CN101930378A - Segmentation method of events in parallel large-scale distribution system - Google Patents
Segmentation method of events in parallel large-scale distribution system Download PDFInfo
- Publication number
- CN101930378A CN101930378A CN2010102216695A CN201010221669A CN101930378A CN 101930378 A CN101930378 A CN 101930378A CN 2010102216695 A CN2010102216695 A CN 2010102216695A CN 201010221669 A CN201010221669 A CN 201010221669A CN 101930378 A CN101930378 A CN 101930378A
- Authority
- CN
- China
- Prior art keywords
- affairs
- task
- resources
- new
- former
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a large-scale distributed parallel digitalized data processing and managing system in an interactive network environment. The invention aims to provide a segmentation method of events in a parallel large-scale distribution system; when the whole event is executed slowly, the method still can ensure that a part of resources are unlocked to provide corresponding services to other events. The segmentation method of events in the parallel large-scale distribution system is applied in an open cooperative environment; an event which has a manager is formed in a task completion process; in order to guarantee the consistency of data change in different resources when the task is completed, the event manager is entitled to lock some related resources; and the event is used for the resource management required by the task. The segmentation method is characterized in that: in the resources related to the task, each sub-task can only relate to a part of resources, and the operation of the part of resources can be managed by a newly established event so as to realize the segmentation of the total resources managed by the original event.
Description
Technical field
The present invention relates in an interactively network environment, large-scale distributed parallel digitalized data is handled and management system.
Background technology
After some performance elements in a digital display circuit were activated, this implementation may cause the content of many permanent storage sites in the system (database, file, external system etc.) to change.Often, these variations are success or failure together, otherwise the integrality of system data will be by permanent damage.In present technology, guarantee that this conforming way is to adopt a transaction management system, is collected in all variations that need to take place together in the same affairs; After relevant execution finished, transaction management system determined according to its result whether all changes of collecting take place, and guaranteed the consistance of generation.
Collect in the process that may change in affairs, the use of all some resources (data) that variation involved must be restricted, because, before these affairs finish, resource wherein can not be used, this restriction is called locking, and the intact unitarity of affairs also is to be based upon on the basis of this locking, otherwise just can't really realize the integrality of affairs.This shows that this locking must influence the progress of other its execution (if they need use same resource) in this environment.Time of locking is long more, and this things is also big more to the negative effect of system's operation.
In the environment of an open cooperation, a things possibly can't selectively involve a plurality of never relevant digital display circuits.If wherein any one system carries out slowly, other system just may find because the some of them resource is often caused this system normally to provide normal service to other execution by long-term locking.
Above-described transaction management system is mature technology and is widely used.The example that general things management system is used for the environment of open cooperation does not on a large scale also exist.The long-time things management system that designs for special applications (as workflow) is existing a lot.
Summary of the invention
Technical matters to be solved by this invention is, the method of the cutting of affairs in a kind of parallel large-scale distribution system is provided, in large-scale distributed parallel digitalized data processing and management system, if because of certain system carries out slowly, long-time lock part resource, therefore cause whole affairs to carry out when slow, still can make the resource of part carry out release, for other affairs provide corresponding service.
For finishing the foregoing invention purpose, the present invention is achieved in that the cutting method of affairs in a kind of parallel compartment system.It is applied in the environment of an open cooperation, comprises several server and client sides in the environment that this opening is cooperated; Each client can be submitted task to different server, and each server can be decomposed into task a series of subtasks and provide once more to different servers, by that analogy; The regulation of each task and subtask and the result who finishes transmit by changing the data cell that forms; Can cause the content changing of a series of dot data memories in the process that task is finished, such memory point is called a resource, and a resource is by many task sharings; The operation of all resources forms affairs in task complete process, and these affairs have a supvr; The consistance of the data variation when guaranteeing that task is finished in the different resource, the Service Manager has the right to lock the resource that some relates to, and refuses some operation in these resources of other affairs, finishes until affairs; Described affairs are used for the resource management of management role needs; It is characterized in that: in the resource that this task relates to, each subtask can only relate to a part wherein; The operation of this a part of resource can be stood transaction management in addition, thereby realizes the total resources of the former transaction management of cutting.
This scheme is because of opening new affairs in former affairs inside, and with the corresponding memory point of stylish affairs correspondence, and the unlatching of new affairs can't influence the execution of whole affairs, so has realized former affairs and the execution arranged side by side of affairs newly.In the execution of affairs, if the operation of a certain subsystem needs for a long time, then affairs can lock all resources, and other system can't be used, thereby can influence the performance of total system; And in this scheme owing to started new affairs, the resource of new things management can with corresponding subsystem finish and release makes system be able to these resources are offered other task, increase substantially the availability of system.
Such scheme is done further to limit, when the execution of former affairs do not rely on that son carries out as a result the time, former affairs are preserved executor and the data cell that this son is carried out; After treating that affairs run succeeded, set up new affairs, carry out this new affairs.Carry out failure, its new affairs abandoned as office.After described new affairs were carried out failure, its data cell that produces owing to new office was kept at this son automatically and carries out on the pairing memory point; Can not cause any change to system file.
Such scheme is done further to limit, and as the result that the execution of former affairs needs son to carry out, then former affairs data collection is set up new things, and carries out new affairs.The executor that described data comprises new affairs with return executor and relevant execution and return the data cell of execution.As the then whole transaction rollback of new affairs failure.Finally submit success to as former affairs, then former task manager abandons the person of returning of the new affairs of preserving; Former affairs prove an abortion, and then carry out the execution of returning of new affairs.In the process of implementation, because the relation of interdependence between the system, the execution that a kind of situation may occur and be exactly former affairs need rely on the result that son is carried out, at this moment, just need be by current former affairs data collection, set up new affairs, when setting up new affairs, also need to consider certainly, if affairs do not run succeeded, and the son execution was carried out, and then needed the person of returning of new affairs is preserved, so that rollback.
Technique scheme mainly is to be used for solving, and in the daily programming, have some resources to be taken for a long time, and other system needs these resources.Therefore, these resources could start after can only waiting former affairs to be finished.After adopting the affairs cutting, new things can just be implemented locking when needing resource really, in case and finish the subtask, thereby discharge the utilization in advance that all resources realize resources immediately.
Adopt above-mentioned affairs cutting method to go the cutting affairs, guaranteed the fundamental characteristics of former affairs on macroscopic view, i.e. the variation of all resources is unified successfully or is cancelled; But allow some valuable sources to be fast released again simultaneously.The particularly important is, this method use be fully local operation, that is, except cut-off (place that the subtask is distributed), do not relate to any other system, especially it does not need former Service Manager that the coordination of integral body is had anyly more to become; In a large-scale open system, this is crucial.
The analogy explanation
For this patent advantage is described more accurately, below we are further described the execution principle of this affairs cutting to this patent with the way of analogy.
Suppose to have a large-scale finishing project, it can involve a lot of all kinds of businessmans in set-up procedure, as the building supplies merchant, and household electrical appliances supplier, floor supplier, construction brigade, decoration corporation, or the like, wherein considerable part never cooperation always.Supposing that these businessmans and house-owner have the management system of oneself in above-mentioned open system, finish in this system and this fitting-out work itself is used as a task, is a feasible way of finishing below then.
General assignment is broken down into a series of subtasks, and as floor finiss, ceiling and wall finishing also have the kitchen, multimedia system or the like.Each subtask is divided into a series of subtasks again.For example, wall finishing team needs coating, and it can go the suppliers delivery of cargo of oneself; By that analogy, whole fitting-out work just can have been finished.
Because our example is finished the work by digital display circuit, it requires the success of all activities or failure to be correlated with.Such as, if the paint supply merchant is just in time in short supply, and other tasks are all finished according to plan, floor so, and the kitchen, stereo set just all has been contained in the room that does not whitewash.Digital display circuit is unlike manual system, unless design in advance, it does not have adaptability to changes.
