CN101179516B - Digraph based data distributing method - Google Patents
Digraph based data distributing method Download PDFInfo
- Publication number
- CN101179516B CN101179516B CN2006101144975A CN200610114497A CN101179516B CN 101179516 B CN101179516 B CN 101179516B CN 2006101144975 A CN2006101144975 A CN 2006101144975A CN 200610114497 A CN200610114497 A CN 200610114497A CN 101179516 B CN101179516 B CN 101179516B
- Authority
- CN
- China
- Prior art keywords
- data
- transmission
- dependence
- directed graph
- task
- 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
Abstract
The invention discloses a data distribution method based on an oriented graph, which includes the following steps: A. the oriented graph is used for defining a data distribution job with an oriented graph definition, that is Job = (N, D, T), wherein, N is a set of nodes, each of which stands for a unit or a software system joining in data exchange; D is a set of data elements; T is a set of sides,each of which stands for a data transmission task including data and data flow direction, and each transmission task T is a triad (Src, Dest, Data), wherein, Src belongs to N and stands for a sourcenode; Dest belongs to N and stands for a target node; Src and Dest form an oriented side <SRC, <td>.
Description
Technical field
The present invention relates to the data distributing method of a kind of network information resource in sharing, refer to a kind of data distributing method especially based on directed graph.
Background technology
Data sharing and exchange are basic problems of information technology, and along with rapid development of Internet, development of science and technology, accumulation of knowledge, how with one's own mass data, information security, apace by the Internet and other people share, exchange becomes more complicated.For example, more and more data need be shared by increasing library; Many R﹠D institutions are after generation is the mass data of unit with T, the research institution that these data need be distributed to different regions even country goes to analyze, suchlike resource-sharing problem needs to provide a kind of safer, transmission service efficiently to come swap data urgently, provide more convenient and effective MRP service simultaneously, control resource flowing between the sharer.
The GLOBUS software product provides a kind of reliable file transmission (Reliable File Transfer, be called for short RFT) service, it is a kind of and Open Grid Services Architecture (Open Grid Services Infrastructure, OGSI) Jian Rong mesh services, it can control data adopts the grid file host-host protocols at two, and (GridFile Transfer Protocol transmits between server GridFTP).As operation, each operation can comprise a plurality of files with transfer of data in the RFT service.
The RFT service class that provides with the GLOBUS software product seemingly, the gLite software product also provides a kind of file transfer services, and (File Transfer Service FTS), and also adopts GridFTP as host-host protocol.FTS is the same with RFT, all is the role who plays the part of " third party " outside the source node of transfer of data and destination node.
The shortcoming of RFT and FTS is: can only control the shared and exchange of two data between nodes, promptly so-called 1-to-1 data sharing.Yet the data sharing problem is not only sharing between two nodes, and data have been distributed to M data consumption node from N data source node, and promptly so-called N-to-M data sharing also is very general a kind of phenomenon.For this class problem, RFT service and FTS seem too simple.
The system that meets N-to-M data sharing, distribution is less, and (CompositeEndpoint Protocol CEP) is a general N-to-M data sharing communication protocol to the combining endpoints agreement, it makes S1, S2, S3,, Sn is for providing the node of data, R1, R2, R3 ...., Rn is for receiving the node of data, each Si provides a part of data, can be overlapping between the data, constitute linear data space; Each Ri needs a part of data in the data space.The solution of CEP was divided into for two steps: at first, API (API) by CEP is described transmission demand, comprise data transmitting node tabulation, data reception node tabulation, the data that each sending node can be visited, and the needed data of each receiving node; Secondly, send by the data between the scheduler Control Node.
The subject matter of CEP is not exist between the node structure, therefore, if the A node has data DataX, and B, C, D, four nodes of E all need data DataX, and DataX will be transferred to B respectively from the A node so, C, D, the E node, obviously the load of A node is higher.
The grid work flow technology is the technology that another kind is used for the data distribution.Grid work streaming system such as DAGMan, GridFlow, GrADS, Gridbus workflow, Taverna, Pegasus etc. all are that (Directed Acyclic Graph, DAG) as workflow modeling method, definition of data is distributed operation to the employing directed acyclic graph.When these grid work streaming systems use DAG, use the vertex representation activity usually, represent sequence between the activity, parallel and choice relation with the limit.If the grid work flow system applies is distributed in data, so every transformation task is exactly an activity.For the user, the grid work streaming system is not data-centered, but be the center with the transformation task, need the user to go to define transformation task and the relation between them, this undoubtedly definition mode is not directly perceived, can't directly embody data flowing between node, and, the difficulty that user job defines increased.
Summary of the invention
In order to solve the problem of N-to-M data distribution, main purpose of the present invention provides the data distributing method based on directed graph that a kind of novel representation of data intuitively flows between each node.
Another object of the present invention provides a kind of data distributing method based on directed graph that can check the operation legitimacy.
Another purpose of the present invention provides a kind of data distributing method based on directed graph that can be optimized operation.
For achieving the above object, the present invention by the following technical solutions: a kind of data distributing method based on directed graph, this method may further comprise the steps:
The first step: with directed graph definition of data distribution operation
A) with directed graph definition of data distribution operation (Job), promptly Job=(N, D, T), wherein N is the set of node, each node is represented the unit or the software systems that participate in exchanges data; D is the set of data element; T is the set on limit, and a data transfer task is represented on every limit, comprises the flow direction of data and data, and each transformation task T is that (Data), wherein Src ∈ N represents source node to a tlv triple for Src, Dest; Dest ∈ N, the expression destination node, Src and Dest constitute directed edge＜Src, Dest 〉, Data is the subclass of D, represents the data of this transmission;
B) will convert corresponding task description file to the data distribution operation of directed graph definition;
Second step: task description file is submitted to the transmission Control Server, carry out data distribution transmission operation;
The 3rd step: the transmission Control Server is notified the user with execution result.
Described second step: task description file is submitted to the transmission Control Server, carry out data distribution transmission operation and specifically may further comprise the steps:
A) task description file is submitted to the transmission Control Server, the transmission Control Server is checked the validity of operation;
Transmission at first checks operation after Control Server receives operation, judges between the transformation task of operation whether have dependence; Then, utilize topological sorting algorithm to judge whether dependence constitutes loop,, have the operation of loop, illustrate that then these operations are illegal, can't be performed if some dependence has constituted loop; Otherwise,, illustrate that then these operations are legal, can be performed if there is not loop between these dependences;
B) redundant transmission in the cancel job;
C) transmission of decomposition coarseness makes each transformation task send littler data;
D) actual figure transporting operation reportedly.
In the described step a) between the transformation task definite method of dependence as follows:
Every limit T of the data distribution operation of (1) representing with directed graph is corresponding, increases a corresponding summit Nt in dependence graph, in the data distribution operation that directed graph is represented what transformation tasks is arranged, just what summits of increase in dependence graph;
(2) data of representing for directed graph are distributed any two limits in the operation, if there is public node on any two limits, and the transmission data partially overlap, then illustrate between the data transfer task of representing on these two limits to have dependence, its dependence is as follows: make T1=(Na, Nb, Data1), T2=(Nc, Nd, Data2), the data of T1 and T2 transmission have part to overlap, and promptly the common factor of Data1 and Data2 is not empty:
I) if Nb=Nc is same summit, T2 depends on T1 so, increases by one and point to the limit of Nt1 from Nt2 in dependence graph, be i.e. limit＜Nt2, Nt1 〉;
If ii) Nd=Na is same summit, T1 depends on T2 so, increases by one and point to the limit of Nt2 from Nt1 in dependence graph, be i.e. limit＜Nt1, Nt2 〉.
The method of redundant transmission is as follows in the described step b) cancel job:
(1) if T
Xy∈ T is empty transmission, deletes T so from T
Xy
(2) if T
Ba=(B, A, Data1) ∈ T, T
Ca=(Data2) ∈ T, and Data1 and Data2 have the data that partially overlap for C, A, keep one of them transmission so, with the data deletion that repeats in another transmission, if that selection is T
Ba, T so
CaConstant, T
BaMake (B, A, Data1-Data1 ∩ Data2) into;
(3) repeat above operation, till Job does not have redundant transmission.
Advantage of the present invention is: can represent intuitively that data are each internodal flowing; Can check the legitimacy of operation; Can be optimized operation.
Description of drawings
Fig. 1 is the data distribute work process flow diagram that the present invention is based on directed graph;
Fig. 2 is the data distribution job instances figure that the present invention is based on directed graph;
Fig. 3 is the transmission dependence schematic diagram of data distribution job instances shown in Figure 2.
Embodiment
In order to realize the internodal data distribution of N-to-M, and representation of data is internodal mobile at each intuitively, the data distributing method that representation of data intuitively provided by the invention flows between each node adopts directed graph, and (Directed Graph DG) comes data of description each internodal flowing.
As shown in Figure 1, provided by the invention based on directed graph (Directed Graph, data distributing method DG) may further comprise the steps:
The first step: with directed graph definition of data distribution operation
A) with directed graph definition of data distribution operation Job, promptly Job=(N, D, T), wherein N is the set of node, each node is represented the unit or the software systems that participate in exchanges data; D is the set of data element; T is the set on limit, and a data transfer task is represented on every limit, comprises the flow direction of data and data, and each transformation task T is that (Data), wherein Src ∈ N represents source node to a tlv triple for Src, Dest; Dest ∈ N, the expression destination node, Src and Dest constitute directed edge＜Src, Dest 〉, Data is the subclass of D, represents the data of this transmission.
B) will convert corresponding task description file to the data distribution operation of directed graph definition, such as the XML file; The usually corresponding description document of each operation.
Second step: carry out operation
A) task description file is submitted to the transmission Control Server, the transmission Control Server is checked the validity of operation.
Transmission at first checks operation after Control Server receives operation, judges between the transformation task of operation whether have dependence; Then, judge whether these dependences are legal,, have the operation of loop, illustrate that then these operations are illegal, can't be performed if these dependences have constituted ring; Otherwise,, illustrate that then these operations are legal, can be performed if there is not loop between these dependences.
B) redundant transmission in the cancel job.
Redundant transmission can reduce the efficiency of transmission of system, and brings that conflict waits some problems such as write data, and the transmission Control Server is by redundant technology for eliminating, and the redundant transmission in the cancel job improves the overall performance of operation.
C) transmission of decomposition coarseness makes each transformation task send littler data.
The transmission Control Server decomposes the coarseness transformation task, by increasing new transformation task, makes each transformation task send littler data, so just can improve the degree of parallelism of transformation task, and then improve efficiency of transmission.
D) carry out transmission, and carry out troubleshooting.
The transmission Control Server is realized data between nodes distribution, transmission by calling concrete transmission service such as GridFTP.The transmission Control Server can be found the fault in the operation implementation in explaining the process of carrying out the data distribution, the type of identification fault is also done corresponding processing.Concrete processing method comprises the deletion transformation task, repeats to transmit and select automatically new processing modes such as source node continuation transmission.
The 3rd step: carry out and finish, the transmission Control Server is notified the user with execution result.
Describe the data distributing method that the present invention is based on directed graph in detail below in conjunction with accompanying drawing.
As shown in Figure 2, the present invention is for representation of data is internodal mobile at each intuitively, adopt digraph description operation (being the distribution/transmission of data), each operation Job is defined as a directed graph, i.e. Job=(N, D, T,), wherein N is the set of node, each node is represented the unit or the software systems that participate in exchanges data; D is the set of data element; T is the set on limit, and a data transfer task is represented on every limit, comprises the flow direction of data and data.Fig. 2 is the example of data distribution, N={A wherein, B, C, D, E, F, G}, D={d1, d2, d3, d4, d5}, T={T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11}.Each transformation task T is that (Data), wherein Src ∈ N represents source node to a tlv triple for Src, Dest; Dest ∈ N, the expression destination node, Src and Dest constitute directed edge＜Src, Dest 〉; Data is the subclass of D, represents the data of this transmission.For example, ({ d2 is transferred to destination node B from source node A to T1=for d1, d2}) expression data d1 for A, B.T3=(B, D, d1, and d2, d3}) expression data d1, d2, d3 is transferred to destination node D from source node B.
Directly perceived by directed graph, vividly described data distribution operation after, convert the directed graph that generates to task description file such as XML file, description document of each Job correspondence; Then, task description file is submitted to the transmission Control Server, explain by this server and carry out.
After the transmission Control Server receives data distribution task description file, at first the validity of task description file is checked, whether whether judgment data distribution operation legal can the execution, promptly by whether there being dependence between the transformation task of judging operation, and whether these dependences have constituted ring, if have loop between some operation transformation task then these operations are illegal, can't be performed; The operation that does not have loop is legal, can be performed.Can find that from Fig. 2 T3 depends on T1, T7; T4 depends on T2, T8; T5 depends on T3, T10; Or the like.If all dependences of Fig. 2 are showed with diagram, so just can obtain transmission dependence graph as shown in Figure 3.In Fig. 3, each summit correspondence the limit of Fig. 2, just each summit correspondence of Fig. 3 a data transfer task, such as the corresponding T1 of Nt1, the corresponding T2 of Nt2, the Nt3 correspondence T3, Two dependences between the transmission are represented on every limit among Fig. 3, depend on T1 such as the T3 among Fig. 2, T7, and so corresponding Fig. 3 just has from Nt3, has pointed to Nt1 respectively, two limits of Nt7.Can find that from Fig. 3 T5, T10, T11 interdepend, constitute a circulation, so T5, T10, T11 are illegal.
The method of transmitting dependence graph GT from the data distribution operation Job structure of representing with directed graph is as follows:
1, for every limit T of the data distribution operation Job that represents with directed graph, in GT, increases a corresponding summit Nt.What transformation tasks are arranged among the Job, just in GT, increase what summits.
2, any two limits among the data distribution operation Job that represents for directed graph are T1 for example, T2, make T1=(Na, Nb, Data1), T2=(Nc, Nd, Data2), the data of T1 and T2 transmission have part to overlap, just the common factor of Data1 and Data2 is not empty, at this moment considers in two kinds of situation:
If a) Nb=Nc is same summit, T2 depends on T1 so, increases by one and point to the limit of Nt1 from Nt2 in GT, be i.e. limit＜Nt2, Nt1 〉;
B) if Nd=Na is same summit, T1 depends on T2 so, increases by one and point to the limit of Nt2 from Nt1 in GT, be i.e. limit＜Nt1, Nt2 〉.
Had after the transmission dependency graph, can utilize topological sorting (Topological Sort) algorithm to judge in the transmission dependency graph whether have loop, if there is loop, operation is illegal so, can not be performed; Otherwise operation is legal, can be performed.
After the transmission Control Server checked out the validity of operation, ensuing task was transmission redundant in the cancel job.So-called redundant transmission is meant transfer of data sky or that repeat that may exist in the operation, and for example the T6 among Fig. 2 does not transmit data, and T6 is exactly the transmission of a sky so; In addition, if there are two transmission to send identical data item to same summit, this also is a kind of of redundant transmission.Redundant transmission can reduce the efficiency of transmission of system, and brings that conflict waits some problems such as write data.Therefore need clear up the redundant transmission among the Job, the method for clearing up is as follows:
1. if T
Xy∈ T is empty transmission, deletes T so from T
Xy
2. if T
Ba=(B, A, Data1) ∈ T, T
Ea(Data2) ∈ T, and Data1 and Data2 have the data that partially overlap for C, A, keep one of them transmission so, with the data deletion that repeats in another transmission.If that select is T
Ba, T so
CaConstant, T
BaMake (B, A, Data1-Data1 ∩ Data2) into;
3. repeat above operation, till Job does not have redundant transmission.
After the transmission Control Server checks out transmission redundant in the validity, cancel job of operation, in order to improve data transmission efficiency, improve the degree of parallelism of transformation task, the transmission Control Server decomposes the coarseness transformation task by increasing new transformation task, makes each transformation task send littler data.For example among Fig. 2, T7 depends on T2, T2 such as has only could start T7 after being finished.If T2 resolve into T21=(A, C, d3}) and T22=(A, C, d4}), T7 will only depend on T21 so, as long as T21 is complete, this moment no matter the executing state of T22 how, T7 can start.Same T3 can resolve into T31=(B, D, d1}) and T32={B, D, { d2, d3}}.By such step, little transmission is resolved in big transmission, make each transformation task all transmit littler data, so just can improve the degree of parallelism between all transformation tasks of entire job, reduce the stand-by period of transformation task, and then total time of implementation of shortening entire job, the treatment effeciency of raising system.
After above step, the transmission Control Server begins really to carry out transformation task.The transmission Control Server is realized the data between nodes exchange, the way of similar RFT of this process or FTS by triggering concrete transmission service such as GridFTP.
In the middle of the process of transfer of data, certain transformation task may break down, and causes this bust this.At this moment, the transmission Control Server can be found the fault in the operation implementation, and the type of identification fault is also done corresponding processing.Concrete processing method comprises the deletion transformation task, repeats to transmit and select automatically new processing modes such as source node continuation transmission.No matter be which kind of processing mode, its final objective is that the former destination node of assurance can be obtained required data.In these three kinds of processing modes, be redirected relative complex.The T8 that supposes Fig. 2 has failed, and after T8 carries out failure, the way according to being redirected will need to obtain data from other nodes so.From Fig. 2, can find out intuitively, can not select D, because select D to repeat failure possibly as new data source; Can not select E as data source, because T4 depends on T8, when T8 did not finish, E did not have data, and if select E, T 4 and this new transmission will constitute loop so.A or B can be used as data source, and this moment, A and B should have data; F also can be used as data source, although this moment, F may also not have data.
Realize comparing with the design of other products in N-to-M Data dissemination process, method for interchanging data provided by the invention has following Some features:
1, with a data distribution of the formal description operation of digraph, each node represents the unit or the software systems that participate in exchanges data, and every limit represents a data transfer task (as shown in Figure 2). And grid workflow system is used the vertex representation activity when using DAG usually, represents sequence between the activity, parallel and choice relation with the limit. For the user, grid workflow system is not data-centered, but centered by transformation task, need the user to go to define transformation task and the relation between them, undoubtedly this definition mode is not directly perceived, can't directly embody data flowing between node, increase the difficulty of operational definition. The present invention adopts the operation of digraph description Data dissemination, can the Direct function description data flowing between node, thereby when describing operation, seem very directly perceived, convenient.
2, legitimacy that can Inspection especially can be judged the data dependence between the transmission easily.
3, can eliminate redundant transformation task, remove unnecessary transmission, save the processing time of operation; Can decompose big transformation task, improve degree of parallelism, thereby improved the overall performance of operation.
4, support automatic fault to process, can fail after the bust this and can adopt re-transmission, redirection mechanism, the raising transmission reliability after failing.
5, can be tied to different bottom host-host protocols, realize third party's control.
Claims (3)
1. data distributing method based on directed graph, this method may further comprise the steps:
The first step: with directed graph definition of data distribution operation;
A) with directed graph definition of data distribution operation Job, promptly (T), wherein N is the set of node to Job=for N, D, and each node is represented a unit that participates in exchanges data; D is the set of data element; T is the set on limit, and a data transfer task is represented on every limit, comprises the flow direction of data and data, and each transformation task T is that (Data), wherein Src ∈ N represents source node to a tlv triple for Src, Dest; Dest ∈ N, the expression destination node, Src and Dest constitute directed edge＜Src, Dest 〉, Data is the subclass of D, represents the data of this transmission;
B) will convert corresponding task description file to the data distribution operation of directed graph definition;
Second step: task description file is submitted to the transmission Control Server, carry out data distribution transmission operation;
A) task description file is submitted to the transmission Control Server, the transmission Control Server is checked the validity of operation;
Transmission at first checks operation after Control Server receives operation, judges between the transformation task of operation whether have dependence; Then, utilize topological sorting algorithm to judge whether dependence constitutes loop,, have the operation of loop, illustrate that then these operations are illegal, can't be performed if some dependence has constituted loop; Otherwise,, illustrate that then these operations are legal, can be performed if there is not loop between these dependences;
B) redundant transmission in the cancel job;
C) transmission of decomposition coarseness makes each transformation task send littler data;
D) actual figure transporting operation reportedly;
The 3rd step: the transmission Control Server is notified the user with execution result.
2. the data distributing method based on directed graph according to claim 1 is characterized in that: definite method of dependence is as follows between the transformation task in described second step in the step a):
Every limit T of the data distribution operation of (1) representing with directed graph is corresponding, increases a corresponding summit Nt in dependence graph, in the data distribution operation that directed graph is represented what transformation tasks is arranged, just what summits of increase in dependence graph;
(2) data of representing for directed graph are distributed any two limits in the operation, if there is public node on any two limits, and the transmission data partially overlap, then illustrate between the data transfer task of representing on these two limits to have dependence, its dependence is as follows: make T1=(Na, Nb, Data1), T2=(Nc, Nd, Data2), the data of T1 and T2 transmission have part to overlap, and promptly the common factor of Data1 and Data2 is not empty:
I) if Nb=Nc is same summit, T2 depends on T1 so, increases by one and point to the limit of Nt1 from Nt2 in dependence graph, be i.e. limit＜Nt2, Nt1 〉;
If ii) Nd=Na is same summit, T1 depends on T2 so, increases by one and point to the limit of Nt2 from Nt1 in dependence graph, be i.e. limit＜Nt1, Nt2 〉.
3. the data distributing method based on directed graph according to claim 1 is characterized in that: in described second step in the step b) cancel job method of redundant transmission as follows:
(1) if T
Xy∈ T is empty transmission, deletes T so from T
Xy
(2) if T
Ba=(B, A, Data1) ∈ T, T
Ca=(Data2) ∈ T, and Data1 and Data2 have the data that partially overlap for C, A, keep one of them transmission so, with the data deletion that repeats in another transmission, if that selection is T
Ba, T so
CaConstant, T
BaMake (B, A, Data1-Data1 ∩ Data2) into;
(3) repeat above operation, till Job does not have redundant transmission.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006101144975A CN101179516B (en) | 2006-11-10 | 2006-11-10 | Digraph based data distributing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006101144975A CN101179516B (en) | 2006-11-10 | 2006-11-10 | Digraph based data distributing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101179516A CN101179516A (en) | 2008-05-14 |
| CN101179516B true CN101179516B (en) | 2010-06-09 |
Family
ID=39405606
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2006101144975A Expired - Fee Related CN101179516B (en) | 2006-11-10 | 2006-11-10 | Digraph based data distributing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101179516B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102209369B (en) * | 2011-05-20 | 2014-03-19 | 北京航空航天大学 | Method based on wireless network interface selection to improve a smart phone user experience |
| CN102970245B (en) * | 2012-11-21 | 2016-10-05 | 北京奇虎科技有限公司 | Data transmission method and system |
| CN103440339B (en) * | 2013-09-09 | 2017-01-18 | 中国农业银行股份有限公司 | Method, device and system for searching application system change association objects |
| CN107395722A (en) * | 2017-07-21 | 2017-11-24 | 郑州云海信息技术有限公司 | A kind of group system event interaction processing method |
| CN108170717B (en) * | 2017-12-05 | 2020-12-04 | 东软集团股份有限公司 | Data exploration mode conversion method and device, storage medium and electronic equipment |
| CN109379441A (en) * | 2018-12-07 | 2019-02-22 | 华中科技大学 | Chain rule combined method and system are serviced in a kind of cloud environment |
| CN113569184A (en) * | 2021-07-16 | 2021-10-29 | 众安在线财产保险股份有限公司 | Configurable data calculation method, device, equipment and computer readable medium |
| CN116301735B (en) * | 2023-05-19 | 2023-07-21 | 华南理工大学 | Method, device and storage medium for organizing software elements into software data links |
| CN117348852B (en) * | 2023-12-05 | 2024-03-29 | 华南理工大学 | Method, device and medium for constructing data link based on fine-grained software elements |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5802289A (en) * | 1992-12-21 | 1998-09-01 | Apple Computer, Inc. | Method for propagating preemptive bus initialization on an acyclic directed graph |
| CN1781092A (en) * | 2003-03-17 | 2006-05-31 | 米特里昂尼克斯股份公司 | Data flow machine |
-
2006
- 2006-11-10 CN CN2006101144975A patent/CN101179516B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5802289A (en) * | 1992-12-21 | 1998-09-01 | Apple Computer, Inc. | Method for propagating preemptive bus initialization on an acyclic directed graph |
| CN1781092A (en) * | 2003-03-17 | 2006-05-31 | 米特里昂尼克斯股份公司 | Data flow machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101179516A (en) | 2008-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101179516B (en) | Digraph based data distributing method | |
| US10795737B2 (en) | Generic distributed processing for multi-agent systems | |
| CN101471885B (en) | Virtual multicast routing for a cluster having state synchronization | |
| CN101854338B (en) | Subscriber equipment, subscription management method thereof, and real-time communication method and system | |
| US9153001B2 (en) | Approach for managing distribution of automated demand response events in a multi-site enterprise | |
| US7346893B2 (en) | Exchange infrastructure system and method | |
| Guerraoui et al. | Dynamic byzantine reliable broadcast [technical report] | |
| CN100514314C (en) | Network equipment system | |
| CN104753817A (en) | Local analogy method and local analogy system for cloud computing message queue service | |
| CN102355413A (en) | Method and system for unifying message space on large scale in real time | |
| CN101917394A (en) | Middleware system for sharing data in mobile phone equipment and working method | |
| CN104808606A (en) | Method for providing functions within an industrial automation system and industrial automation system | |
| CN101933291A (en) | Initiation and expiration of objects in a knowledge based framework for a multi-master synchronization environment | |
| CN1973282A (en) | System and method for managing a change to a cluster configuration | |
| CN102884508A (en) | Hierarchically disassembling messages | |
| WO2013097363A1 (en) | Method and system for scheduling data sharing device | |
| Han et al. | CALM: an intelligent agent-based middleware for community computing | |
| Kühn | Reusable coordination components: reliable development of cooperative information systems | |
| AU2003223040B2 (en) | Exchange infrastructure system and method | |
| Prist et al. | Cyber-physical manufacturing systems: An architecture for sensor integration, production line simulation and cloud services | |
| Ghate et al. | Collaborative distributed communication in heterogeneous environments: A comprehensive survey | |
| Blanchet et al. | Supporting adaptive web-service orchestration with an agent conversation framework | |
| CN101163106A (en) | Method of executing combined service in wireless self-organizing network | |
| CN114331113A (en) | Method and system for managing industrial production based on micro-service architecture and process engine | |
| CN101801111A (en) | Acentric tuple space system facing cooperation |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100609 Termination date: 20121110 |