CN101013381B - Distributed lock based on object memory system - Google Patents
Distributed lock based on object memory system Download PDFInfo
- Publication number
- CN101013381B CN101013381B CN200710051406A CN200710051406A CN101013381B CN 101013381 B CN101013381 B CN 101013381B CN 200710051406 A CN200710051406 A CN 200710051406A CN 200710051406 A CN200710051406 A CN 200710051406A CN 101013381 B CN101013381 B CN 101013381B
- Authority
- CN
- China
- Prior art keywords
- lock
- data
- module
- metadata
- lock request
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
It's an object-based storage system to model attributes of distributed locks, a distributed storage system for storage of multi-client visit to the sharing of resources, aimed at enhancing the efficiency of Reading lock visit. The present invention include data locking and metadata management for the management of locks, running on the object-based storage system, data management lock-operation in various storage devices on the target, including data objects name space module, access patterns attribute value module installed, local Queue processing module and functional processing module; Each data lock manager responsible for data management object locking, metadata management lock-operation in various metadata servers, including metadata objects name space module, local Queue processing module and functional processing modules, each metadata management lock-lock for metadata management. The invention uses the object storage technology by expanding access pattern attribute value and improves the efficiency of Reading lock visit.
Description
Technical field
The invention belongs to distributed memory system, be specifically related to distributed lock, be used to realize the share and access of many clients storage resources based on object storage system access module attribute.
Background technology
Along with infotech, rapid development of network technology, various sharing with the data message that exchanges sharply increases with high speed, increasing user to the capacity and the bandwidth demand of storage products present routinely, volatile growth, network store system towards mass data emerges in large numbers one after another: as NAS, SAN, OBS.
In the distributed network storage system, a plurality of clients can be by the shared resource on each node of computer network access.In order to make correctly and effectively operation of distributed memory system, must control of the concurrent visit of a plurality of clients to same shared resource, so that prevent uncontrolled, the concurrent or while access shared resources of some probabilistic behavior, cause the consistance of data and metadata to be destroyed, lock is the conventional means that realizes concurrent access control.At present, being used to of having developed realizes that the lock of concurrent control is to be based upon on the basis of file or piece storage mostly, as pNFS, and GPFS, Blue Whale Distributed File System or the like.Yet, all there are a deficiency in file and piece interface, be exactly the enough descriptions that all lack self, shared resource all is passive acceptance management, shared resource can not give full expression to the characteristics relevant with access module of self, makes that the efficient of concurrent control particularly is not very high under the situation of read operation.Based on object storage is a kind of newer memory technology, to as if an interface that is rich in expressive force, but the abundant extended attribute of object can make based on the concurrent control of object storage system more efficient.Lustre, panFS be present industry famous based on object-based file system, yet Lustre adopts expansion lock (extent lock) to file, and metadata is used the lock (intent lock) of band intention, and the expansion lock be in fact one based on other lock of piece level; The lock of panFS realizes being based on equally piece, file.Therefore, they do not utilize the extensibility of object on the realization mechanism of lock, just fail to have given play to the advantage that object storage brings for concurrent control yet.
Summary of the invention
The invention provides a kind of distributed lock based on object storage system, purpose is the advantage of utilizing object storage technology to bring, adopts the access module property value of expansion, improves the access efficiency of read lock.
