CN102760161A - Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof - Google Patents
Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof Download PDFInfo
- Publication number
- CN102760161A CN102760161A CN2012101915668A CN201210191566A CN102760161A CN 102760161 A CN102760161 A CN 102760161A CN 2012101915668 A CN2012101915668 A CN 2012101915668A CN 201210191566 A CN201210191566 A CN 201210191566A CN 102760161 A CN102760161 A CN 102760161A
- Authority
- CN
- China
- Prior art keywords
- page
- log
- daily record
- master chip
- transaction
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a log organization structure clustered according to transaction aggregation and a recovery protocol based on the log organization structure clustered, which can be applied to a transactional data management system of a large-sized computer. A log file is sequentially organized to a plurality of log fragments and each log fragment is used for storing the log content of the same transaction and reserving a transaction number as well as a preceding log fragment pointer of the transaction; a data page number involved in a log entry of the same fragment is stored in the form of an array. When the system is operating, each transaction only writes its own log fragment and writes the log fragment in the log file when the transaction is submitted. In a recovering process, the system can be recovered to a lasting and consistent state by scanning all the log fragments for remake and returning the log fragments of all the active transactions for return. The problem of producing bottlenecks during writing logs in the traditional transactional data management system is resolved and log amount of the system can be effectively reduced.
Description
Technical field
The present invention's design belongs to the processing data information field, and particularly the usage log technology guarantees the transactional data management system of data consistency.
Background technology
In the Feng Shi Computer Architecture; Permanent data is stored in the external equipment slow in speed, that volume is bigger (like hard disk); And when data were visited, it was read in the less memory device of rapid speed, volume (like RAM), and visited by central processing unit.Yet because performance demands, data are only upgraded its memory mapping when upgrading, and the renewal of external memory is undertaken by the time-delay of buffer management unit.
In the transactional data management system; Single affairs---are promptly upgraded the single job process of being formed by some---, and be regarded as a groundwork unit, it has two important properties, and one is consistance; Promptly all must be simultaneously in sight to updating data, perhaps can't see fully; The 2nd, persistence, promptly when affairs were submitted to, all must forever be saved in the system updating data.Under the situation that a large amount of concurrent transactions are worked simultaneously, consistance and persistence are the essential conditions that guarantees the total system true(-)running.But because the external memory of data reflection always postpones the latest edition in internal memory; Therefore when system breaks down; This moment, the versions of data of external memory possibly lost the renewal (destruction persistence) of having submitted affairs to, also possibly preserve the renewal (destruction consistance) of not submitting affairs to.
All transactional data management systems all adopt based on the recovery agreement of daily record and come recovery system at present, to realize consistance and persistence.This agreement can briefly be described below: daily record is a flow structure file on the external memory, and system safeguards a log buffer district in internal memory, wherein deposits the daily record of nearest generation.When affairs A preparation was carried out a renewal U to data page P, system produced daily record (S, an A in the log buffer district; P, U, Cp; Cs, Sprev), wherein S is the unique logged sequence numbering (monotone increasing in time) that aims at this day in the daily record stream; Cp is the data updated preceding paragraph, and Cs is that data updated is consequent, and Sprev is the last logged sequence numbering of upgrading daily record of affairs A.Affairs write on S on the P before renewal.And following principle must be followed in data buffer administrative unit and transaction management unit: 1) when page of data is write back external memory by the data buffer administrative unit, the daily record before the unique logged sequence numbering on its page must be by first materialization in the external memory journal file.2) when affairs are submitted to, all daily records in the log buffer district must be by materialization in the external memory journal file at this moment.
When total system breaks down, at first scan the journal file of external memory materialization, every daily record S corresponding page P is read into internal memory; And the relation between the logged sequence on inspection page numbering S ' and the S, if S ' less than S, then the external memory version of instruction page is before this daily record is upgraded; This daily record this moment need be reformed, and page sequence number is updated to S, if S ' is greater than S; The renewal materialization of this daily record correspondence then has been described, need have not been reformed.Because the necessary all daily records of materialization when affairs are submitted to, so the persistence of affairs is guaranteed.After all daily records all scannings and are reformed as required; Also made up simultaneously the transaction table that enlivens when breaking down; Because the renewal daily record of same affairs is linked by the Sprev territory through daily record, therefore can all renewal daily records of rollback, make database recovery to consistent state.
Above-mentioned log-structured with recover agreement and exist certain problem in the computer system in modern times.At first; The volume of wall scroll daily record is bigger, since every necessary log sequence numbering S of daily record, transaction number A; Transaction journal preceding paragraph Sprev; The length of adding up in these territories often surpasses 20 bytes, and if the data of a renewal operation itself more in short-term, shared space, these territories often approaching whole journal file space 1/3 even 1/2; Secondly, owing to produce the unique incremental of the necessary global synchronization of the action of daily record with the assurance logged sequence, and current mainframe computer often has dozens or even hundreds of central processing unit, when upgrading intensive generation, can cause the bottleneck of ductility.
Summary of the invention
The object of the invention is to provide a kind of daily record institutional framework by transaction aggregation, and the implementation method of base recovery agreement on it, reduces the journal file volume to reach, the target of the performance of raising system under the concurrent condition of height.
For reaching above-mentioned target, the present invention realizes through following technical scheme:
A kind of daily record institutional framework of pressing transaction aggregation.Journal file is several days master chips by sequential organization, each day the length of master chip can be different, but all be the integral multiple of external equipment fundamental block element length.
Day master chip in the said journal file institutional framework is deposited the log content of same affairs, and its basic structure is: (S, A, Pn, Ln, Pa, La, Sprev).Wherein, S represents the unique logged sequence numbering of Ben Ri master chip in journal file, and A represents the relevant issues number of Ben Ri master chip, and Pn represents the related page of data number of Ben Ri master chip, the daily record number that on behalf of the Ben Ri master chip, Ln comprised, and they all are the territories of fixed length.Face Pa is the page number array of all pages of relating to of the daily record in this day master chip, and on behalf of an independent page of data, the item of total Pn the fixed length of array number for every.La is all journal entry arrays, and total Ln the elongated item of array represented an independent daily record for every.Sprev be these affairs last one day master chip sequence number.
Page number array Pa in the said day master chip sorts according to page number, so that locate visit fast through dichotomy.
Among the journal entry array La in the said day master chip institutional framework of wall scroll daily record be (seq, i, U, Cp, Cs).Wherein, seq is the page sequence number on the related data page, and this sequence number increases progressively when each page of data takes place to upgrade.I is in that [0, the numeral between Pn) is represented the subscript of the page number array Pa on day master chip, and Pa [i] is the page of data that this daily record takes place.U, the related meanings of Cp and Cs is consistent with the definition in the background technology.
A kind of realization comprises based on the method for the recovery agreement of above-mentioned journal file structure:
Each affairs is safeguarded a privately owned log buffer district, and its structure is consistent with the day master chip of journal file, and the visit in privately owned log buffer district is not needed synchronously.
When affairs A upgrades a page of data, in the journal entry array La in privately owned log buffer district, insert a new daily record, and possibly upgrade page number array Pa.On each page of data, safeguard a monotonically increasing page sequence number seq, seq is recorded in the daily record during renewal, upgrade back seq and add 1, and transaction number A is recorded on the page.
When affairs were submitted to, its privately owned log buffer district was write back journal file, and obtained unique logged sequence numbering S.
When the data buffer administrative unit writes back data file with certain page of data; If enliven transaction number on the transaction number that writes down on the page; Then the privately owned log buffer district of these affairs is write back journal file, and obtains unique logged sequence numbering S, and this privately owned log buffer district is reset then.
When after system makes mistakes, recovering, at first reform.The process of reforming is: according to the order of sequence scanning day master chip and in all journal entry arrays; And compare according to the seq on the seq ' in the journal entry content and the related pages; If seq ' equals seq; Then reform this daily record and the seq on the page added 1, if seq ' equals seq, this daily record of then not reforming.
After the completion of reforming, all are enlivened affairs carry out rollback.Back off procedure is: enliven affairs for each, the day master chip all according to Sprev territory backward scanning to all journal entry arrays of each day master chip backward scanning, and carries out rollback operation to each the bar daily record in the journal entry content array.
The invention solves in the transactional data management system on the mainframe computer server in modern times, the performance bottleneck problem that is run into when utilizing the daily record technology to come restore data has following beneficial effect:
1) each affairs is when parallel the renewal, need not take place when writing daily record anyly to have improved the operational efficiency of affairs self effectively synchronously, and greatly remove in contingent wait relation between affairs under the concurrent environment of height, improved the throughput of system.
The preceding paragraph daily record pointer Sprev that must preserve in 2) will original every daily record is reduced to each day one of master chip; It is shared that the page number that original every daily record must be preserved is reduced to the daily record in the master chip in same day; Above measure has effectively reduced the volume of daily record, has reduced the carrying cost of system.
Description of drawings
Fig. 1 is a journal file one-piece construction exemplary plot.
Fig. 2 is a daily record chip architecture synoptic diagram.
Fig. 3 is a transaction write daily record schematic flow sheet.
Fig. 4 is the system failure recovery schematic flow sheet.
Embodiment
Below in conjunction with accompanying drawing the present invention is described in further detail:
See also Fig. 1, this figure is the global tissue topology example figure of journal file, journal file be divided into differ in size one by one the day master chip, each day the length of master chip be the integral multiple of external equipment elementary cell.Each day, master chip was that an office monopolizes, and the Sprev territory of passing through on day master chip between the different day master chips of same affairs links to each other.
See also Fig. 2, this figure is single day master chip institutional framework synoptic diagram, on each daily record head some fixed-length fields is arranged, and they comprise:
1) this day master chip unique sequence numbering S, can uniquely navigate to this day master chip according to this numbering.Be noted that in some transactional data management system and locate log record with absolute displacement, S can not preserve in this case;
2) the respective transaction A of this daily record is used for when recovering, making up the active transaction table;
3) the page of data number P n that relates on day master chip;
4) the daily record number Ln that deposits on day master chip.
In fixed-length field back, followed by store data page number array Pa, each deposits a page of data number array, and the length of array is Pn.In all data management systems, the length of page of data number all is certain, therefore can calculate the length of Pa according to Pn.
Deposit journal entry array La at the back in the page of data array,, here can employing deposit in proper order because the length of journal entry is unfixing.The content of every daily record all comprises following territory:
When 1) this daily record produces, the page sequence number seq on the related data page, fixed length;
2) page of data array index i points to the corresponding entry of page of data array in this day master chip, fixed length;
3) Log Types U, fixed length;
4) daily record data updated preceding paragraph Cp, elongated, can and resolve according to the U assembling;
5) the consequent Cs of daily record data updated, elongated, can and resolve according to the U assembling.
Each affairs is opened up the privately owned log buffer district of oneself in internal memory, when Updating Information the page, in buffering, write log information.The Update Information method flow of the page of the affairs based on above-mentioned daily record institutional framework that propose for the present invention shown in Figure 3 may further comprise the steps:
301 upgrade when beginning, and affairs at first obtain the exclusive access rights of page of data, promptly do not allow other this pages of transactions access this moment, judge to upgrade whether can carry out according to the type of upgrading then.
302 prepare to write daily record, the current page sequence number seq on the read data pages, the preceding paragraph Cp of data, the consequent Cs of data, and the type U of this renewal.
303 judge among the page number array Pa of this page number in the privately owned log buffer of these affairs district whether exist, if do not exist, then call 304, otherwise directly get into 305.Because the page number array is orderly and fixed length, therefore uses binary chop can judge rapidly whether this page number exists.
304 insert the page number P of this page in page number array Pa, notice that this insertion should still keep the order of Pa.This operation simultaneously can cause La after move.
305 confirm the subscript i of page number in the page number array, construct complete journal entry.And be stored into the end of La.
306 pages that Update Information are recorded in this transaction number on the page, and the seq of the page are added 1.
When system normally moves, also to follow following two rules, at first, when a page of data is excluded out buffer zone, need write back journal file to the still privately owned buffer zone of active affairs in the transaction number that writes down on the page; Secondly, when affairs are submitted to, oneself privately owned buffer zone is write back journal file.These two processes all need synchronously, whenever write a daily record in the legacy protocol and all need synchronously but compare, and performance is greatly improved.
When system breaks down, abide by the flow process of Fig. 4 signal and recover, recover beginning, in internal memory, set up the transaction table that enlivens of a sky, carry out following steps then:
Whether 401 cycle criterions are current also has a day master chip not scanned, if having, then gets into 402, if do not have, then gets into 403;
402 scanning next day master chip S judge at first whether the transaction number A on day master chip exists in enlivening transaction table, if do not exist, then A is added and enliven transaction table, and last day master chip of putting A is S; If last day master chip that exists then upgrade A is S;
If whether the journal entry of institute do not have, then gets into 404 by whole scannings on the current day master chip of 403 cycle criterions, otherwise returns 401 judgements.
404 scan next journal entry L, and obtain the corresponding data page access power in the data file according to the page number P that journal entry is associated
Seq in the 405 audit log items and the relation between the seq on the page, if the former equals the latter, then instruction page is in the state before upgrading, get into 406 this moment, otherwise return 404, visits next bar daily record.
406 this daily records of reforming are updated on the page the data of daily record are consequent, and the seq of the page are added 1.
407 accomplish to this explanation process of reforming, and this moment, inspection enlivened transaction table, and circulation is with all transaction rollback of enlivening.
Whether the current affairs of 408 circular tests rollback are accomplished, and the condition of inspection is whether last day master chip of affairs is empty, if not then enter into 409.
Current day master chip of 409 rollbacks pressed the backward rollback with all journal entries, and last day master chip of affairs is changed to the Sprev of this day master chip.
Claims (11)
1. daily record institutional framework based on transaction aggregation, wherein:
Journal file is several days master chips by sequential organization, each day the length of master chip can be different, but all be the integral multiple of external equipment fundamental block element length.
2. require described daily record institutional framework according to right 1, it is characterized in that, the day master chip in the said journal file institutional framework is deposited the log content of same affairs, and it substance that comprises comprises at least:
1) day master chip sequence numbering,
2) transaction number
3) page number array length
4) journal entry array length,
5) with the previous day master chip sequence numbering of affairs,
6) page number array,
7) journal entry array.
3. daily record institutional framework according to claim 1 and 2 is characterized in that, the page number array Pa in the said day master chip sorts according to page number, so that locate visit fast through dichotomy.
4. daily record institutional framework according to claim 1 and 2 is characterized in that, the institutional framework of the wall scroll daily record in the journal entry array in the said day master chip comprises at least:
1) page sequence number on the page of data,
2) page number is in the subscript of page number array,
3) updating type,
4) preceding paragraph that Updates Information,
5) Update Information consequent.
5. a realization is based on the method for the recovery agreement of the described daily record institutional framework of claim 1, and wherein the transactional system follows protocol requirement when normal operation, and recovers according to agreement in the back of makeing mistakes.
6. method according to claim 5 is characterized in that, each affairs is safeguarded a privately owned log buffer district, and its structure is consistent with the day master chip of journal file, and the visit in privately owned log buffer district is not needed synchronously.
7. according to claim 5 or 6 described methods, it is characterized in that, when affairs are upgraded a page of data, in the journal entry array in privately owned log buffer district, insert a new daily record, and possibly upgrade the page number array.On each page of data, safeguard a monotonically increasing page sequence number, upgrade the time series quilt and record in the daily record, upgrade the back sequence number and add 1, and transaction number is recorded on the page.
8. according to claim 5 or 6 described methods, it is characterized in that when affairs were submitted to, its privately owned log buffer district was write back journal file, and obtain unique logged sequence numbering.
9. according to claim 5 or 6 described methods; It is characterized in that; When the data buffer administrative unit write back data file with certain page of data, if enliven transaction number on the transaction number that writes down on the page, then the privately owned log buffer district of these affairs was write back journal file; And obtaining unique logged sequence numbering, this privately owned log buffer district is reset then.
10. according to claim 5 or 6 described methods, it is characterized in that, when after system makes mistakes, recovering, at first reform.The process of reforming is: according to the order of sequence scanning day master chip and in all journal entry arrays, and determine whether reform according to the relation between the page sequence number that writes down in the journal entry content and the page sequence number on the related pages.
11. according to claim 5 or 6 described methods, it is characterized in that, after accomplishing when reforming, need enliven affairs to all and carry out rollback.Back off procedure is: enliven affairs for each, the day master chip that backward scanning is all to all journal entry arrays of each day master chip backward scanning, and carries out rollback operation to each the bar daily record in the journal entry array.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101915668A CN102760161A (en) | 2012-06-12 | 2012-06-12 | Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101915668A CN102760161A (en) | 2012-06-12 | 2012-06-12 | Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102760161A true CN102760161A (en) | 2012-10-31 |
Family
ID=47054618
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012101915668A Pending CN102760161A (en) | 2012-06-12 | 2012-06-12 | Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102760161A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102945278A (en) * | 2012-11-09 | 2013-02-27 | 华为技术有限公司 | Method and device for redoing logs of database records |
| CN104424351A (en) * | 2013-08-22 | 2015-03-18 | 国际商业机器公司 | Log data store that stores data across a plurality of storage devices using non-disjoint layers |
| CN105022676A (en) * | 2014-04-22 | 2015-11-04 | 大唐软件技术股份有限公司 | Recovery method and device of main memory database redo log files |
| CN107544866A (en) * | 2016-06-29 | 2018-01-05 | 北京忆恒创源科技有限公司 | A kind of daily record update method and device |
| CN107544912A (en) * | 2016-06-29 | 2018-01-05 | 北京忆恒创源科技有限公司 | A kind of log recording method, loading method and its device |
| CN108536619A (en) * | 2017-03-03 | 2018-09-14 | 北京忆恒创源科技有限公司 | The method and apparatus of fast quick-recovery FTL tables |
| CN110222035A (en) * | 2019-06-10 | 2019-09-10 | 天津神舟通用数据技术有限公司 | A kind of efficient fault-tolerance approach of database page based on exclusive or check and journal recovery |
| CN110659091A (en) * | 2018-06-29 | 2020-01-07 | 国际商业机器公司 | Session proxy dialog flow user interface |
| CN110750587A (en) * | 2019-10-15 | 2020-02-04 | 大连莱克科技发展有限公司 | Method for processing large amount of high-speed logs of debugging equipment |
| CN111858502A (en) * | 2020-06-02 | 2020-10-30 | 武汉达梦数据库有限公司 | Log reading method and log reading synchronization system based on log analysis synchronization |
| CN112256656A (en) * | 2019-07-22 | 2021-01-22 | 中兴通讯股份有限公司 | Transaction rollback method and device, database, system and computer storage medium |
| CN111858502B (en) * | 2020-06-02 | 2024-07-02 | 武汉达梦数据库股份有限公司 | Log reading method and system based on log analysis synchronization |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006095356A1 (en) * | 2005-03-10 | 2006-09-14 | Hewlett-Packard Development Company, L.P. | A method of logging transactions and a method of reversing a transaction |
| CN101794247A (en) * | 2010-03-26 | 2010-08-04 | 天津理工大学 | Real-time database failure recovery method under nested transaction model |
| CN102306168A (en) * | 2011-08-23 | 2012-01-04 | 成都市华为赛门铁克科技有限公司 | Log operation method and device and file system |
-
2012
- 2012-06-12 CN CN2012101915668A patent/CN102760161A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006095356A1 (en) * | 2005-03-10 | 2006-09-14 | Hewlett-Packard Development Company, L.P. | A method of logging transactions and a method of reversing a transaction |
| CN101794247A (en) * | 2010-03-26 | 2010-08-04 | 天津理工大学 | Real-time database failure recovery method under nested transaction model |
| CN102306168A (en) * | 2011-08-23 | 2012-01-04 | 成都市华为赛门铁克科技有限公司 | Log operation method and device and file system |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102945278B (en) * | 2012-11-09 | 2015-12-09 | 华为技术有限公司 | A kind of method and apparatus of data-base recording redo log |
| CN102945278A (en) * | 2012-11-09 | 2013-02-27 | 华为技术有限公司 | Method and device for redoing logs of database records |
| CN104424351A (en) * | 2013-08-22 | 2015-03-18 | 国际商业机器公司 | Log data store that stores data across a plurality of storage devices using non-disjoint layers |
| CN104424351B (en) * | 2013-08-22 | 2017-11-24 | 国际商业机器公司 | Across the method and system of the daily record data thesaurus of multiple storage device data storages |
| CN105022676A (en) * | 2014-04-22 | 2015-11-04 | 大唐软件技术股份有限公司 | Recovery method and device of main memory database redo log files |
| CN105022676B (en) * | 2014-04-22 | 2018-08-17 | 大唐软件技术股份有限公司 | A kind of memory database redoes log the restoration methods and device of file |
| CN107544866B (en) * | 2016-06-29 | 2021-01-05 | 北京忆恒创源科技有限公司 | Log updating method and device |
| CN107544866A (en) * | 2016-06-29 | 2018-01-05 | 北京忆恒创源科技有限公司 | A kind of daily record update method and device |
| CN107544912A (en) * | 2016-06-29 | 2018-01-05 | 北京忆恒创源科技有限公司 | A kind of log recording method, loading method and its device |
| CN108536619A (en) * | 2017-03-03 | 2018-09-14 | 北京忆恒创源科技有限公司 | The method and apparatus of fast quick-recovery FTL tables |
| CN108536619B (en) * | 2017-03-03 | 2021-12-14 | 北京忆恒创源科技股份有限公司 | Method and device for rapidly recovering FTL table |
| CN110659091A (en) * | 2018-06-29 | 2020-01-07 | 国际商业机器公司 | Session proxy dialog flow user interface |
| CN110222035A (en) * | 2019-06-10 | 2019-09-10 | 天津神舟通用数据技术有限公司 | A kind of efficient fault-tolerance approach of database page based on exclusive or check and journal recovery |
| CN112256656A (en) * | 2019-07-22 | 2021-01-22 | 中兴通讯股份有限公司 | Transaction rollback method and device, database, system and computer storage medium |
| CN110750587A (en) * | 2019-10-15 | 2020-02-04 | 大连莱克科技发展有限公司 | Method for processing large amount of high-speed logs of debugging equipment |
| CN110750587B (en) * | 2019-10-15 | 2023-04-25 | 大连莱克科技发展有限公司 | Method for processing large amount of high-speed logs of debugging equipment |
| CN111858502A (en) * | 2020-06-02 | 2020-10-30 | 武汉达梦数据库有限公司 | Log reading method and log reading synchronization system based on log analysis synchronization |
| CN111858502B (en) * | 2020-06-02 | 2024-07-02 | 武汉达梦数据库股份有限公司 | Log reading method and system based on log analysis synchronization |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102760161A (en) | Log organization structure clustered based on transaction aggregation and method for realizing corresponding recovery protocol thereof | |
| JP7410181B2 (en) | Hybrid indexing methods, systems, and programs | |
| CN103412803B (en) | The method and device that data are recovered | |
| CN102521269B (en) | Index-based computer continuous data protection method | |
| EP3117349B1 (en) | System and method for massively parallel processing database | |
| CN101930404B (en) | Memory device and operation method thereof | |
| CN103765393B (en) | Storage system | |
| KR101259557B1 (en) | Cluster data management system and method for data recovery using parallel processing in cluster data management system | |
| CN100583050C (en) | Method for protecting and recovering continuous data based on time stamp diary memory | |
| US10896156B2 (en) | Flexible synchronous file system replication | |
| CN104391903A (en) | Distributed storage and parallel calculation-based power grid data quality detection method | |
| CN101794247A (en) | Real-time database failure recovery method under nested transaction model | |
| CN105930500A (en) | Transaction recovery method in database system, and database management system | |
| CN102725739A (en) | Distributed database system by sharing or replicating the meta information on memory caches | |
| CN102012849B (en) | Flash memory-based database restoring method | |
| JP6296442B2 (en) | Highly efficient genome read alignment in in-memory databases | |
| US10650024B2 (en) | System and method of replicating data in a distributed system | |
| CN103488710B (en) | The non-fixed-length data method of efficient storage in big data page | |
| CN110389942A (en) | A kind of the key assignments separate-storage method and system of no garbage reclamation | |
| CN102750322A (en) | Method and system for guaranteeing distributed metadata consistency for cluster file system | |
| CN102779138A (en) | Hard disk access method of real time data | |
| CN101458655B (en) | Internal memory management method, system and internal memory control device | |
| CN105956041A (en) | Data model processing method based on Spring Data for MongoDB cluster | |
| Li et al. | FluteDB: An efficient and scalable in-memory time series database for sensor-cloud | |
| CN102750376A (en) | Multi-version database storage engine system and related processing implementation method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| DD01 | Delivery of document by public notice |
Addressee: Tianjin Shenzhou General Data Co., Ltd. Document name: Notification of Publication and of Entering the Substantive Examination Stage of the Application for Invention |
|
| DD01 | Delivery of document by public notice |
Addressee: Zhang Ye Document name: the First Notification of an Office Action |
|
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20121031 |