CN101901237B - Type-Index-Value distributed database built based on SQIite - Google Patents
Type-Index-Value distributed database built based on SQIite Download PDFInfo
- Publication number
- CN101901237B CN101901237B CN2009102165440A CN200910216544A CN101901237B CN 101901237 B CN101901237 B CN 101901237B CN 2009102165440 A CN2009102165440 A CN 2009102165440A CN 200910216544 A CN200910216544 A CN 200910216544A CN 101901237 B CN101901237 B CN 101901237B
- Authority
- CN
- China
- Prior art keywords
- index
- ddb
- request
- dls
- client
- 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
Images
Abstract
The invention discloses a Type-Index-Value manner distributed database built based on SQIite, which can be applied to an air traffic control (ATC) system. The invention comprises a distributed lock service (DLS) progress and a distributed database (DDB) server, wherein the DLS progress is used for providing distributed exclusive lock management and assignment for the whole system; the DDB server is used for utilizing the DLS and providing service for all progresses required by database update; the DLS progress is provided with a programming interface LibDLS base and the DDB server is provided with a programming interface LibDDB base; a programming interface LibDDB client is called in a client progress which utilizes a Sqlite database assignment table (type) to carry out data inserting, deleting and updating; a request IPC is sent to the DDB server by the programming interface LibDDB client; and after the lock is obtained, the data is submitted to the DDB by the programming interface LibDDB client. The DLS progress is based on R-A algorithm, and the index generated from the DLS is directly generated by an improved logical timer and local IP information.
Description
Technical field
The present invention relates to the distributed data base in the computer application field distributed system, can be used in air traffic control (the being called for short ATC) system with " type-index-content " (being called for short Type-Index-Value) form distributed earth preservation data.
In the large-scale distributed in modern times system, the database server of each main frame on fixed host computer of visit, but each database is installed on each main frame, database presents many copies state, distributed data base that Here it is simultaneously.The database of Type-Index-Value form is a kind of relevant database, and this type database exists data volume big, inserts, inquiry is frequent, but must satisfy the consistance of multimachine data.The ATC system requirements can move in 24 hours continuously, simultaneously stronger fault-tolerant ability can be arranged, and distributed data base can satisfy such requirement.But in practicality, there is the requirement that one type of data is stored, inquired about, and is not only general " key-Value " mode, promptly require to store with " Type-Index-Value " form.Based on the structure of " type-index-content " mode distributed data base of SQLite then be solve in the air traffic control (ATC) system with " type-index-content " (Type-Index-Value) the form distributed earth preserve a kind of approach of data demand, this invention has been applied in the distributed deployment of flight information data in system of air traffic control automation system.One Chinese patent application number 200810097631.4 " a kind of memory database system and application that realize distributed storage " more then are different fully though related to database storing in the document and utilized SQL request problem from its purpose, solution and the present invention.
The objective of the invention is solving with the ATC system is the realization of " type-index-content " mode distributed data base of representative.
The technical scheme that realizes the object of the invention is following:
A kind of " type-index-content " mode distributed data base that makes up based on SQlite is characterized in that comprising: the distributed lock service processes (DLS) that the management and the assignment of distributed mutual exclusion lock are provided for total system; Use the distributed data base service end (being called for short DDB) of DLS; Described distributed lock service processes (DLS) provides DLL LibDLS the storehouse, and described distributed data base service end (DDB) provides DLL LibDDB the storehouse; Utilizing sqlite database named list (type) to carry out the client process that operation was inserted, deletes, upgraded to data, need call DLL LibDDB client; Described DLL LibDDB client will ask IPC (being interprocess communication) to send to the DDB server, and after acquisition locks, submit the data to DDB again; After the DDB database server is received the request of client, send the request that locks to specified type designated recorder (being the synthetic character string of Type_index) to DLS lock server, and the asynchronous wait result that locks; Distributed lock service processes (DLS) returns to DDB with the index value Index that generates; DDB refreshes after this locality (promptly writing in the local Sqlite database) and these data of other main frames; Add the result to LibDDB asynchronous transmission data, index value Index returns to the process of using client as parameter.