In order to prevent that this situation from occurring, current system is collected in all change in resources in the affairs.Any one failure all causes the variation of all other resources not take place in the affairs, is called rollback.In the above in the example, if the coating supply shortage, then whole finishing rollback, all tasks are return former state, just as anything all not have generation.In order to accomplish this point, affairs need lock the resource that it needs.For example, can not run out of goods in order to guarantee coating, affairs find supplier in stockit is available in the locking supplier, make it to give other client's delivery.
Since affairs must be after task be only all finished could release, so, this supplier just can not deliver to other client during this fitting-out work.This is unacceptable.
The affairs cutting method has here been arranged, and this difficulty just can followingly solve: wall finishing team directly removes delivery of cargo the supplier, uses when preserving invoice in order to the return of goods then.If the delivery of cargo failure, whole fitting-out work failure, rollback comprehensively.If success, the delivery of cargo back supplier of release at once makes it normally provide service for other clients.Other reason rollback if fitting-out work is last, the coating after wall finishing team brings back and rolls and the invoice of preservation go the supplier to locate return goods (anti-affairs).
Claims (10)
1. the cutting method of affairs in the parallel compartment system, it is applied in the environment of an open cooperation, comprises several server and client sides in the environment that this opening is cooperated; Each client can be submitted task to different server, and each server can be decomposed into task a series of subtasks and provide once more to different servers, by that analogy; The regulation of each task and subtask and the result who finishes transmit by changing the data cell that forms; Can cause the content changing of a series of dot data memories in the process that task is finished, such memory point is called a resource, and a resource is by many task sharings; The operation of all resources forms affairs in task complete process, and these affairs have a supvr; The consistance of the data variation when guaranteeing that task is finished in the different resource, the Service Manager has the right to lock the resource that some relates to, and refuses some operation in these resources of other affairs, finishes until affairs; Described affairs are used for the resource management of management role needs; It is characterized in that: in the resource that this task relates to, each subtask can only relate to a part wherein; The operation of this a part of resource can be stood transaction management in addition, thereby realizes the total resources of the former transaction management of cutting.
2. the cutting method of affairs in a kind of parallel compartment system according to claim 1, it is characterized in that: the locking time that certain task is carried out the part in the used resource be when need be different from all the other resources, with the corresponding subtask of these resources can in addition upright new things management.
3. the cutting method of affairs in a kind of parallel compartment system according to claim 1 is characterized in that: when finishing of former task do not rely on the subtask as a result the time, former affairs are preserved executor and the data cell that this son is carried out; After treating that former task runs succeeded, set up new affairs to finish the subtask.
4. the cutting method of affairs in a kind of parallel compartment system according to claim 1 is characterized in that: when finishing of former task need the subtask as a result the time, former affairs data collection is set up new things and subtasking.
5. the cutting method of affairs in a kind of parallel compartment system according to claim 3 is characterized in that: carry out failure, its new affairs abandoned as task under the former office.
6. the cutting method of affairs in a kind of parallel compartment system according to claim 3, it is characterized in that: after described new affairs are carried out failure, it carries out on the server of operation no information loss because the data cell that new office produces is kept at this son automatically.
7. the cutting method of affairs in a kind of parallel compartment system according to claim 4 is characterized in that: the executor that described data comprises new task with return executor and relevant execution and return the data cell of execution.
8. the cutting method of affairs in a kind of parallel compartment system according to claim 4 is characterized in that: as the then whole transaction rollback of new affairs failure.
9. the cutting method of affairs in a kind of parallel compartment system according to claim 4 is characterized in that: run succeeded as new affairs, then the task manager person of returning that preserves new affairs continues to carry out.
10. the cutting method of affairs in a kind of parallel compartment system according to claim 9, it is characterized in that: former affairs are finally submitted success to, and then former task manager abandons the person of returning of the new affairs of preserving; Former affairs prove an abortion, and then carry out the execution of returning of new affairs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102216695A CN101930378A (en) | 2010-07-08 | 2010-07-08 | Segmentation method of events in parallel large-scale distribution system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102216695A CN101930378A (en) | 2010-07-08 | 2010-07-08 | Segmentation method of events in parallel large-scale distribution system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101930378A true CN101930378A (en) | 2010-12-29 |
Family
ID=43369573
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010102216695A Pending CN101930378A (en) | 2010-07-08 | 2010-07-08 | Segmentation method of events in parallel large-scale distribution system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101930378A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111752685A (en) * | 2020-05-22 | 2020-10-09 | 清华大学 | Persistent memory transaction submitting method under multi-core architecture |
| CN112003827A (en) * | 2020-07-27 | 2020-11-27 | 山东英信计算机技术有限公司 | Configurable Redfish service analysis method and system based on parallel computing |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1507597A (en) * | 2001-05-15 | 2004-06-23 | �Ҵ���˾ | Method for managing distributed savepoints across multiple DBMS's within a distributed transaction |
| CN1801737A (en) * | 2005-01-06 | 2006-07-12 | 中兴通讯股份有限公司 | Method for gateway service node processing business service node restarting |
| CN101184106A (en) * | 2007-11-30 | 2008-05-21 | 北京航空航天大学 | Associated transaction processing method of mobile database |
-
2010
- 2010-07-08 CN CN2010102216695A patent/CN101930378A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1507597A (en) * | 2001-05-15 | 2004-06-23 | �Ҵ���˾ | Method for managing distributed savepoints across multiple DBMS's within a distributed transaction |
| CN1801737A (en) * | 2005-01-06 | 2006-07-12 | 中兴通讯股份有限公司 | Method for gateway service node processing business service node restarting |
| CN101184106A (en) * | 2007-11-30 | 2008-05-21 | 北京航空航天大学 | Associated transaction processing method of mobile database |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111752685A (en) * | 2020-05-22 | 2020-10-09 | 清华大学 | Persistent memory transaction submitting method under multi-core architecture |
| CN111752685B (en) * | 2020-05-22 | 2022-09-23 | 清华大学 | Persistent memory transaction submitting method under multi-core architecture |
| CN112003827A (en) * | 2020-07-27 | 2020-11-27 | 山东英信计算机技术有限公司 | Configurable Redfish service analysis method and system based on parallel computing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102306197B (en) | Device and method for guaranteeing consistency of data-source-crossing operation results | |
| CN109194538A (en) | Test method, device, server and storage medium based on distributed coordination | |
| CN102523254B (en) | Method for automatically distributing systematic resources of cloud computing environment | |
| CN102655532A (en) | Distributed heterogeneous virtual resource integration management method and system | |
| CN106817408B (en) | Distributed server cluster scheduling method and device | |
| CN107168779A (en) | A kind of task management method and system | |
| CN106033373A (en) | A method and a system for scheduling virtual machine resources in a cloud computing platform | |
| CN105210326B (en) | Cloud application processing method and application dispositions method and relevant apparatus and system | |
| CN101960439B (en) | Client environment creates system and client environment creation method | |
| CN103095845B (en) | A kind of method and system realizing distributed communication | |
| CN105700958A (en) | Method and system for automatic splitting of task and parallel execution of sub-task | |
| CN103197952A (en) | Management system and method aiming at maintenance and deployment of application system based on cloud infrastructure | |
| CN105404549B (en) | Scheduling virtual machine system based on yarn framework | |
| CN102495759A (en) | Method for scheduling job in cloud computing environment | |
| CN110661842B (en) | Resource scheduling management method, electronic equipment and storage medium | |
| CN104410511B (en) | A kind of server management method and system | |
| CN105404530A (en) | System and method for realizing simple deployment and using private cloud | |
| CN101930378A (en) | Segmentation method of events in parallel large-scale distribution system | |
| CN102243653A (en) | Method and device for managing database connections | |
| CN102243655A (en) | Method and device for managing database connections | |
| CN107589972B (en) | Method for sharing IPC (Internet protocol Camera) name space by Pod in Kubernetes | |
| CN112003931B (en) | Method and system for deploying scheduling controller and related components | |
| CN101834729A (en) | Business process distributing and processing system | |
| CN108810107B (en) | Cloud management platform and service method of two-state IT (information technology) business | |
| CN102868594B (en) | Method and device for message processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20101229 |