Distributed lock based on object storage system of the present invention, be used for guaranteeing data object and the consistance of metadata object in the share and access process, comprise data lock manager and metadata lock manager, run on based on the object storage system, describedly comprise meta data server, object storage equipment, client and be connected network based on object storage system; Wherein:
(1) the data lock manager runs on each object storage equipment, and this data lock manager comprises that data object resource name word space module, access module property value are provided with module, the first local queue processing module and first functional processing module; Each data lock manager is responsible for the management of data lock, comprises the mandate and the recovery of data lock;
(1.1) all data objects on the object storage equipment of data object resource name word space module management place are found out the relevant information of this data object according to data object ID value;
(1.2) the access module property value is provided with the access module attribute that module is used to be provided with data object;
The processing that (1.3) first local queue processing modules are responsible for having authorized lock request queue and are waited for lock request queue: all the lock requests in the lock request queue of waiting for all are and one or more certain sharing data objects resource of lock request mutual exclusion accessing of authorizing in the lock request queue, only ought authorize in the lock request queue certain lock request service to finish, when discharging its data object resource of occupying, if this data object resource is identical with the data object resource that the lock request of waiting for lock request queue requires, according to the FIFO mode, the request of should locking is put into and is authorized lock request queue, waits for that simultaneously lock request queue deletes this lock and ask;
(1.4) first functional processing modules are responsible for handling the lock request that application process is sent, and comprise the mandate and the recovery of data lock; At first obtain the relevant information of data object from data object resource name word space module according to lock request corresponding data objects ID, the access module attribute that comprises data object, after process access module property value is provided with resume module, whether the corresponding data objects resource of being judged the lock request by the first local queue processing module can obtain, if can, this lock request is inserted into authorizes lock request queue, obtains mandate, otherwise the wait lock request queue is inserted in this lock request;
(2) the metadata lock manager runs on each meta data server, this metadata lock manager comprises metadata object resource name word space module, the second local queue processing module and second functional processing module, each metadata lock manager is responsible for the management of metadata lock, comprises the mandate and the recovery of metadata lock;
(2.1) all metadata objects on the meta data server of metadata object resource name word space module management place are found out the relevant information of this metadata object according to metadata object ID value;
The first local queue processing module of the function of (2.2) second local queue processing modules and (1.3) data lock manager is identical;
The function of (2.3) second functional processing modules is not provided with the function of module except there being the access module property value, (1.4) first functional processing module of data lock manager is identical, and mandate, removal process that mandate, removal process and the data of metadata lock are locked are identical.
Described distributed lock based on object storage system is characterized in that, the access module property value of described data lock manager is provided with module, and that the access module attribute process of data object is set is as follows:
(1) after the access module property value initialization of each data object is set to RDWT, starts timer;
(2) at timer in pre-set time, if the visit of this data object all is read operation, then the access module property value of this data object is set to READ;
(3) in follow-up access process, run into write operation, after then the access module property value of this data object is set to RDWT, restart timer.
The present invention utilizes the object accesses pattern attribute of expansion, the data lock manager receives a new request, if read operation, then find the corresponding object resource information of this request lock fast by data object resource name word space module after, check the access module property value of this object, when the value of its pattern attribute was READ, this data lock manager was directly returned one and is obtained the read lock success, and no longer really apply for a read lock resource needed, thereby greatly improve the efficient of read lock.
Description of drawings
Fig. 1 is based on object distribution formula storage system synoptic diagram;
Fig. 2 represents based on one in the object distribution formula storage system simple concurrent problem;
The process that problem produces in Fig. 3 presentation graphs 2;
Fig. 4 represents that the present invention is used for the situation based on object distribution formula storage system;
Fig. 5 is the composition structural representation of data lock manager of the present invention;
Fig. 6 is the treatment scheme synoptic diagram of local queue processing module of the present invention;
Fig. 7 is the access module property value setting up procedure synoptic diagram of data object;
Fig. 8 is the treatment scheme synoptic diagram of data lock manager of the present invention;
Embodiment
Fig. 1 is based on object distribution formula storage system synoptic diagram.Comprise four assemblies based on object distribution formula storage system: 102,104,106,108, three client nodes 122,124,126 of 112,114,116, four object storage equipment of three metadata server nodes and TCP/IP network 132.In system as shown in Figure 1, the data object storage resources on the object storage equipment 102-108 can be by client 122-126 share and access; Meta data server 112-116 preserves and manages the metadata information of the data object of total system.The flow process of client-access system is: at first, obtain the metadata information that it wants access object to meta data server, then according to the metadata information that returns directly and object storage equipment carry out communication, carry out associative operations such as data transmission, after operation is finished, client feeds back to meta data server to status information, upgrades relevant metadata information.Therefore, the metadata on the meta data server also is a shared resource.
Fig. 2 represents based on one in the object distribution formula storage system simple concurrent problem.First row at Fig. 2 is the partial data resource 202 of certain data object of sharing, and comprises 8 continuous in logic data storage block 41-48; Second row and the third line of Fig. 2 have provided 203,204 two processes are operated the 43-45 data block of shared data resource 202, the situation that contention occurs, first process 203 writes " X " respectively to the 43-45 of sharing data resources 202 data block, " Y " and " Z ", second process 204 writes " A " respectively to the 43-45 of sharing data resources 202 data block, " B " and " C ", when these two processes simultaneously concomitantly when wherein writing, the result that may occur is unpredictable: acceptable correct result both may occur, also error result may occur.The fourth line of Fig. 2 and fifth line show the correct two kinds of possible outcomes 205,206 of shared resource, to result 205, be that process 204 writes " A " toward 43-45 earlier, " B " and " C ", then be that process 203 writes " X " toward 43-45, " Y " and " Z ", after the write operation in the time of through two processes, the net result of sharing data resources is an acceptable; To result 206, process to write precedence relationship just opposite, still, its result remains acceptable.The nethermost delegation of Fig. 2 shows a kind of mistake and chaotic result 207, causes this result's possible cause to illustrate at Fig. 3, and step 301-306 shows the data variation process of shared resource 43-45:
Step 301: process 204 writes " A " toward 43 data blocks;
Step 302: process 203 writes " X " toward 43 data blocks, covers original " A ";
Step 303: process 203 writes " Y " toward 44 data blocks;
Step 304: process 204 writes " B " toward 44 data blocks, covers original " Y ";
Step 305: process 204 writes " C " toward 45 data blocks;
Step 306: process 203 writes " Z " toward 45 data blocks, covers original " C ";
207 only is a kind of chaotic situation that may occur in the contention process as a result, also have the multiple unacceptable result that may occur, and under most of situation, the result is unexpected.
Fig. 4 represents that the present invention is used for the situation based on object distribution formula storage system.In order to prevent to cause the invalid of shared resource and unacceptable result occurs owing to the contention of resource, at characteristics, adopted the mode of distributed lock to guarantee the validity and the consistance of data object resource and metadata object based on object distribution formula storage system.Be service data object resources effective, increased data lock manager 411-417 respectively on object storage equipment 102-108, each data lock manager is responsible for the management of data lock, comprises the mandate and the recovery of data lock; Equally, for safeguarding the metadata object resources effective, increased metadata lock manager 401-405 respectively on object storage equipment 112-116, each metadata lock manager is responsible for the management of metadata lock, comprises the mandate and the recovery of metadata lock.Each data lock manager and metadata lock manager all have a unique identifier GUID of the overall situation to identify respectively, and, the node at data lock manager and metadata lock manager and its place is to concern one to one, that is to say, each object storage equipment has and only has a data lock manager, and each meta data server has and only have a metadata lock manager.
Fig. 5 represents the composition framework of data lock manager 501, comprising: data object resource name word space module 504, access module property value are provided with module 505, local queue processing module 507 and functional processing module 512.The GUID of data lock manager correspondence is shown 502, and it has represented a data lock manager that the overall situation is unique; Data object resource name word space module illustrates by 504, because in based on object distribution formula storage system, the quantity of shared resource (comprising data object resource and metadata resource) all is magnanimity, for the ease of the relevant information of managing these data and in so huge data, can retrieve certain resource fast, adopt the resource name word space to manage; Local queue processing module 507 comprises two different lock request queues: authorized lock request queue 506 and waited for lock request queue 508, authorized the current lock formation that has obtained certain shared resource granted access of lock request queue 506 expressions, and waited for that lock request queue 508 expressions temporarily can't obtain the lock formation of this shared resource access authorization.These two formations have inherent contact, and the lock request among both will transform under a stable condition, and this conversion is realized by local queue processing module 507.
The metadata lock manager is not provided with the module except there being the access module property value, and its module frame is identical with data lock manager 501.
Fig. 6 represents the treatment scheme of local queue processing module, and under initial conditions, 506 and 508 these two formations all are empty, and first lock request is can obtain access authorization certainly, and still, follow-up lock request will have been handled accordingly.Treatment scheme is:
(1) wins a lock request the lock request queue from authorizing;
(2) activate the corresponding process of this lock, allow this process that the lock requested resource is conducted interviews;
(3) finish when the process access resources, discharge the resource that it occupies, its lock is also deleted from authorizing the lock request queue;
(4) check the wait lock request queue then, seeing if there is request resource is the lock request that discharged just now; If have then change step (5), then do not change step (1);
(5) this lock request is authorized lock request queue by waiting for that lock request queue changes over to, change step (1).
Concerning based on object storage system, this interface that is rich in expressive force of object give distributed lock the performance boost band opportunity, utilize the characteristics of object properties, expanded the attribute of the access module of object, write down the access module one section pre-set time range the inside of object, this schedule time is set to T, and Fig. 7 is the access module property value setting up procedure synoptic diagram of data object.The access module property value of data object illustrates by 602, it has only two kinds of possible value: READ and RDWT, when property value 602 is READ, the accessing operation that is illustrated in data object in the T time period all is to read, follow-up accessing operation 604-608 reads (RD), it is constant that the value of access module is kept READ, and the timeing closing device; In case number of write access operations (WT) occurred, as the 612-614 among Fig. 6, the value of access module is set to RDWT at once, as long as a read operation 616 occurs, just start timer, if at timer all is read operation in the T at the fixed time, is READ with regard to the value of resetting access module so, and the timeing closing device.
Fig. 8 represents the treatment scheme of data lock manager:
(1) handles the lock request that arrives;
(2) lock corresponding data objects ID finds this data object from data object resource name word space information according to data, comprise object accesses pattern property value;
(3) judgment data lock requested operation type is carried out different branch process;
(4) read operation change (5), otherwise the access module property value that data object is set is RDWT, and the cancellation timer changes (8);
(5) the access module value of inspection data object is judged it;
(6) if READ returns the lock request that (1) handles the back; Otherwise be RDWT, check whether being provided with of timer;
(7) if timer is provided with, check whether timer has arrived a little, has arrived a little, and the access module property value that data object is set is READ, the timeing closing device changes (8) simultaneously; Otherwise directly change (8);
If timer is not provided with, the access module property value that data object is set is READ, and timer is set, and changes (8);
(8) judge that whether this lock request and the arbitrary lock request in the lock request queue of waiting for conflict mutually, if conflict is mutually inserted the wait lock request queue to this lock request, change step (1); Otherwise lock request queue has been authorized in this lock request insertion, changeed step (1).
Claims (2)
1. distributed lock system based on object storage system, be used for guaranteeing data object and the consistance of metadata object in the share and access process, comprise data lock manager and metadata lock manager, run on based on the object storage system, describedly comprise meta data server, object storage equipment, client and the network that is connected based on object storage system; Wherein:
(1) the data lock manager runs on each object storage equipment, and this data lock manager comprises that data object resource name word space module, access module property value are provided with module, the first local queue processing module and first functional processing module; Each data lock manager is responsible for the management of data lock, comprises the mandate and the recovery of data lock;
(1.1) all data objects on the object storage equipment of data object resource name word space module management place are found out the relevant information of this data object according to data object ID value;
(1.2) the access module property value is provided with the access module attribute that module is used to be provided with data object;
The processing that (1.3) first local queue processing modules are responsible for having authorized lock request queue and are waited for lock request queue: all the lock requests in the lock request queue of waiting for all are and one or more certain sharing data objects resource of lock request mutual exclusion accessing of authorizing in the lock request queue, only ought authorize in the lock request queue certain lock request service to finish, discharge the data object resource that it occupies, and when it is deleted the lock request queue from authorizing, if this data object resource is identical with the data object resource that the lock request of waiting for lock request queue requires, according to the FIFO mode, the request of should locking is put into and is authorized lock request queue, waits for that simultaneously lock request queue deletes this lock and ask;
(1.4) first functional processing modules are responsible for handling the lock request that application process is sent: the relevant information of at first obtaining data object according to lock request corresponding data objects ID value from data object resource name word space module, the access module attribute that comprises data object, after process access module property value is provided with resume module, whether the corresponding data objects resource of being judged the lock request by the first local queue processing module can obtain, if can, this lock request is inserted into authorizes lock request queue, obtain mandate, otherwise the wait lock request queue is inserted in this lock request;
(2) the metadata lock manager runs on each meta data server, this metadata lock manager comprises metadata object resource name word space module, the second local queue processing module and second functional processing module, each metadata lock manager is responsible for the management of metadata lock, comprises the mandate and the recovery of metadata lock;
(2.1) all metadata objects on the meta data server of metadata object resource name word space module management place are found out the relevant information of this metadata object according to metadata object ID value;
The first local queue processing module of the function of (2.2) second local queue processing modules and (1.3) data lock manager is identical;
The function of (2.3) second functional processing modules is not provided with the function of module except there being the access module property value, and is identical with first functional processing module of (1.4) data lock manager.
2. the distributed lock system based on object storage system as claimed in claim 1 is characterized in that, the access module property value of described data lock manager is provided with module, and that the access module attribute process of data object is set is as follows:
(1) after the access module property value initialization of each data object is set to RDWT, starts timer;
(2) at timer in pre-set time, if the visit of this data object all is read operation, then the access module property value of this data object is set to READ;
(3) in follow-up access process, run into write operation, after then the access module property value of this data object is set to RDWT, restart timer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710051406A CN101013381B (en) | 2007-01-26 | 2007-01-26 | Distributed lock based on object memory system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710051406A CN101013381B (en) | 2007-01-26 | 2007-01-26 | Distributed lock based on object memory system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101013381A CN101013381A (en) | 2007-08-08 |
| CN101013381B true CN101013381B (en) | 2010-05-19 |
Family
ID=38700913
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200710051406A Expired - Fee Related CN101013381B (en) | 2007-01-26 | 2007-01-26 | Distributed lock based on object memory system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101013381B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104603762A (en) * | 2012-09-28 | 2015-05-06 | 国际商业机器公司 | Supporting coordinated access to file system's shared storage using automatic alignment of parallel file access protocol and metadata management |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101187930B (en) * | 2007-12-04 | 2010-06-09 | 浙江大学 | Distribution type file system dummy directory and name space implementing method |
| CN101252603B (en) * | 2008-04-11 | 2011-03-30 | 清华大学 | Cluster distributed type lock management method based on storage area network SAN |
| US20100042659A1 (en) * | 2008-08-15 | 2010-02-18 | Electronic Data Systems Corporation | Deployment overview management system, apparatus, and method |
| US8161019B2 (en) * | 2009-05-01 | 2012-04-17 | Microsoft Corporation | Cross-channel coauthoring consistency |
| CN101631328B (en) * | 2009-08-14 | 2011-11-16 | 北京星网锐捷网络技术有限公司 | Synchronous method performing mutual exclusion access on shared resource, device and network equipment |
| CN101997823B (en) * | 2009-08-17 | 2013-10-02 | 联想(北京)有限公司 | Distributed file system and data access method thereof |
| CN102508720B (en) * | 2011-11-29 | 2017-02-22 | 中能电力科技开发有限公司 | Method for improving efficiency of preprocessing module and efficiency of post-processing module and system |
| CN103067471B (en) * | 2012-12-24 | 2019-01-11 | 北京华科力扬科技有限公司 | A kind of data capture method of distributed storage |
| US9658899B2 (en) * | 2013-06-10 | 2017-05-23 | Amazon Technologies, Inc. | Distributed lock management in a cloud computing environment |
| CN104702655B (en) * | 2014-03-21 | 2018-04-27 | 杭州海康威视系统技术有限公司 | Cloud storage resource allocation methods and its system |
| CN105635082A (en) * | 2014-11-12 | 2016-06-01 | 北大方正集团有限公司 | Dynamic authorization method and system, authorization center, and authorization client |
| CN104461705B (en) * | 2014-11-17 | 2019-02-19 | 华为技术有限公司 | A kind of method and storage control, cluster storage system of business access |
| CN105095353B (en) * | 2015-06-19 | 2018-12-04 | 中国科学院计算技术研究所 | The equal system and method that does after small documents is pre-read in a kind of parallel network file system |
| CN105631023B (en) * | 2015-12-30 | 2019-03-26 | 华为技术有限公司 | The method and apparatus of distributed lock service |
| CN108733477B (en) * | 2017-04-20 | 2021-04-23 | 中国移动通信集团湖北有限公司 | Method, device and equipment for data clustering processing |
| CN109669632A (en) * | 2018-12-10 | 2019-04-23 | 浪潮电子信息产业股份有限公司 | Metadata wiring method, device and medium based on distributed objects storage system |
