the class center of three classes is obtained respectively, i.e. minimum value, class center, maximum value, and left in an array, and then this array is sorted, thus by all class borders and class center according to depositing in array by little to big order, if A1 is the minimum value of class A, A2 is the class center of class A, A3 is the maximum value of class A, same B1, B2, B3 is the minimum value of class B respectively, class center and maximum value, C1, C2, C3 is the minimum value of class C respectively, class center and maximum value, then according to by little to numerical ordering in array being greatly: A1 < A2 < A3 < B1 < B2 < B3 < C1 < C2 < C3,
5) from user profile table, again read LoginCount field, determine the duplicate of the document number of all users according to the class border of A, B, C tri-classes:
If a) user logs in number of times and is less than or equal to A3, then RepNum=1;
If b) user logs in number of times and is more than or equal to B1 and is less than or equal to B3, then RepNum=2;
If c) user logs in number of times and is more than or equal to C1, then RepNum=3;
Again the value of the RepNum field of all users is upgraded;
6) for each user, using the value of UserName field as " key ", using the value of the RepNum field of its correspondence as " value ", deposit in user file copy list with the form of key value pair;
7) each memory node is as the server end of TCPSocket, opens monitoring, waits the connection of client terminal;
8) user's storage directory of network disk is carried out monitor in real time by Controlling vertex, once there be user to upload file, just proceeds to step 9) continue to perform;
9) obtaining the upper transmitting file path of this user, intercept user name and filename from this path, traversal user file copy list, until finding the position identical with this user name, reads the duplicate of the document number of its correspondence;
10) Controlling vertex is according to duplicate of the document number, Distribution Algorithm is utilized to obtain the memory node list (node number in this list with duplicate of the document number identical) corresponding with this filename from clustered node list, and using filename as " key ", using the memory node list that obtains as " value ", by this key value to joining in file storage location list;
11) Controlling vertex is as the client terminal of TCPSocket, sends Socket connection request with the pattern of multi-thread journey to each memory node in memory node list, and after setting up connection, transmits this file to the memory node of correspondence. Thus, the many copies completing this file store.
Above-mentioned steps 3) in utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class, it is as follows that it runs step:
<1>number by preparation classification is appointed as 3, and the path simultaneously arranging input file is above-mentioned steps 2) in the storage location of LoginCount.txt file;
<2>from input file, read each row data line by line, obtain all sampled datas and deposit in lists;
<3>from sampled data list, select the initial classes center of 3 sampled datas as 3 classes at random;
<4>travel through sampled data list, calculate the distance of each data to 3 Ge Lei centers, and these data are referred in the class nearest with 3 class width between centers;
<5>recalculate the mean value of each class, using this mean value as Xin Lei center, and upgrade the center of each class;
Whether the new and old center<6>comparing each class is identical, if identical, then algorithm terminates; Otherwise proceed to step<4>, continue circulation execution and go down.
Claims (2)
1. a dynamic copies storage means for network file, is realized by following system, and this system comprises network disk, MySQL database server and distributed storage server cluster, and wherein network disk is the network file memory disk of sing on web, MySQL database server is the database server being provided with MySQL, distributed storage server cluster is a distributed file storage system, it comprises a Controlling vertex and multiple memory node, wherein, Controlling vertex is used for the file operation of supervisory user storage directory, and safeguard user file copy list, file storage location list and clustered node list, memory node is the real deposit position of user file, communicated by TCPSocket between Controlling vertex and memory node, pass between this system three part is: the user profile table that store network disk on MySQL database server, the Apache of distributing network disk and user's storage directory on the Controlling vertex of distributed storage server cluster, the network request of all users is all that maincenter processes taking Controlling vertex, the file of user is finally stored on memory node, these storage means concrete steps are as follows:
1) on MySQL database server, user profile table is created, UserName, LoginCount and RepNum field is set, wherein UserName represents the user name of network disk, LoginCount represents that user's logs in number of times, RepNum represents the duplicate of the document number of user, and LoginCount can log in the increase of number of times along with user and constantly update;
2) Controlling vertex reads LoginCount field from user profile table, and is write and output in local LoginCount.txt file by the number of times that logs in of all users;
3) utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class;
4) class border and the class center of three classes is obtained respectively, i.e. minimum value, class center, maximum value, and left in an array, and then this array is sorted, thus by all class borders and class center according to depositing in array by little to big order, if A1 is the minimum value of class A, A2 is the class center of class A, A3 is the maximum value of class A, same B1, B2, B3 is the minimum value of class B respectively, class center and maximum value, C1, C2, C3 is the minimum value of class C respectively, class center and maximum value, then according to by little to numerical ordering in array being greatly: A1 < A2 < A3 < B1 < B2 < B3 < C1 < C2 < C3,
5) from user profile table, again read LoginCount field, determine the duplicate of the document number of all users according to the class border of A, B, C tri-classes:
If a) user logs in number of times and is less than or equal to A3, then RepNum=1;
If b) user logs in number of times and is more than or equal to B1 and is less than or equal to B3, then RepNum=2;
If c) user logs in number of times and is more than or equal to C1, then RepNum=3;
Again the value of the RepNum field of all users is upgraded;
6) for each user, using the value of UserName field as " key ", using the value of the RepNum field of its correspondence as " value ", deposit in user file copy list with the form of key value pair;
7) each memory node is as the server end of TCPSocket, opens monitoring, waits the connection of client terminal;
8) user's storage directory of network disk is carried out monitor in real time by Controlling vertex, once there be user to upload file, just proceeds to step 9) continue to perform;
9) obtaining the upper transmitting file path of this user, intercept user name and filename from this path, traversal user file copy list, until finding the position identical with this user name, reads the duplicate of the document number of its correspondence;
10) Controlling vertex is according to duplicate of the document number, Distribution Algorithm is utilized to obtain the memory node list corresponding with this filename from clustered node list, node number in this list is identical with duplicate of the document number, and using filename as " key ", using the memory node list that obtains as " value ", by this key value to joining in file storage location list;
11) Controlling vertex is as the client terminal of TCPSocket, send Socket connection request with the pattern of multi-thread journey to each memory node in memory node list, and after setting up connection, transmit this file to the memory node of correspondence, thus, the many copies completing this file store.
2. the dynamic copies storage means of a kind of network file as claimed in claim 1, it is characterised in that step 3) in utilize K-means algorithm, all users are logged in how many number of times is divided into A according to logging in, B and C tri-class, it is as follows that it runs step:
<1>number by preparation classification is appointed as 3, and the path simultaneously arranging input file is above-mentioned steps 2) in the storage location of LoginCount.txt file;
<2>from input file, read each row data line by line, obtain all sampled datas and deposit in lists;
<3>from sampled data list, select the initial classes center of 3 sampled datas as 3 classes at random;
<4>travel through sampled data list, calculate the distance of each data to 3 Ge Lei centers, and these data are referred in the class nearest with 3 class width between centers;
<5>recalculate the mean value of each class, using this mean value as Xin Lei center, and upgrade the center of each class;
Whether the new and old center<6>comparing each class is identical, if identical, then algorithm terminates; Otherwise proceed to step<4>, continue circulation execution and go down.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310282367.2A CN103327116B (en) | 2013-07-05 | 2013-07-05 | The dynamic copies storage means of a kind of network file |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310282367.2A CN103327116B (en) | 2013-07-05 | 2013-07-05 | The dynamic copies storage means of a kind of network file |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103327116A CN103327116A (en) | 2013-09-25 |
| CN103327116B true CN103327116B (en) | 2016-06-01 |
Family
ID=49195637
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310282367.2A Active CN103327116B (en) | 2013-07-05 | 2013-07-05 | The dynamic copies storage means of a kind of network file |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103327116B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104539741B (en) * | 2015-01-26 | 2019-10-15 | 北京奇艺世纪科技有限公司 | A kind of reminding method and device of Account Logon |
| CN106033434B (en) * | 2015-03-12 | 2019-06-14 | 中国人民解放军国防科学技术大学 | Fictitious assets data copy processing method based on data scale and popularity degree |
| CN104869361B (en) * | 2015-05-20 | 2018-06-05 | 浙江宇视科技有限公司 | A kind of Video Monitoring Terminal equipment in video monitoring system |
| CN106302625B (en) * | 2015-06-26 | 2019-10-25 | 阿里巴巴集团控股有限公司 | Data-updating method, device and related system |
| CN107105050B (en) * | 2017-05-11 | 2020-01-31 | 北京奇艺世纪科技有限公司 | Storage and downloading method and system for service objects |
| CN108900509B (en) * | 2018-06-29 | 2020-06-02 | 华中科技大学 | Copy selector based on programmable network equipment |
| CN110365783B (en) * | 2019-07-18 | 2022-10-21 | 深圳市网心科技有限公司 | File deployment method and device, network node and storage medium |
| CN116305220B (en) * | 2023-05-18 | 2023-08-08 | 天云融创数据科技（北京）有限公司 | Big data-based resource data processing method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102012933A (en) * | 2010-12-02 | 2011-04-13 | 清华大学 | Distributed file system and method for storing data and providing services by utilizing same |
| CN102999633A (en) * | 2012-12-18 | 2013-03-27 | 北京师范大学珠海分校 | Cloud cluster extraction method of network information |
| CN103150347A (en) * | 2013-02-07 | 2013-06-12 | 浙江大学 | Dynamic replica management method based on file heat |
-
2013
- 2013-07-05 CN CN201310282367.2A patent/CN103327116B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102012933A (en) * | 2010-12-02 | 2011-04-13 | 清华大学 | Distributed file system and method for storing data and providing services by utilizing same |
| CN102999633A (en) * | 2012-12-18 | 2013-03-27 | 北京师范大学珠海分校 | Cloud cluster extraction method of network information |
| CN103150347A (en) * | 2013-02-07 | 2013-06-12 | 浙江大学 | Dynamic replica management method based on file heat |
Non-Patent Citations (1)
| Title |
|---|
| Parallel K-PSO Based on MapReduce;Junjun Wang;《IEEE》;IEEE;20121111;第48卷;1203-1208 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103327116A (en) | 2013-09-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103327116B (en) | The dynamic copies storage means of a kind of network file | |
| CN107003906B (en) | Type-to-type analysis of cloud computing technology components | |
| US10601660B2 (en) | Auto discovery of configuration items | |
| CN102915331B (en) | The coordination engine selected for cloud | |
| US9081837B2 (en) | Scoped database connections | |
| CN106534291B (en) | Voltage monitoring method based on big data processing | |
| JP2018088293A (en) | Database system providing single tenant environment and a plurality of tenant environments | |
| CN106662986B (en) | The browser render process of optimization | |
| CN105681104B (en) | Method, system and the computer readable storage devices of network and machine are managed for online service | |
| WO2017204951A1 (en) | Optimizing read and write operations in object schema-based application programming interfaces (apis) | |
| US20070266029A1 (en) | Recovery segment identification in a computing infrastructure | |
| CA3025493A1 (en) | Optimizing read and write operations in object schema-based application programming interfaces (apis) | |
| CN103473696A (en) | Method and system for collecting, analyzing and distributing internet business information | |
| JP2017529633A (en) | Method and system for movably deploying an application to one or more cloud systems | |
| CN103605698A (en) | Cloud database system used for distributed heterogeneous data resource integration | |
| CN103430150A (en) | Technique for resource creation in a cloud computing system | |
| CN102404381A (en) | Software deployment system and deployment method based on workflow in cloud computing environment | |
| US10158709B1 (en) | Identifying data store requests for asynchronous processing | |
| CN102073683A (en) | Distributed real-time news information acquisition system | |
| CN103425511A (en) | System and method of installing and deploying application software in cloud computing environment | |
| CN103473636A (en) | System data components for collecting, analyzing and distributing internet business information | |
| CN103942281A (en) | Method and device for operating object persistently stored | |
| CN110661842A (en) | Resource scheduling management method, electronic equipment and storage medium | |
| CN106713391A (en) | Session information sharing method and sharing system | |
| CN110516076A (en) | A kind of the cloud computing management method and system of knowledge based map |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |