TWI552079B - Intelligent multi - layer job assignment system and its - Google Patents

Description

Intelligent multi-layer work assignment system and method thereof

The invention relates to a smart multi-layer work assignment system and a method thereof, in particular to a system that can dynamically augment a dispatch level according to system load to accelerate the execution performance of the work, and can also dynamically converge the dispatch level and reduce the system. A smart multi-layer work assignment system and method for effectively controlling system performance.

Efficient use of system resources has always been an important design consideration for information system execution module design. A good execution module can not only effectively apply system resources, but also improve system service performance and maximize system output. On the contrary, if it is wrong The execution module design will cause a waste of system resources, resulting in poor system output performance, resulting in poor user evaluation of the system.

Large-scale information systems often use a decentralized architecture as the basic architecture for performance considerations, but the cost of building a decentralized architecture is often not affordable for a small project. The hardware cost of distributed architecture, software maintenance cost, software development cost, etc., are many obstacles that many small projects can't step into the decentralized architecture, which makes these projects encounter some work that requires large data operations. In addition to the execution of the back end, the schedule of the task must be defined in the system leisure time to ensure that the system operation is not affected. This strategy not only sacrifices the timeliness of the task, but also makes it due to the fixed schedule. The flexibility of the system is reduced.

In view of the above problems of the prior art, the object of the present invention is to provide a smart multi-layer work assignment system and method thereof, thereby improving the flexibility of scheduling tasks and effectively using limited resources to improve system output. .

In accordance with the purpose of the present invention, a smart multi-layer work assignment system is proposed for use in an information system. The intelligent multi-layer work assignment system includes a work management module, a multi-tree architecture module, and a dynamic amplification and convergence dispatch level module. Work management module The system receives a delivery job and parses out one of the execution parameters corresponding to the delivery work. The multi-tree architecture module receives the delivery work and execution parameters to construct a multi-tree structure corresponding to the delivery work according to the execution parameters, wherein the multi-tree structure is formed into a plurality of levels in a tree branch manner, and the next hierarchical level is At least one node in the previous hierarchy is formed by branching down to a plurality of nodes. The dynamic amplification and convergence dispatching level module performs the delivery work, and the dynamic amplification and convergence dispatching level module will selectively prioritize one of the hierarchical structures corresponding to the delivery work according to the load state of the information system. And the dynamic amplification and convergence dispatching level module periodically detects the load state of the information system to selectively amplify the processing to the next level or to the upper level until the nodes in all the levels are processed. .

Preferably, the intelligent multi-layer work assignment system of the present invention further comprises a database module for storing delivery work, execution parameters and multi-tree structure.

Preferably, when the load state of the dynamic amplification and convergence dispatch level module detection information system is less than a set lower limit value, at least one node in the current processing level is dynamically selected for amplification to perform amplification. Multiple branch nodes in a hierarchy below the selected node; if the detected load state is greater than a set upper limit, the dynamic amplification and convergence dispatch level module will converge some of the current processing levels or All nodes to the nodes of a hierarchy above them.

Preferably, the work management module parses the execution parameters according to the work type of the delivery work and stores them in the database module, and the work management module selectively transmits the delivery work to the plurality according to the load status of the information system. The tree architecture module enables the multi-tree architecture module to obtain execution parameters from the database module to construct a multi-tree structure for delivery work. .

Preferably, the load status of the information system includes memory usage and CPU usage.

According to the object of the present invention, a smart multi-layer work assignment method is further applied to an information system, the method comprising the steps of: receiving a delivery work via a work management module, and parsing one of the execution parameters corresponding to the delivery work Receiving delivery work and execution parameters by a multi-tree architecture module to construct a multi-tree structure corresponding to the delivery work according to the execution parameters, wherein the multi-tree structure is formed into a plurality of levels in a tree branch manner. The next level is formed by branching down at least one of the nodes in the previous level into a plurality of nodes; Using a dynamic amplification and convergence dispatching level module to selectively prioritize one of the multi-tree structures corresponding to the delivery work according to the load state of the information system, and the dynamic amplification and convergence dispatching hierarchical module periodicity The load state of the information system is detected to selectively augment the process to the next level or to the next level until the nodes in all the levels that have processed the delivery work are executed.

As described above, the intelligent multi-layer work assignment system and method thereof according to the present invention have one or more of the following advantages:

1. This intelligent multi-layer work assignment system and its methods ensure that the tasks delivered are indeed performed.

2. This intelligent multi-layer work assignment system and its method provide dynamic augmentation dispatch technology to improve the system capacity of the server, make full use of system resources, and accelerate the output of the work.

3. This intelligent multi-layer work assignment system and its method provide dynamic convergence dispatch technology to ensure the stability of the server. By controlling the level of execution work to ensure that the system load does not exceed the high value, it can ensure that all delivery work is completed. Tasks can be completed under effective control.

S11~S13, S1~S8‧‧‧ steps

101‧‧‧Work Management Module

102‧‧‧Multiple Tree Architecture Module

103‧‧‧Dynamic amplification and convergence dispatch level module

104‧‧‧Database Module

Job‧‧‧ delivery work

L0~LN‧‧‧ level

A~C, A1~A3, B1~B2, C1~C2, xx‧‧‧ nodes

1, 2A~2B, 3A~3C, 4A~4B, 5, 6x, 6y, 7~9‧‧‧ messages

Figure 1 is a flow chart of the smart multi-layer work assignment method of the present invention.

Figure 2 is a block diagram of the intelligent multi-layer work assignment system of the present invention.

Figure 3 is another flow chart of the smart multi-layer work assignment method of the present invention.

Figure 4 is a schematic diagram of an embodiment of a multi-tree architecture for constructing a multi-tree architecture of the present invention.

Figure 5 is a schematic diagram of an embodiment of the dynamic amplification and convergence dispatch level module of the present invention performing dynamic amplification or convergence.

Figure 6 is still another flow chart of the intelligent multi-layer work assignment system and method thereof.

The technical features, contents, and advantages of the present invention, as well as the advantages thereof, can be understood by the present inventors, and the present invention will be described in detail with reference to the accompanying drawings. The subject matter is for illustrative purposes only and is intended to be used in the specification. The true proportions and precise configurations after the implementation of the invention are not to be construed as limiting the scope and configuration of the accompanying drawings, and limiting the scope of the invention in the actual implementation.

Please refer to FIG. 1 , which is a flow chart of the smart multi-layer work assignment method of the present invention. The intelligent multi-layer work assignment method is implemented by the intelligent multi-layer work assignment system of the present invention, and is applicable to an information system, such as an accounting information system. The invention provides a good development module for a resource-limited information system, in particular for an information system that requires a large amount of data analysis or scheduling tasks, such as an accounting information system closing account and some related reports or decision analysis systems. The data analysis can use the invention to improve the output of the system, and also consider the stability and timeliness of the system, and effectively improve the system resource utilization rate. The smart multi-layer work assignment system of the present invention mainly includes a work management module 101, a multi-tree architecture module 102, a dynamic amplification and convergence dispatch level module 103, and a database module 104, as shown in FIG. Shown. The process steps of this smart multi-layer work assignment method are:

Step S11: Receive a delivery job via the work management module, and parse out one of the execution parameters corresponding to the delivery work. Among them, the work management module parses the execution parameters according to the type of work of the delivery work.

Step S12: Receive the delivery work and the execution parameters by the multi-tree architecture module to construct a multi-tree structure corresponding to the delivery work according to the execution parameters. The multi-tree structure is formed by a plurality of levels in a tree-like branch manner, and the next hierarchical level is formed by branching at least one node of the previous hierarchical level into a plurality of nodes.

Step S13: using the dynamic amplification and convergence dispatching level module to selectively prioritize one of the multi-tree structures corresponding to the delivery work according to the load state of the information system, and dynamically augment and converge the hierarchical module system. The load status of the information system is periodically detected to selectively augment processing to the next level or to the next level until the nodes in all levels that have processed the delivery are processed. Among them, the load status of the information system includes memory usage and CPU usage. When the load state of the dynamic amplification and convergence dispatch level module detection information system is less than a set lower limit value, at least one node in the current processing level is dynamically selected for amplification to amplify execution of the selected node. Multiple branch nodes in the next level; if the detected load state is greater than a set upper limit, the dynamic amplification and convergence dispatch level module will converge some or all of the nodes in the current processing level to Corresponding to a level Node.

In the above, the delivery work, the execution parameters, and the multi-tree structure are stored in the database module. In addition, the method further includes the steps of: selectively transmitting the delivery work to the multi-tree architecture module according to the load status of the information system by the work management module, so that the multi-tree architecture module is executed in the database module. The parameters are structured in a multi-tree structure for delivery work.

In the system architecture of the present invention, each delivery work is generated by a multi-tree building module to generate a multi-tree structure of its work before being executed, and the multi-tree has the characteristics of downward divergence and upward convergence. The feature dynamic amplification and convergence dispatching level module can control the number of threads according to the current information system load and the number of execution authorizations of the delivery work, and the nodes corresponding to each thread have independence due to the characteristics of the multi-tree architecture. , does not affect each other and ensures the integrity of the delivery work.

In detail, the system of the present invention is mainly composed of four modules, as shown in FIG. 2, the names and main functions are as follows:

Work Management Module 101: This module is responsible for storing the work stored by the user in the front-end system interface, and assigning the task to be executed to the multi-tree architecture module 102.

Multi-tree architecture module 102: This module will build a multi-tree architecture for delivery work based on the work category.

The dynamic augmentation and convergence dispatching level module 103: responsible for the execution of the delivery work, determines the specific level of the working multi-tree to be executed according to the system load, and dynamically detects whether to augment or converge the dispatch level during execution. To increase system output or reduce system load.

The database module 104 is responsible for storing relevant information required by the work management module 101, the multi-tree architecture module 102, and the dynamic amplification and convergence dispatch level module 103.

The invention utilizes a multi-tree structure as a method for cutting delivery work, and the multi-tree has a layered architecture, that is, the upper node has a set relationship with its child nodes, and this relationship serves as a basis for dynamic amplification and convergence dispatching levels, which ensures The integrity of the delivery work and the independence of the individual work thread processing data. After the current end user delivers the work performed by the plurality of backends, the work management module 101 stores the respective delivery work in the database module 104, and selects the work that can be executed in the waiting for delivery work list to be diversified. Tree architecture module 102, multi-tree architecture module Group 102 will deliver a multi-tree structure based on its work category. After the structural construction is completed, the dynamic amplification and convergence dispatching level module 103 takes over the execution of the delivery work. The module determines the multi-tree hierarchy that can be performed according to the current situation of the system, and performs dynamic amplification and convergence during execution. The dispatch level module 103 determines whether to augment or converge according to the information system load threshold until the delivery work is completed. Among them, the database module 104 plays the state of execution of the data temporary storage and delivery work.

Please refer to FIG. 3, which is another flow chart of the smart multi-layer work assignment method of the present invention, illustrating a detailed execution flow of a delivered work. First, after the user performs the execution of the front-end browser interface, the application server receives the request and transmits the work-related information to the work management module 101 for processing. The work management module 101 analyzes the delivery work according to the work category, and completes the work. After the parsing, the execution parameters input by the user are retrieved and stored in the database module 104. After the completion, the work management module 101 determines whether to deliver the delivery work according to the system status. If the system load is still executable, the execution flag of the delivery work is set to TRUE, and the delivery work is transmitted to the plurality of components. The tree architecture module 102; if the delivery cannot be performed under the system load, the execution flag is maintained at FALSE, waiting for the next execution judgment. After receiving the task of delivering the work execution, the multi-tree architecture module 102 obtains the relevant execution parameters of the delivery work from the database module 104, constructs the multi-tree structure of the work according to the parameters, and stores the multi-tree structure. To the database module 104. The dynamic augmentation and convergence dispatch level module 103 periodically obtains the delivery work to be performed in the database module 104, and then starts the execution of the work. Firstly, the dynamic amplification and convergence dispatch level module 103 obtains the multi-tree structure of the work, and determines the executable multi-tree hierarchy according to the current situation of the information system, and then dispatches several threads to start the parallel execution of the delivery work, and in a fixed period. Decide whether to augment this delivery function thread or to converge the dispatch level until the delivery is completed.