| CN109344136A (en) * | 2018-12-13 | 2019-02-15 | 浪潮(北京)电子信息产业有限公司 | A kind of access method of shared-file system, device and equipment |
| CN111782609B (en) * | 2020-05-22 | 2023-10-13 | 北京和瑞精湛医学检验实验室有限公司 | Method for rapidly and uniformly slicing fastq file |
| CN113050893B (en) * | 2021-03-30 | 2022-08-30 | 重庆紫光华山智安科技有限公司 | High-concurrency file storage method, system, medium and electronic terminal |
| CN115114305B (en) * | 2022-04-08 | 2023-04-28 | 腾讯科技（深圳）有限公司 | Lock management method, device, equipment and storage medium for distributed database |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6041383A (en) * | 1996-07-22 | 2000-03-21 | Cabletron Systems, Inc. | Establishing control of lock token for shared objects upon approval messages from all other processes |
| CN1728665A (en) * | 2005-07-26 | 2006-02-01 | 华中科技大学 | Expandable storage system and control method based on objects |
| CN1786918A (en) * | 2004-12-10 | 2006-06-14 | 惠普开发有限公司 | Distributed lock |
-
2007
- 2007-01-26 CN CN200710051406A patent/CN101013381B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6041383A (en) * | 1996-07-22 | 2000-03-21 | Cabletron Systems, Inc. | Establishing control of lock token for shared objects upon approval messages from all other processes |
| CN1786918A (en) * | 2004-12-10 | 2006-06-14 | 惠普开发有限公司 | Distributed lock |
| CN1728665A (en) * | 2005-07-26 | 2006-02-01 | 华中科技大学 | Expandable storage system and control method based on objects |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104603762A (en) * | 2012-09-28 | 2015-05-06 | 国际商业机器公司 | Supporting coordinated access to file system's shared storage using automatic alignment of parallel file access protocol and metadata management |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101013381A (en) | 2007-08-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101013381B (en) | Distributed lock based on object memory system | |
| CN108268208B (en) | RDMA (remote direct memory Access) -based distributed memory file system | |
| CN107888657A (en) | Low latency distributed memory system | |
| US9710407B2 (en) | Congestion control in storage systems | |
| CN103116634B (en) | System for supporting high concurrent cache task queue and asynchronous batch operation method thereof | |
| US9495478B2 (en) | Namespace management in distributed storage systems | |
| US10372685B2 (en) | Scalable file storage service | |
| US9569459B1 (en) | Conditional writes at distributed storage services | |
| CN101019105B (en) | Method and apparatus for data storage using striping | |
| CN100517335C (en) | Distributed file system file writing system and method | |
| US7647454B2 (en) | Transactional shared memory system and method of control | |
| CN104346317B (en) | Shared resource access method and device | |
| US20150277969A1 (en) | Atomic writes for multiple-extent operations | |
| US7783607B2 (en) | Decentralized record expiry | |
| US20070143340A1 (en) | System and method of time-based cache coherency maintenance in user file manager of object-based storage system | |
| US20150242311A1 (en) | Hybrid dram-ssd memory system for a distributed database node | |
| US20170256023A1 (en) | Solid state storage local image processing system and method | |
| CN102984256B (en) | Processing method and system for metadata based on authorization manner | |
| CN116501267B (en) | Independent redundant disk array card control method and device | |
| CN107832159A (en) | Manage the method and computer-readable recording medium of distributed lock | |
| CN107832017B (en) | Method and device for improving storage IO performance of seismic data | |
| CN104461705B (en) | A kind of method and storage control, cluster storage system of business access | |
| CN110162395B (en) | Memory allocation method and device | |
| CN111435286A (en) | Data storage method, device and system | |
| CN106202271A (en) | The read method of the product database of OTA |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100519 Termination date: 20220126 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |