CN102111451B - Reactor mode-based multi-task processing method - Google Patents
Reactor mode-based multi-task processing method Download PDFInfo
- Publication number
- CN102111451B CN102111451B CN201110050327.6A CN201110050327A CN102111451B CN 102111451 B CN102111451 B CN 102111451B CN 201110050327 A CN201110050327 A CN 201110050327A CN 102111451 B CN102111451 B CN 102111451B
- Authority
- CN
- China
- Prior art keywords
- task
- reactor
- event
- module
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention relates to a reactor mode-based distributed multi-task cooperative processing network system, which comprises a plurality of reactors used for realizing multi-path cooperative event separation and distribution processing, wherein the reactors are connected with one another by a network, and each reactor is provided with at least one independent central processing unit (CPU). The invention also relates to a method for realizing multi-task processing based on a reactor mode by utilizing the network system. In the method, after the reactors divide received task messages into various types of events, allocate the events to corresponding CPUs for processing, and feed task processing results back to the network according to event processing results returned by the CPUs. By the reactor mode-based distributed multi-task cooperative processing network system and the reactor mode-based distributed multi-task cooperative processing network method provided by the invention, a great calculated amount, high extendibility, higher cost performance for hardware cost and system performance, and calculation indexes which cannot be achieved by a conventional centralized system can be achieved.
Description
Technical field
The present invention relates to network system field, particularly task in enormous quantities is processed network system field, specifically refers to distributed multi-task cooperative processing network system and method based on reactor pattern.
Background technology
Along with informationization technology is more and more universal, existing centralized service solution can not adapt to the requirement that the public is more and more higher to informationization technology service.Traditional centralized service solution as shown in Figure 1, processed by thread A, B, C through decomposing by same task.On to the access of shared resource, need to lock contention, avoid data destroyed.When thread A, B and C need shared resource to access simultaneously, can only take lock access by one, all the other two in wait state; On monokaryon list cpu server, because the operation of three threads is all completed by a CPU, so the waste of resource is not obvious.But at present, multi-core CPU is more and more general, on the server of the many CPU of multinuclear, because thread A, B, C can have independently processor, adopt this centralized service solution will cause the free time waste of two other CPU computing capability.
Therefore, centralized service solution is in the past because the Bottleneck Restrictions of separate unit server hardware performance, current large-scale service is solved merely more and more difficult by centralized solution, and the existing multitask solution being based upon on monokaryon list CPU also can not adapt to new situation.
At present, this field need badly a kind of complicated for service function, operand large and possess the distributed multi-task concurrent cooperation solution of excellent extensibility, it can pass through rational network topology layout, reasonably service distribution formula is disposed, and makes the number of servers of its deployment and the service ability providing can have linear matching relationship.
Reactor is a kind of efficiently for event multichannel is separated and assignment provides extendible object-oriented framework.Reactor is in order to provide one group of senior programming abstract, and it has simplified the Design and implementation of event driven Distributed Application, also the multichannel separation of some different types of events is integrated in wieldy event handler.
Summary of the invention
The object of the invention is to have overcome above-mentioned shortcoming of the prior art, provide a kind of based on reactor pattern, controlled message route, controlled Task-decomposing and distributing, more complicated service function can be provided, operand is large, and possesses the distributed multi-task cooperative processing network system based on reactor pattern and the processing method of excellent extensibility.
In order to realize above-mentioned object, the distributed multi-task cooperative processing network system based on reactor pattern of the present invention has following formation:
This network system comprises that a plurality of each described reactor interconnects by network in order to realize the separated reactor with distributing processing of event of duplex cooperation, and each reactor all has at least one independently central processing unit CPU.
In this distributed multi-task cooperative processing network system, described reactor comprises that transceiver module, coupling routing module, task distribute module, event processing module, described transceiver module sequentially distributes module by described coupling routing module and task and is connected described event processing module, and described transceiver module is interconnection network also.
In this distributed multi-task cooperative processing network system, described event processing module comprises at least one event response processor.