The method of said structure " type-index-content " mode distributed data base is characterized in that: the request of locking can also can be used the process of client from other host nodes from client, and its request step that locks is following:
1. receive after the request of locking from client, stamp logical time for it and stab, with the request tabulation that joins request, and to other nodes transmission requests of locking;
2. receive after the request of locking from other nodes; Timestamp in the request of relatively locking and local logical time stab; Get maximum and add 1; Also collect whose of timestamp of timestamp (being the timestamp reckling) and this request of locking of the highest lock request of priority that neat other main frames reply when judging earlier from local client morning more; If should ask early; Then send it back and answer, otherwise, then directly add the REQUEST(request) formation;
3. receive after the unlocking request from client, by lock object and the timestamp in the message, corresponding item in the deletion request list, and check the highest request of next priority.
The method of described a kind of structure " type-index-content " mode distributed data base; It is characterized in that: the starting point that described logical time stamp adopts stab actual time and is used as a node logical timer is before a node begins to get into duty; Stab the actual time that obtains current this locality earlier; The back start main frame time stamp should greater than before the main frame time stamp, increase progressively 1 delicate timestamp when in the lock service incident taking place as increment value.
The method of described a kind of structure " type-index-content " mode distributed data base; It is characterized in that: the index value Index that described distributed lock service processes (DLS) generates directly generates by improving logical timer+local IP information; Its Index that generates out is different between each main frame; Because the time, even same main frame is exactly the phenomenon that shutdown occurred, Index also can not repeat in passage.
The method of described a kind of structure " type-index-content " mode distributed data base, it is characterized in that: DDB is responsible for the operation requests of process user process to the Sqlite database.Consumer process is if insert data, and the table name that only needs to operate Type and data send to DDB through the API of LibDDB, and DDB handles request and bears the application lock, obtains lock afterwards to the task of other DDB node refresh datas; Consumer process only need get final product with the asynchronous wait Refresh Data of LibDDB result and the index Index that obtains to return, and consumer process is if the modification data then need use the API of LibDDB to send " table name, index, data " to DDB, and DDB is responsible for operation.
The method of described a kind of structure " type-index-content " mode distributed data base; It is characterized in that: the task of described DDB refresh data is that version refreshes by two steps completion; Send version heartbeat and version and refresh merging; The treatment scheme that version refreshes is following: flow process begins, and receives the heartbeat message about version information; Relatively; Is judgement the record of deletion? If then do deletion and handle; If not, does is then judgement otherwise receive version (local version) once more? If, judge that once more oneself whether to have IP in this main frame that writes down highest version minimum, restart if not then getting back to the incipient stage; If then adding is updated tabulation, upgrades the old node of these versions after a period of stabilisation at version; Flow process finishes.
Beneficial effect of the present invention is embodied in the practical application in the ATC air traffic control system, solves and has realized in the distributed system carrying out data storage, the problem of retrieval by type.
Description of drawings
Fig. 1 is the structure structured flowchart synoptic diagram of DDB of the present invention.
Fig. 2 is the synoptic diagram of DLS of the present invention based on the R-A algorithm.
Fig. 3 is a locking and de-locking request step schematic flow sheet of the present invention.
Fig. 4 is the synchronous treatment scheme synoptic diagram of edition data of the present invention.
Fig. 5 is embodiment according to the invention and SQLite attended operation synoptic diagram.
Embodiment
Can find out that referring to Fig. 1 distributed data base of the present invention makes up distribution.Comprise distributed lock service routine DLS (Distributed Lock Service) that adopts the distributed mutual exclusion algorithms realization and the distributed data base service end DDB (Distributed DataBase) that uses DLS; Distributed lock service routine DLS provides DLL LibDLS the storehouse, and each need all need use this client application distributed lock to the DDB process that mutual exclusive resource is operated, and obtains the result that locks through the asynchronous call mode.DDB provides DLL LibDDB the storehouse, and each need submit the data to DDB by the Type-Index-Value mode to the process that database upgrades, and obtains operating result through the synchronization call mode.
The configuration file of DLS and DDB is effective to total system; The DLS of each main frame and DDB use the configuration file of unified content respectively; LibDLS and LibDDB can use the part of configuration file; They need know definition of T ype in the system, and promptly which the table that exists of database has, and DLS, the DDB information such as port numbers that are used to monitor as service end.
The present invention uses the embedded relational database Sqlite that increases income, and stores data by the mode of Type-Index-Value.Type is a table name, and list item has Index-Value-VersionNumber.Carry out the client process that operation was inserted, deletes, upgraded to data in sqlite database named list (type), should transfer the LibDDB client.
If newly-increased certain categorical data, and the request that will comprise Type-Value information submits to LibDDB, and LibDDB will ask IPC to send to the DDB server, and after obtaining to lock, submits the data to DDB again.After DDB data base consistency(-tance) server is received the request of client, send specifying the request that locks of Type to DLS lock server, and the asynchronous wait result that locks; DLS returns to DDB with the index value Index that generates; DDB refreshes after this locality (writing in the local Sqlite database) and these data of other main frames; Add the result to LibDDB asynchronous transmission data, index value Index returns to the process of using client as parameter.
If inquire about certain Type assigned I ndex data, then directly directly read local Sqlite database through using LibDDB by process.
When other host modifications, the data that increased, deleted certain Type newly, for this Type having been carried out the process of subscribing to, DDB can notify the improvement journey, tell its type that incident takes place comprise modifications, newly-increased, delete.
Described distributed lock service processes (DLS) is based on the DLS of R-A algorithm, moves its DLS process and can lock service be provided for whole distributed system, and each will apply for the process of locking, and all pass through the DLS of client end AP I default access this locality.The node that the R-A algorithm requires to get into the critical section must send request to every other node, and must wait for that the message that all nodes are replied just has the qualification that gets into the critical section afterwards.In DLS, we claim that this state is for obtaining lock.Fig. 2 is a R-A algorithm synoptic diagram, can find out S1, S2; Three same critical resources of main frame application of S3 conduct interviews, and Time To Event from left to right can know that S1 at first sends the request of being with timestamp to S2 and S3; S2 and S3 receive after the request; S3 has also proposed the request to this resource to S1 and S2, and S2 just replys S1 immediately because intranodal own does not have other to the request more early of this critical resource.Though and S3 has also proposed request, the time of request is later than the request time of S1, just replys S1.So the S1 that is all replied gets into the critical section.For the request of S3, S2 self is request not, replys immediately, and S1 has got into the critical section, wouldn't reply, and then S3 is in the stage of waiting for that all answers collect, does not promptly obtain lock.When S1 finishes the visit critical resource, just can reply S3, S3 has also collected all answers so, and S3 gets into the critical section.S1 finishes the visit critical resource, and the process of replying S3 is in DLS, and we are referred to as to discharge lock.
For the sequencing of the time of following, generally adopt logical timer to be used as the generation event time reference of total system in the distributed system.Logical timer can adopt simple register algorithm, but can have some defectives, such as the DLS that starts at certain midway, with other DLS through before the heartbeat unified clock, logical timer during this period of time all is in inaccurate state.Logical time that the request of receiving is during this period of time stamped stab can't and system synchronization; Because the lock of lock service distributes the R-A algorithm that adopts; The priority that obtains locking that timestamp is more little is high more, and these requests can make the order that locks of system chaotic so, thereby cause Refresh Data chaotic.The starting point that described logical time stamp adopts stab actual time and is used as a node logical timer is before a node begins to get into duty; Stab the actual time that obtains current this locality earlier; The back start main frame time stamp should greater than before the main frame time stamp, increase progressively 1 delicate timestamp when in the lock service incident taking place as increment value.When taking place in the lock service, following incident increases progressively 1 delicate timestamp:
1. receive the lock request from client, self timestamp increases progressively
2. receive lock request, compare with self timestamp and get maximal value and increase progressively from other nodes
3. receive response request, compare with self timestamp and get maximal value and increase progressively from other nodes
4. receive heartbeat message, compare with self timestamp and get maximal value and increase progressively from other nodes
5. send the heartbeat message of oneself, self timestamp increases progressively
When 6. testing other node break-makes, self timestamp increases progressively
In addition, in order further to improve reliability, nodal point separation can be set in configuration file start to the time interval that gets into steady state (SS).Before getting into steady state (SS), do not handle request from client.
The index value Index that described distributed lock service processes (DLS) generates directly generates by improving logical timer+local IP information; Its Index that generates out is different between each main frame; Because the time, even same main frame is exactly the phenomenon that has shutdown, Index also can not repeat in passage.Therefore the generation of Index is very crucial.For " Type-Index-Value " database; If Index is according to generate obviously existing problems since 1 order that increases progressively; Like two different main frames in some data that do not had under the network connection situation each self-generating; Index increases progressively from 1, when these two main frames connect through network, just has the problem that data merge under the distributed situation so; And if two main frames are connected in online two main frames, and each increases progressively and begins to generate Index and also can produce ambiguity since 1.We notice that Index just plays a part to distinguish data in fact, and whether unimportant continuously, the method that provides here is; Index is generated by DLS, directly uses above-mentioned improvement logical timer+local IP information to generate, and concrete grammar is: when obtaining lock; DLS can therefore generating a new logical timer gives DDB; New logical timer=MAX (logical timer, the logical timer in last REPLY message)+1, this logical timer send DDB by DLS; Its microsecond partly is 99999 to the maximum; In 4 bytes of microsecond, only need take three bytes, last byte that last byte is got the IP address get final product, and last byte of IP address promptly as " 192.168.18.16 " and last byte of IP address be 16 (0x0A); Still there is different local area for the 3rd byte in IP address (promptly going up 18 in the example); Can 4 byte per second information among the Index (since on January 1st, 1,980 zero time) partial fixing be deducted a bigger value, make second information Control in 3 bytes, with the 3rd byte of IP address portion the 4th byte as information second; Generating the Index that comes out like this is different between each main frame; Because the time, even same main frame is exactly the phenomenon that shutdown occurred, Index also can not repeat in passage.
The task of described DDB refresh data is that version refreshes by two steps completion, sends version heartbeat and version and refreshes merging, and the treatment scheme that version refreshes is following: flow process begins, and receives the heartbeat message about version information; Relatively; Is judgement the record of deletion? If then do deletion and handle; If not, does is then judgement otherwise receive version (local version) once more? If, judge that once more oneself whether to have IP in this main frame that writes down highest version minimum, restart if not then getting back to the incipient stage; If then adding is updated tabulation, upgrades the old node of these versions after a period of stabilisation at version; Flow process finishes.
Though the assurance of distributed mutual exclusion lock is arranged; But the situation that restart after can existing new adding main frame or main frame to withdraw from the system, cutting apart appears in network; The data of these main frame disappearances need be updated, and the data of different each self-generating of main frame need to merge, and DDB has born the task of this Refresh Data.The version of DDB refreshes by two steps to be accomplished, and sends version heartbeat and version and refreshes merging.
Version number described in the present invention (Version) and Index-Value count the Sqlite database as a record together; Version number is the logical time that obtains when using DLS to obtain lock and stabs; According to the R-A algorithm, this timestamp can reflect that the variation of data in time lapse promptly can write down the time that this version is updated.
The content of version heartbeat comprises sends this node certain shows the Index information of all and the timestamp of Version information and the last this Type Delete All operation (the logical time stamp that obtains from DLS when locking when carrying out full deletion action); If what carry out is the deletion of certain bar record, the version of this record is changed to 0.Other nodes receive after the version heartbeat, and this that contrasts own database shown the version number and the timestamp of corresponding list item.Following several kinds of situations are carried out handled.
Below promptly do not deleted for 0(for version number) processing:
1. to same Index, version number in the heartbeat of receiving＜local node version number
The version of assigned I ndex that this heartbeat main frame is described is old excessively, and this main frame of DDB record adds legacy version Host List, exists if when record, find to exist than own high version, and then deletion should tabulation; When record find to exist another with own the same all be the main frame of highest version, through the IP address of this machine and this machine relatively, the little main frame in IP address is deleted and is changed tabulation; After 15 seconds (configurable) after receiving heartbeat message, the application lock carries out version updating to the old the type All hosts of these versions then.
2. receive version number in the heartbeat＞this node version number
Ignore, wait for the renewal of highest version main frame
3. receive version number in the heartbeat=this node version number
Ignore, do not handle
Below for the processing of full deletion time stamp occurring:
Full timestamp＜local node index the version number of deleting in the heartbeat of 1. receiving
The record that keeps this index
Full timestamp＞=local node index the version number of deleting in the heartbeat of 2. receiving
Delete this index record
Below for version number be not the coverlet record deletion for 0() processing
The Index=local node index that deletes in the heartbeat of 1. receiving
The version of local this Index is changed to 0, and broadcasts heartbeat message immediately.
The initiation main frame of single deletion is collected the information of other these record deletions of main frame, and main frame deletes all that (have two kinds of situation, the 1.vetsion field is changed to 0 in the discovery system; 2. in the time of the information of not having this index), be the order that 0 main frame sends this recorded information of deletion to version, each main frame is received thoroughly this record of deletion of back, when finding that other main frames do not have this index record again, and local this record of initiation machine deletion.
Fig. 5 is embodiment according to the invention and database SQL ite interface attended operation synoptic diagram.
Sqlite is an embedded relational database of increasing income, and it can reduce the expense of application program management data, and the SQLite portability is good, is easy to use, and is very little, efficient and reliable.SQLite is embedded in the application program of using it, their shared identical process spaces, rather than an independent process.Inner in process, it is complete, self-contained database engine; One big benefit of embedded database is exactly not need network configuration in your program inside, does not also need management.Because client and server moves in the same process space.The data base authority of SQLite only depends on file system, does not have the notion of user account.SQLite has the database level locking, does not have the webserver.It uses expense very little, is suitable for embedded device.
Preserve this locality that main use Sqlite does data in the database among the present invention, is used for the direct read of client.
Among the present invention the Sqlite database being implemented write operation is mainly accomplished by DDB; And read operation mainly is to be accomplished by the client process of using libDDB; Shown this relation among Fig. 5, client process has only been drawn one among this figure, but all client process are all by this completion.Because the lock service is not used and is not drawn the DLS process among this figure of Sqlite.
Each said DDB node is that each type creates a table in Sqlite, and the field of each table comprises: [dataIndex1, dataIndex2; Version1, version2, dataValue; DataBinary], dataIndex1 and dataIndex2 are two ingredients of index, the version1 in the field; Version2 is data when being modified, and increases the corresponding logical time of version number and stabs.
During operation sqlilte database, should open database, and when finishing closing database.
Claims (7)
1. a method that makes up " type-index-content " mode distributed data base based on Sqlite is characterized in that comprising: the distributed lock service processes DLS that distributed mutual exclusion lock management is provided for whole distributed data base; Use the distributed data base service end DDB of DLS; Described distributed lock service processes DLS provides DLL LibDLS the storehouse; Described distributed data base service end DDB provides DLL LibDDB the storehouse; Carrying out data insertion, deletion, upgrading the client process of operating, need call DLL LibDDB client; Described DLL LibDDB client will send to the DDB server through interprocess communication IPC to the operation requests of database, and after acquisition locks, submit the data to DDB again; After the DDB database server was received the operation requests to database of client, sending the specified type designated recorder to DLS lock server was the request that locks of the synthetic character string of Type_index, the asynchronous then wait result that locks; After obtaining lock, carry out in the distributed refresh process; The DDB that upgrade to initiate node is refreshing after this locality promptly writes local these data of other main frames of Sqlite database neutralization; Add the result to LibDDB asynchronous transmission data, index value Index returns to the process of using client as parameter.
2. the method for structure according to claim 1 " type-index-content " mode distributed data base; It is characterized in that: described distributed lock service processes DLS is based on the DLS of R-A algorithm; Move it and can the lock service be provided for whole distributed data base; Each will apply for the process of locking, and all passes through the local DLS of client end AP I default access.
3. the method for structure according to claim 1 " type-index-content " mode distributed data base, it is characterized in that: described DDB is responsible for the operation requests of process user process to the Sqlite database; Consumer process is if insert data, and the table name that only needs to operate Type and data send to DDB through the API of LibDDB, and DDB handles request and bears the application lock, obtains lock afterwards to the task of other DDB node refresh datas; Consumer process only need get final product with the asynchronous wait Refresh Data of LibDDB result and the index Index that obtains to return.
4. the method for structure according to claim 1 " type-index-content " mode distributed data base; It is characterized in that: the said request of locking can be from client; Also can use the process of client from other host nodes, its locking and de-locking request step is following:
1. receive after the request of locking from client, stamp logical time for it and stab, with the request tabulation that joins request, and to other nodes transmission requests of locking;
2. receive after the request of locking from other nodes; Timestamp in the request of relatively locking and local logical time stab; Get maximum and add 1; The timestamp of also collecting the highest request of locking of priority that neat other main frames reply when judging earlier from local client be timestamp reckling and this request of locking timestamp who more early; If should ask early; Then send it back and answer, otherwise the REQUEST formation then directly joins request;
3. receive after the unlocking request from client, by lock object and the timestamp in the message, corresponding item in the deletion request list, and check the highest request of next priority.
5. the method for structure according to claim 4 " type-index-content " mode distributed data base; It is characterized in that: the starting point that described logical time stamp adopts stab actual time and is used as a node logical timer is before a node begins to get into duty; Stab the actual time that obtains current this locality earlier; The back start main frame time stamp should greater than before the main frame time stamp, increase progressively 1 delicate timestamp when in the lock service incident taking place as increment value.
6. the method for structure according to claim 1 " type-index-content " mode distributed data base; It is characterized in that: the index value Index that described distributed lock service processes DLS generates directly generates by improving logical timer+local IP information; Its Index that generates out is different between each main frame; Because the time, even the phenomenon of shutdown has appearred in same main frame, Index also can not repeat in passage.
7. the method for structure according to claim 3 " type-index-content " mode distributed data base; It is characterized in that: the task of described DDB refresh data is that version refreshes by two steps completion; Send version heartbeat and version and refresh merging; The treatment scheme that version refreshes is following: flow process begins, and receives the heartbeat message about version information; Relatively; Judgement is the record of deletion, if then do deletion and handle; If not, then judge once more to be otherwise to receive version, if, judge that once more oneself whether to have IP in this main frame that writes down highest version minimum, restart if not then getting back to the incipient stage; If then adding is updated tabulation, upgrades the old node of these versions after a period of stabilisation at version; Flow process finishes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102165440A CN101901237B (en) | 2009-12-04 | 2009-12-04 | Type-Index-Value distributed database built based on SQIite |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102165440A CN101901237B (en) | 2009-12-04 | 2009-12-04 | Type-Index-Value distributed database built based on SQIite |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101901237A CN101901237A (en) | 2010-12-01 |
| CN101901237B true CN101901237B (en) | 2012-05-23 |
Family
ID=43226776
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009102165440A Active CN101901237B (en) | 2009-12-04 | 2009-12-04 | Type-Index-Value distributed database built based on SQIite |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101901237B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102654878B (en) * | 2011-03-04 | 2014-03-05 | 迈普通信技术股份有限公司 | Data operation method for distributed database in embedded system and board card |
| CN102591979A (en) * | 2012-01-05 | 2012-07-18 | 厦门市美亚柏科信息股份有限公司 | Method and system for mining deleted records of free space of sqlite |
| CN105224361B (en) * | 2015-09-15 | 2018-05-25 | 华讯方舟科技有限公司 | A kind of method and system upgraded to sqlite3 type embedded databases |
| CN106708608B (en) * | 2015-11-16 | 2020-08-11 | 阿里巴巴集团控股有限公司 | Distributed lock service method, acquisition method and corresponding device |
| CN107146075B (en) * | 2016-03-01 | 2020-11-10 | 创新先进技术有限公司 | Request processing method and device |
| US10891291B2 (en) * | 2016-10-31 | 2021-01-12 | Oracle International Corporation | Facilitating operations on pluggable databases using separate logical timestamp services |
| CN106874341B (en) * | 2016-12-23 | 2022-04-05 | 中科星图股份有限公司 | Database synchronization method |
| CN106709005B (en) * | 2016-12-23 | 2020-11-24 | 北京奇虎科技有限公司 | Method, device and system for processing redundant index in database system |
| CN109033359A (en) * | 2018-07-26 | 2018-12-18 | 北京天地和兴科技有限公司 | A kind of method of multi-process secure access sqlite |
| CN109542867B (en) * | 2018-11-26 | 2020-07-24 | 成都四方伟业软件股份有限公司 | Distributed data acquisition method and device |
| CN110334823B (en) * | 2019-06-17 | 2022-04-05 | 北京大米科技有限公司 | Reservation method, reservation device, electronic equipment and medium |
| CN110865977A (en) * | 2019-11-05 | 2020-03-06 | 中盈优创资讯科技有限公司 | Method and device for operating HDFS directory by multiple programs |
| CN111125129A (en) * | 2019-12-10 | 2020-05-08 | 珠海格力电器股份有限公司 | Data processing method and device, storage medium and processor |
| CN110968646B (en) * | 2019-12-20 | 2023-06-06 | 广东睿住智能科技有限公司 | Embedded system database synchronization method, device and storage medium |
| CN111159106B (en) * | 2019-12-30 | 2023-04-07 | 亚信科技（中国）有限公司 | Data query method and device |
| CN113407635B (en) * | 2021-07-09 | 2022-06-03 | 明度智云(浙江)科技有限公司 | Multi-terminal data synchronization method and system for intelligent manufacturing |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6336117B1 (en) * | 1999-04-30 | 2002-01-01 | International Business Machines Corporation | Content-indexing search system and method providing search results consistent with content filtering and blocking policies implemented in a blocking engine |
| CN101266622A (en) * | 2008-05-07 | 2008-09-17 | 北京邮电大学 | Trans-platform mobile storage medium data system |
-
2009
- 2009-12-04 CN CN2009102165440A patent/CN101901237B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6336117B1 (en) * | 1999-04-30 | 2002-01-01 | International Business Machines Corporation | Content-indexing search system and method providing search results consistent with content filtering and blocking policies implemented in a blocking engine |
| CN101266622A (en) * | 2008-05-07 | 2008-09-17 | 北京邮电大学 | Trans-platform mobile storage medium data system |
Non-Patent Citations (1)
| Title |
|---|
| JP特表2004-505380AT 2004.02.19 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101901237A (en) | 2010-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101901237B (en) | Type-Index-Value distributed database built based on SQIite | |
| KR102307371B1 (en) | Data replication and data failover within the database system | |
| US8224860B2 (en) | Database management system | |
| US9779128B2 (en) | System and method for massively parallel processing database | |
| US10747745B2 (en) | Transaction execution commitment without updating of data row transaction status | |
| US10430298B2 (en) | Versatile in-memory database recovery using logical log records | |
| CN110196885B (en) | Cloud distributed real-time database system | |
| US11442961B2 (en) | Active transaction list synchronization method and apparatus | |
| US20190108166A1 (en) | Replicable differential store data structure | |
| EP2304609B1 (en) | Paging hierarchical data | |
| US20130110767A1 (en) | Online Transaction Processing | |
| CN105183400B (en) | It is a kind of based on content addressed object storage method and system | |
| US20090089313A1 (en) | Decentralized record expiry | |
| CN108491332A (en) | A kind of real-time buffering updating method and system based on Redis | |
| CN102724304A (en) | Information warehouse federation in subscription/release system and data synchronization method | |
| CN104113587A (en) | Client metadata buffer optimization method of distributed file system | |
| CN106021468A (en) | Updating method and system for distributed caches and local caches | |
| CN101232514A (en) | Metadata synchronization method of network additional memory node and network additional memory node | |
| CN108600284A (en) | A kind of virtual machine high availability implementation method and system based on Ceph | |
| US9201685B2 (en) | Transactional cache versioning and storage in a distributed data grid | |
| CN114254016A (en) | Data synchronization method, device and equipment based on elastic search and storage medium | |
| CN106210038B (en) | The processing method and system of data operation request | |
| CN109947592A (en) | A kind of method of data synchronization, device and relevant device | |
| CN113742336A (en) | Data processing method and device and storage medium | |
| Lin et al. | A low-cost checkpointing technique for distributed databases |
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 |