In the system of the present invention, the multi-tree architecture module 102 is responsible for the multi-tree architecture of the architecture delivery work. The multi-tree architecture module 102 first obtains the architecture factor of the delivery work by the database module 104, and the architecture factor has clear identification. The characteristics of data grouping, various types of delivery work will have different multi-trees due to different architectural factors. Referring to FIG. 4, there is shown a schematic diagram of an embodiment of the multi-tree architecture module of the present invention for constructing a multi-tree structure. In the figure, after a delivery job job passes through the multi-tree architecture module 102, a hierarchy of L1~LN will be generated. The multi-tree architecture, the data contained in each parent node in the structure will be equal to the sum of the data contained in the child nodes, and the nodes of the multi-tree will be used by the dynamic amplification and convergence dispatching level module 103. The augmentation and dispatching level module 103 determines the execution dispatch level of the delivery job job based on the information system load. The dispatch level refers to the L1~LN layer of the multi-tree structure, and the 1~N represents the dynamic level. When the first delivery work is performed, the dynamic amplification and convergence dispatch level module 103 will find the best. The dispatching level is executed, and during the execution process, the dynamic augmentation and convergence dispatching level module 103 periodically detects the usage rate of the information system, including the memory usage rate and the CPU usage rate, and if the usage rate exceeds the criticality. Value, dynamic amplification and convergence dispatch level module 103 will discard part of the thread and will converge the dispatch level to its parent level; if the system is found to be lighter, the number of threads can still be increased, dynamic amplification and The convergence dispatch level module 103 will dynamically select some nodes for the augmentation dispatch level to speed up the execution time of the delivery work. Referring to FIG. 5, it is a schematic diagram of an embodiment of performing dynamic amplification or convergence of a dynamic amplification and convergence dispatching level module of the present invention, illustrating an execution delivery work in a dynamic amplification and convergence dispatch level module. The life cycle performed by group 103. The S1 step represents the initial multi-tree architecture; in the S2 step phase, the dynamic amplification and convergence dispatch level module 103 selects the L2 layer as the dispatch level for performing the delivery work, after a period of detection, when the information system usage rate The SUR usage rate is low, and the dynamic amplification and convergence dispatch level module 103 starts the dynamic amplification dispatch level, as shown in steps S3, S4, and S5, which represents the continuous dynamic amplification of the dispatch level; When the dynamic amplification and convergence dispatch level module 103 detects that the system usage exceeds UH, the dispatch level will be converged as shown in step S6. When some nodes complete the execution, the data contained on the node has been processed. As shown in step S7, until all the nodes in execution are completed, the delivery is completed, as shown in step S8. Show.

In addition, please refer to FIG. 6, which is a flow chart of the smart multi-layer work assignment system and method thereof according to the present invention, which describes a delivery work after the user delivers to the system and executes the work management module. After the analysis, the process of starting the thread to the end of the dynamic amplification and convergence dispatch level module is started.

Step 1: The user sends a delivery work message 1 to the work management module 101 through the front-end device.

Step 2: The work management module 101 will deliver the work related information to the message 2A. The request database module 104 is stored. After the database module 104 finishes storing the delivery working parameters, the reply message 2B is stored for completion by the work management module 101.

Step 3: After obtaining the 2B message confirming that the delivery work has been completed, the work management module 101 requests the multi-tree architecture module 102 to perform the multi-tree structure for the delivery work. After receiving the multi-tree setup request, the multi-tree architecture module 102 sends a 3A message to the database module 104 to obtain the relevant execution parameters of the delivery work. The database module 104 obtains the data and then replies to the multi-tree architecture module through the 3B message. 102. The multi-tree architecture module 102 starts the multi-tree structure construction of the delivery work. After the construction is completed, the multi-tree structure is stored into the database module 104 by using the 3C message, and is set to wait for the execution state.

Step 4: The work management module 101 periodically uses the message 4A to query the database module 104 whether there is a waiting for execution to be performed. After receiving the 4A message, the database module 104 will wait for and perform the delivery work. The work management module 101 is replied to with 4B.

Step 5: After receiving the 4B executable delivery work message, the work management module 101 sends a message 5 to notify the dynamic augmentation and convergence dispatch level module 103 to start the execution of the delivery work, and the message 5 includes the multi-function tree of the delivery work. structure.

Step 6: The dynamic augmentation and convergence dispatching level module 103 determines the initial execution delivery level according to the system load after obtaining the multi-tree structure of the delivery work, and then obtains a sufficient number of executions from the thread pool through the 6x message. After a period of execution, the dynamic amplification and convergence dispatch level module 103 will decide whether to augment or converge the delivery level according to the system load. If the system load is found to be too low, and the delivery work can still be expanded downward, Then, the 6x message is sent again to the thread pool to obtain the still idle thread; if the system load is found to exceed the set maximum value, the dynamic amplification and convergence dispatch level module 103 will converge the delivery level of the work, through 6y. The message is returned to the thread pool until all multi-tree nodes have been executed.

Step 7: After the delivery work is completed, the dynamic amplification and convergence dispatch level module 103 will write the relevant information for completion of execution to the database module 104 via the message 7.

Step 8: After the delivery work is completed and the result is completed and stored in the database module 104, the dynamic augmentation and convergence dispatching level module 103 informs the work management module 101 by the message 8, and the delivery work has been completed, and the work management is required. Module 101 performs the next delivery work arrangement.

Step 9: After receiving the message 8, the work management module 101 will execute the result. Message 9 tells the user to complete the execution.

In summary, the system and method used in the present invention cuts the work by a hierarchical multi-tree, and can clearly distinguish the range of data contained in each node, and combines the dynamic amplification and convergence dispatching module, which can be based on the system load. To determine the level of dispatch, dynamic amplification or convergence of the dispatch level in the implementation process, this method is of significant help for online Analytical Processing or Data Warehousing type systems, especially in applications The system under the cloud architecture can greatly increase the system utilization rate and output, and also reduce the delay caused by system overload and mutual competition for resources.

The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of the invention are intended to be included in the scope of the appended claims.

S11~S13‧‧‧Steps

Claims (6)

A smart multi-layer work assignment system, which is applicable to an information system, comprising: a work management module, which receives a delivery work and parses out one of the execution parameters corresponding to the delivery work; a multi-tree architecture module The group receives the delivery work and the execution parameter to construct a multi-tree structure corresponding to the delivery work according to the execution parameter, wherein the multi-tree structure is formed in a tree branch manner to form a plurality of levels, the next level It is formed by branching at least one node in the previous hierarchy into a plurality of nodes; a dynamic amplification and convergence dispatching level module performs the delivery work, and the dynamic amplification and convergence dispatching level module will be based on The load state of the information system selectively prioritizes one of the levels of the multi-tree structure corresponding to the delivery work, and the dynamic amplification and convergence dispatch level module periodically detects the load status of the information system. To selectively amplify processing to the next level or to the next level until the nodes in all levels are processed; a database module And storing the delivery work, the execution parameter, and the multi-tree structure; and wherein when the dynamic amplification and convergence dispatch level module detects that the load state of the information system is less than a set lower limit, the current selection is dynamically selected Processing at least one node in the hierarchy for amplification to augment execution of a plurality of branch nodes in a hierarchy below the selected node; if the detected load state is greater than a set upper limit, the dynamic amplification The convergence dispatch level module will converge some or all of the nodes in the current processing hierarchy to the nodes above it. The smart multi-layer work assignment system according to claim 1, wherein the work management module parses the execution parameter according to the work type of the delivery work and stores the execution parameter into the database module, and the work The management module selectively transmits the delivery work to the multi-tree architecture module according to the load status of the information system, so that the multi-tree architecture module obtains the execution parameter from the database module to deliver the Work on the architecture of the multi-tree structure. The smart multi-layer work assignment system of claim 1, wherein the load status of the information system includes memory usage and CPU usage. A smart multi-layer work assignment method suitable for an information system, the method comprising the steps of: receiving a delivery work via a work management module, and parsing one of the delivery work Executing parameters; receiving, by the multi-tree architecture module, the delivery work and the execution parameter, to construct a multi-tree structure corresponding to the delivery work according to the execution parameter, wherein the multi-tree structure is a tree branch The method is formed by a plurality of levels, and the next level is formed by branching at least one node of the previous level into a plurality of nodes; using a dynamic amplification and convergence dispatching level module according to load state selectivity of the information system Priorityly processing one of the levels of the multi-tree structure corresponding to the delivery work, and the dynamic amplification and convergence dispatching level module periodically detects the load state of the information system for selective amplification processing Go to the next level or converge to the previous level until the nodes in all the levels that have processed the delivery work are executed; wherein the delivery work, the execution parameters, and the multi-tree structure are stored via a database module; and wherein the dynamic When the amplification and convergence dispatch level module detects that the load status of the information system is less than a set lower limit, the current processing will be dynamically selected. At least one node in the stage for performing amplification to amplify execution of a plurality of branch nodes in a lower level of the selected node; if the detected load state is greater than a set upper limit value, the dynamic amplification and The convergence dispatch level module will converge some or all of the nodes in the current processing hierarchy to the nodes above it. The smart multi-layer work assignment method of claim 4, further comprising the steps of: parsing the execution parameter according to the work type of the delivery work and storing the execution parameter to the database module; And selectively transmitting the delivery work to the multi-tree architecture module according to the load status of the information system, so that the multi-tree architecture module obtains the execution from the database module The parameters are structured to perform the multi-tree structure for the delivery work. . The smart multi-layer work assignment method according to claim 4, wherein the load status of the information system includes memory usage and CPU usage.