In this distributed multi-task cooperative processing network system, the central processing unit CPU that described event response processor is corresponding single transaction.
In this distributed multi-task cooperative processing network system, described reactor is that arithmetic element, service escape unit, Task-decomposing unit, message routing unit or task converge unit, wherein, described arithmetic element is processed computing in order to the task message that this cooperative processing network system is received, and according to the routing iinformation comprising in task message, result is returned to described task and converge unit.
In this distributed multi-task cooperative processing network system, described service escape unit in order to task message escape that this cooperative processing network system is received for the task message that can be processed by this collaborative process network configuration and by forwarded to other reactor.
In this distributed multi-task cooperative processing network system, described Task-decomposing unit in order to the main task message receiving is resolved into each independently subtask message and by forwarded to other reactor.
In this distributed multi-task cooperative processing network system, described message routing unit in order to by subtask message according to the loading condition of each described arithmetic element, be routed in corresponding arithmetic element, and set up request message route and reply message route.
In this distributed multi-task cooperative processing network system, described task converges unit in order to the task result of returning is converged as service processing result.
The present invention also provides a kind of method of utilizing described network system to realize the multitasking based on reactor pattern, and the method comprises the following steps:
(1) described reactor receives task message from network;
(2) described reactor is divided into all kinds of events by the task message of receiving;
(3) reactor described in is assigned to corresponding CPU by each event and processes;
(4) described reactor obtains the transaction result that CPU returns;
(5) described reactor according to transaction result to network-feedback task result.
At this, realize in the method for the multitasking based on reactor pattern, described reactor comprises that transceiver module, coupling routing module, task distribute module, event processing module, described transceiver module sequentially distributes module by described coupling routing module and task and is connected described event processing module, described transceiver module is interconnection network also, and described step (2) specifically comprises the following steps:
(21) described coupling routing module is done synchronous processing by received mission bit stream;
(22) described coupling routing module matches mission bit stream and event, and information is sent to task distributes module.
At this, realize in the method for the multitasking based on reactor pattern, described step (3) specifically comprises the following steps:
(31) task distributes the event that module mates according to information and produces corresponding affairs;
(32) task distributes module and event is sent to described event processing module according to its corresponding affairs processes.
At this, realize in the method for the multitasking based on reactor pattern, described event processing module comprises at least one event response processor, and described step (32) is specially:
(32-1) task distributes module event is sent to its corresponding event response processor according to its corresponding affairs;
(32-2) described event response processor is processed received event;
(32-3) described event response processor is sent to task by result and distributes module.
At this, realize in the method for the multitasking based on reactor pattern, the CPU that described event response processor is corresponding single affairs, described step (32-2) also specifically comprises the following steps:
(32-2a) this event response processor is divided into each subevent by this event, and each subevent is sent to other event response processor;
(32-2b) other event response processor is processed each subevent, and returns to subevent result to this event response processor;
(32-2c) this event response processor is processed event according to obtained subevent result.
At this, realize in the method for the multitasking based on reactor pattern, described transceiver module is that transceiver module is processed in asynchronous unblock, and the step of a plurality of tasks (1) and step (5) run on described asynchronous unblock simultaneously and process transceiver module.
The distributed multi-task cooperative processing network system based on reactor pattern and the processing method of this invention have been adopted, because this network system comprises a plurality of in order to realize the separated reactor with distributing processing of event of duplex cooperation, each reactor interconnects by network, and all has at least one independently central processing unit CPU.When utilizing this network system to carry out multitask collaborative process, described reactor is divided into the task message of receiving after all kinds of events, be redistributed to corresponding CPU and process, the transaction result that reactor returns according to CPU is to network-feedback task result.Thereby can provide a kind of based on reactor pattern, there is controlled message route, controlled Task-decomposing and distributing, complicated service function can be provided, operand is large, and the distributed multi-task cooperative processing network system that possesses excellent extensibility, adopts this network system can in hardware cost and systematic function, obtain better cost performance, and can reach parameter that existing integrated system is beyond one's reach.
Accompanying drawing explanation
The schematic diagram of centralized service solution in Fig. 1 prior art.
Fig. 2 is the structural representation of the distributed multi-task cooperative processing network system based on reactor pattern of the present invention.
Fig. 3 is the structural representation of reactor in the distributed multi-task cooperative processing network system based on reactor pattern of the present invention.
Fig. 4 is the flow chart of steps that the network system described in utilization of the present invention realizes the method for the multitasking based on reactor pattern.
Fig. 5 utilizes network system of the present invention to realize the schematic diagram of first embodiment of method of the multitasking based on reactor pattern.
Fig. 6 utilizes network system of the present invention to realize the schematic diagram of second embodiment of method of the multitasking based on reactor pattern.
Fig. 7 utilizes network system of the present invention to realize the schematic diagram of the 3rd embodiment of method of the multitasking based on reactor pattern.
In Fig. 8 distributed multi-task cooperative processing network system based on reactor pattern of the present invention, utilize reactor as the structural representation of Task-decomposing and message route.
Embodiment
In order more clearly to understand technology contents of the present invention, especially exemplified by following examples, describe in detail.
Referring to shown in Fig. 2, is the structural representation of the distributed multi-task cooperative processing network system based on reactor pattern of the present invention.
In one embodiment, this network system comprises that a plurality of each described reactor interconnects by network in order to realize the separated reactor with distributing processing of event of duplex cooperation, and each reactor all has at least one independently central processing unit CPU.Wherein, described reactor as shown in Figure 3, comprise that transceiver module, coupling routing module, task distribute module, event processing module, described transceiver module sequentially distributes module by described coupling routing module and task and is connected described event processing module, and described transceiver module is interconnection network also.Described event processing module comprises at least one event response processor.The central processing unit CPU that this event response processor is corresponding single transaction.
A kind of preferred embodiment in, described reactor is that arithmetic element, service escape unit, Task-decomposing unit, message routing unit or task converge unit.
Wherein, described arithmetic element is processed computing in order to the task message that this cooperative processing network system is received, and according to the routing iinformation comprising in task message, result is returned to described task and converge unit.
Described service escape unit in order to task message escape that this cooperative processing network system is received for the task message that can be processed by this collaborative process network configuration and by forwarded to other reactor.
Described Task-decomposing unit in order to the main task message receiving is resolved into each independently subtask message and by forwarded to other reactor.
Described message routing unit in order to by subtask message according to the loading condition of each described arithmetic element, be routed in corresponding arithmetic element, and setting up described in request message route and reply message route of task converges unit in order to the task result of returning is converged as service processing result.
The present invention also provides a kind of method of utilizing described network system to realize the multitasking based on reactor pattern, and in one embodiment, the method as shown in Figure 4, comprises the following steps:
(1) described reactor receives task message from network;
(2) described reactor is divided into all kinds of events by the task message of receiving;
(3) reactor described in is assigned to corresponding CPU by each event and processes;
(4) described reactor obtains the transaction result that CPU returns;
(5) described reactor according to transaction result to network-feedback task result.
In a kind of more preferably execution mode, described reactor comprises that transceiver module, coupling routing module, task distribute module, event processing module, described transceiver module sequentially distributes module by described coupling routing module and task and is connected described event processing module, described transceiver module is interconnection network also, and described step (2) specifically comprises the following steps:
(21) described coupling routing module is done synchronous processing by received mission bit stream;
(22) described coupling routing module matches mission bit stream and event, and information is sent to task distributes module.
In a kind of further preferred embodiment, described step (3) specifically comprises the following steps:
(31) task distributes the event that module mates according to information and produces corresponding affairs;
(32) task distributes module and event is sent to described event processing module according to its corresponding affairs processes.
In a kind of preferred execution mode, described event processing module comprises at least one event response processor, and described step (32) is specially:
(32-1) task distributes module event is sent to its corresponding event response processor according to its corresponding affairs;
(32-2) described event response processor is processed received event;
(32-3) described event response processor is sent to task by result and distributes module.
A kind of further preferred embodiment in, the CPU that described event response processor is corresponding single affairs, described step (32-2) also specifically comprises the following steps:
(32-2a) this event response processor is divided into each subevent by this event, and each subevent is sent to other event response processor;
(32-2b) other event response processor is processed each subevent, and returns to subevent result to this event response processor;
(32-2c) this event response processor is processed event according to obtained subevent result.
In another preferred embodiment, described transceiver module is that transceiver module is processed in asynchronous unblock, and the step of a plurality of tasks (1) and step (5) run on described asynchronous unblock simultaneously and process transceiver module.
In actual applications, distributed multi-task cooperative processing network system based on reactor pattern of the present invention adopts existing blade server and Ethernet topology to dispose, adopt multi-modal distributed multi-task cooperation solution to build the hardware platform of large-scale service system, and adopt the arithmetic elements of language development based on reactor pattern such as C++ or JAVA.
While utilizing the distributed multi-task cooperative processing network system based on reactor pattern of the present invention, as shown in Figure 2, same task is distributed to corresponding reactor processing through the form of decomposing with message, on each reactor, register one or more event handlers, a class or multiclass message are processed; Due to not to the fighting for of shared resource, and each reactor has independently CPU, the assignment of multitask and carry out the computing capability that can effectively utilize CPU.Complete when a plurality of tasks carryings, message is all returned on the reactor that distributes at first task, and tasks carrying finishes.Each reactor can move the subtask of a plurality of multiclass simultaneously, and the failure of single reactor subtasking or abnormal, can not cause serious impact to the execution performance of overall task.
As shown in Figure 3, the basic composition of reactor comprises: transceiver module, coupling routing module, task distribute module, event handler module.Transceiver module receipt message and transmission message, asynchronous unblock is processed, and can receive and dispatch multiple messages simultaneously.Coupling routing module is done synchronous processing by the message of asynchronous receiving-transmitting, by message and event matches, is routed to corresponding " task " upper, and the response message of task result is matched on other reactor.Task distributes module to the event establishment affairs that produce, and affairs is distributed on corresponding processor and is carried out, repeatedly messaging generation event and repeatedly scheduling events processor in affairs.Event handler module is for processing the minimum logical block of subtask, and event handler is only responsible for response events, carries out calculation task, not responsible transaction management.In logic, a reactor under physical environment can be arranged to the virtual event handler in another reactor.
As shown in Figure 5, the distributed multi-task cooperative processing network system based on reactor pattern of the present invention is used to realize the embodiment of a cluster of mathematical operation, and it shows the actual motion effect of distributed multi-task cooperation solution.Portal server externally provides interface service, receives the calculation task of " 4+(5 * 6)-(18 ÷ 3) ", first divides and tasks " adding " arithmetic element (arithmetic element a: reactor of physics actual deployment); " add " arithmetic element and first " 5 * 6 " multiplication subtask is given to the processing of " taking advantage of " arithmetic element, " taking advantage of " arithmetic element synchronously returns to 30; " add " arithmetic element and calculate 4+30=34, then " 34-(18 ÷ 3) " subtraction subtask is given to " subtracting " arithmetic element and processed, " subtracting " arithmetic element is handed over the processing of " removing " arithmetic element by " 18 ÷ 3 " division subtask again; " remove " arithmetic element and return results 6, " subtracting " arithmetic element is calculated 34-6=28, returns; " add " arithmetic element final calculation result 28 returned to portal server, portal server can by result of calculation by request language conversion and return to service requester.
As shown in Figure 6, it is and an embodiment of computing request same shown in Fig. 5 that this example has been shown another effect of " mathematical operation cluster " task resolution." add " arithmetic element subtraction subtask " 5 * 6-18 ÷ 3 " friendship " is subtracted " to arithmetic element processing, " subtracting " arithmetic element transfers to respectively " taking advantage of " and " removing " arithmetic element to process " 5 * 6 " and " 18 ÷ 3 " subtask; " subtract " arithmetic element and return results and calculate 30-6=24 according to two unit, and return to " adding " arithmetic element; " add " arithmetic element and calculate 4+24=28, and return results to portal server.
This embodiment compares with Fig. 5, Task-decomposing with distribute order different, " subtracting " arithmetic element, the process that " taking advantage of " and " removing " arithmetic element is processed is transferred to respectively in " 5 * 6 " and " 18 ÷ 3 " subtask is to carry out in synchronization, so compared to Fig. 5, the computing of Fig. 6 embodiment has been lacked a message and has been sent a reply the cycle, has namely saved the running time of general assignment.
As shown in Figure 7, this example has illustrated an application that recursive function solves, and can find out that " F function " and " adding " arithmetic element bear the processor active task of self separately, by subtask, decomposes and message communication, has completed distributed computing; When needs portal server provides responsible calculation function external, can well reuse existing arithmetic element like this.
By above three embodiment, can find out distributed multi-task cooperation solution, compare multitask solution in the past, there is good extensibility, in performance, there is the huge space promoting, when the cpu load of " adding " arithmetic element arrives hardware bottleneck, can in cluster, several " adding " arithmetic elements of multi-section administration carry out load.
In the present invention, divided by distributed reaction heap pattern multitask cooperation scheme, solve complicated huge task, in scheme implementation details, also exist different from scheme of the prior art.
In general scheme, after subtask sends and processed by other arithmetic elements, Shi Yuan returns on road, and as shown in Figure 5 and Figure 6, request and response message are to occur in pairs; And under the actual scene of reality, exist the situation that subtask result does not need former road to return, for example: subtask request is Areq → Breq → Creq, it can be Cres → Bres → Ares that subtask result is returned, and can be also Cres → Ares.
In distributed multi-task cooperation solution, the decomposition of task and message route are the keys that promotes whole cluster performance.As shown in Figure 5 and Figure 6, same calculation task, can have two kinds of Task-decomposing modes, carries out route, two different execution performances for two kinds.And in the embodiment shown in fig. 7, if n is a very large number, between " F function " and " adding " arithmetic element, have so the evitable interacting message of a lot of bases, and in Fig. 7 embodiment, can select better Task-decomposing scheme, improve whole runnability.
In order further to promote cluster overall performance, as shown in Figure 8, in cluster, each reactor, except defining before arithmetic element role, also can increase and be defined as service escape unit, Task-decomposing unit, message routing unit, task and converge unit etc.Message communication is to each other carried out on network link layer, can select the agreements such as TCP, can add the proprietary protocol information such as message serial number, task code clerk, affairs code clerk, to meet the requirement of cluster multitask load in message.By service escape unit, by extraneous service request escape, be accessible task requests in cluster; Task-decomposing unit, according to built-in syntax rule, resolves into each independently subtask by main task, and this is olation, according to the difference of task type, can produce different subtask set; Message routing unit is responsible for subtask message, according to each arithmetic element loading condition in cluster, being routed in corresponding arithmetic element, and sets up correct request message route and reply message route; Task converges the task result of different time points, nonidentity operation unit being returned unit and converges, termination affairs, and return service result is to portal server; Arithmetic element is independently minimum subtask processing unit, can not rely on other unit and independently dispose, and the routing iinformation according in request message, returns to result.
The distributed multi-task cooperative processing network system based on reactor pattern and the processing method of this invention have been adopted, because this network system comprises a plurality of in order to realize the separated reactor with distributing processing of event of duplex cooperation, each reactor interconnects by network, and all has at least one independently central processing unit CPU.When utilizing this network system to carry out multitask collaborative process, described reactor is divided into the task message of receiving after all kinds of events, be redistributed to corresponding CPU and process, the transaction result that reactor returns according to CPU is to network-feedback task result.Thereby can provide a kind of based on reactor pattern, there is controlled message route, controlled Task-decomposing and distributing, complicated service function can be provided, operand is large, and the distributed multi-task cooperative processing network system that possesses excellent extensibility, adopts this network system can in hardware cost and systematic function, obtain better cost performance, and can reach parameter that existing integrated system is beyond one's reach.
In this specification, the present invention is described with reference to its specific embodiment.But, still can make various modifications and conversion obviously and not deviate from the spirit and scope of the present invention.Therefore, specification and accompanying drawing are regarded in an illustrative, rather than a restrictive.
Claims (4)
1. the utilization distributed multi-task cooperative processing network system based on reactor pattern is realized the method for the multitasking based on reactor pattern, described network system comprises a plurality of in order to realize the separated reactor with distributing processing of event of duplex cooperation, each described reactor interconnects by network, each reactor all has at least one independently central processing unit CPU, described reactor comprises transceiver module, coupling routing module, task distributes module, event processing module, described transceiver module sequentially distributes module by described coupling routing module and task and is connected described event processing module, described transceiver module is interconnection network also, it is characterized in that, described method comprises the following steps:
(1) described reactor receives task message from network;
(2) described reactor is divided into all kinds of events by the task message of receiving, specifically comprises the following steps:
(21) described coupling routing module is done synchronous processing by received mission bit stream;
(22) described coupling routing module matches mission bit stream and event, and information is sent to task distributes module;
(3) reactor described in is assigned to corresponding CPU by each event and processes;
(4) described reactor obtains the transaction result that CPU returns;
(5) described reactor according to transaction result to network-feedback task result;
Wherein, described transceiver module is that transceiver module is processed in asynchronous unblock, and the step of a plurality of tasks (1) and step (5) run on described asynchronous unblock simultaneously and process transceiver module.
2. the distributed multi-task cooperative processing network system of utilization according to claim 1 based on reactor pattern realized the method for the multitasking based on reactor pattern, it is characterized in that, described step (3) specifically comprises the following steps:
(31) task distributes the event that module mates according to information and produces corresponding affairs;
(32) task distributes module and event is sent to described event processing module according to its corresponding affairs processes.
3. the distributed multi-task cooperative processing network system of utilization according to claim 2 based on reactor pattern realized the method for the multitasking based on reactor pattern, it is characterized in that, described event processing module comprises at least one event response processor, and described step (32) is specially:
(32-1) task distributes module event is sent to its corresponding event response processor according to its corresponding affairs;
(32-2) described event response processor is processed received event;
(32-3) described event response processor is sent to task by result and distributes module.
4. the distributed multi-task cooperative processing network system of utilization according to claim 3 based on reactor pattern realized the method for the multitasking based on reactor pattern, it is characterized in that, the CPU that described event response processor is corresponding single affairs, described step (32-2) also specifically comprises the following steps:
(32-2a) this event response processor is divided into each subevent by this event, and each subevent is sent to other event response processor;
(32-2b) other event response processor is processed each subevent, and returns to subevent result to this event response processor;
(32-2c) this event response processor is processed event according to obtained subevent result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110050327.6A CN102111451B (en) | 2011-03-02 | 2011-03-02 | Reactor mode-based multi-task processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110050327.6A CN102111451B (en) | 2011-03-02 | 2011-03-02 | Reactor mode-based multi-task processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102111451A CN102111451A (en) | 2011-06-29 |
| CN102111451B true CN102111451B (en) | 2014-03-19 |
Family
ID=44175477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110050327.6A Active CN102111451B (en) | 2011-03-02 | 2011-03-02 | Reactor mode-based multi-task processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102111451B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102523169A (en) * | 2011-12-31 | 2012-06-27 | 南京大学 | Parallelization method based on 2D-Mesh multi-core network architecture |
| CN103577263A (en) * | 2013-11-07 | 2014-02-12 | 广东电网公司佛山供电局 | Power quality data real-time collection method and device |
| CN104580367B (en) * | 2014-12-08 | 2018-03-20 | 北京锐安科技有限公司 | A kind of network data synchronizing method and system based on reactor patterns |
| CN107580197B (en) * | 2017-08-15 | 2021-03-16 | 移康智能科技（上海）股份有限公司 | Task processing method, electronic device and computer readable storage medium |
| CN111124693A (en) * | 2018-11-01 | 2020-05-08 | 烟台翼顺电子科技有限公司 | YSTEC-CS1 multitask instant synchronous processing system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5724584A (en) * | 1994-02-28 | 1998-03-03 | Teleflex Information Systems, Inc. | Method and apparatus for processing discrete billing events |
| CN1602467A (en) * | 2001-12-14 | 2005-03-30 | 皇家飞利浦电子股份有限公司 | Data processing system having multiple processors, a task scheduler for a data processing system having multiple processors and a corresponding method for task scheduling |
| CN101667165A (en) * | 2009-09-28 | 2010-03-10 | 中国电力科学研究院 | Bus sharing method and device for distributed multi-master CPUs |
| CN101741850A (en) * | 2009-12-25 | 2010-06-16 | 北京邮电大学 | Multitask concurrent executive system and method for hybrid network service |
-
2011
- 2011-03-02 CN CN201110050327.6A patent/CN102111451B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5724584A (en) * | 1994-02-28 | 1998-03-03 | Teleflex Information Systems, Inc. | Method and apparatus for processing discrete billing events |
| CN1602467A (en) * | 2001-12-14 | 2005-03-30 | 皇家飞利浦电子股份有限公司 | Data processing system having multiple processors, a task scheduler for a data processing system having multiple processors and a corresponding method for task scheduling |
| CN101667165A (en) * | 2009-09-28 | 2010-03-10 | 中国电力科学研究院 | Bus sharing method and device for distributed multi-master CPUs |
| CN101741850A (en) * | 2009-12-25 | 2010-06-16 | 北京邮电大学 | Multitask concurrent executive system and method for hybrid network service |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102111451A (en) | 2011-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11934883B2 (en) | Computer cluster arrangement for processing a computation task and method for operation thereof | |
| CN102111451B (en) | Reactor mode-based multi-task processing method | |
| CN107566153B (en) | Self-management micro-service implementation method | |
| CN110278249A (en) | A kind of distribution group intelligence system | |
| CN109791509A (en) | High performance computing system and method | |
| CN104239555B (en) | Parallel data mining system and its implementation based on MPP | |
| CN103777950B (en) | Gridding method for resolving AOS (Advanced Orbiting System) telemetering data | |
| JP2011516998A (en) | Workload scheduling method, system, and computer program | |
| CN110308987B (en) | Method for updating connection parameters of distributed training tasks on container cloud | |
| CN104615489A (en) | Multi-node data interaction achieving method | |
| CN106293955A (en) | A kind of communication means between application process | |
| CN105786447A (en) | Method and apparatus for processing data by server and server | |
| CN106161520A (en) | Big market demand platform and exchange method based on it | |
| Behl et al. | Scalable {BFT} for {Multi-Cores}:{Actor-Based} Decomposition and {Consensus-Oriented} Parallelization | |
| CN101304410A (en) | Intelligent information platform for distributed WEB | |
| CN115562846A (en) | Resource scheduling method and device and computing node | |
| Rossant et al. | Playdoh: a lightweight Python library for distributed computing and optimisation | |
| CN107872527B (en) | LVC integrated remote cloud service system and method | |
| Czarnul | A multithreaded CUDA and OpenMP based power‐aware programming framework for multi‐node GPU systems | |
| CN106844021B (en) | Computing environment resource management system and management method thereof | |
| Yang et al. | Building a pervasive SOA based IOT communication middleware using runtime compilation and reflection | |
| US20120198470A1 (en) | Compact node ordered application placement in a multiprocessor computer | |
| Li et al. | Researches on performance optimization of distributed integrated system based on mobile agent | |
| Ng et al. | Framework for developing distributed systems in a peer-to-peer environment | |
| CN111767135A (en) | Data distribution method based on SRIO global memory sharing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |