WO2017125015A1

WO2017125015A1 - Method for processing workflow of distributed system and workflow engine system

Info

Publication number: WO2017125015A1
Application number: PCT/CN2017/071524
Authority: WO
Inventors: 姚远
Original assignee: 中兴通讯股份有限公司
Priority date: 2016-01-18
Filing date: 2017-01-18
Publication date: 2017-07-27
Also published as: CN106980530A

Abstract

Provided are a method for processing a workflow of a distributed system and a workflow engine system. The method comprises: acquiring service information of a distributed system; arranging, according to the acquired service information, multiple workflows, each workflow comprising multiple workflow nodes in a directed acyclic graph; and regularly monitoring an execution state of each workflow node until completion of all workflows.

Description

Distributed system workflow processing method and workflow engine system

Technical field

The present disclosure relates to the field of computer communication technologies, such as a distributed system workflow processing method and a workflow engine system.

Background technique

The management process between nodes in a distributed system can interact with each other through messages, such as a process in which a user applies for a cloud hard disk in a cloud computing project, and the management main process sends a message to create a command to the storage management process, and the storage management process receives and executes the creation. After the execution of the command, the message to the management main process is returned to the management main process, and the management main process receives the response message of the successful execution, and then updates the user resource library, and the process ends. The application process of the above cloud hard disk can be understood as an administrator (such as a management main process) sending a command to an executor (such as a storage management process), the executor receives and executes the command, and replies a response message to the manager, however When implementing the application process with coding, there may be a variety of complex and uncontrollable factors that make the message delivery reliability between the manager and the performer low.

The above-mentioned distributed system with reliable and low message delivery can usually use the AB template to process the work process in the management process. The definition of the AB template can refer to the use of non-reliable message interaction between ABs, and A needs to pay attention to or manage B. The result of the current operation (ie: resource management), or the progress of the operation (ie: state management).

However, in a large-scale distributed system, the business process is complex, and is usually composed of multiple workflow nodes. If each workflow node adopts an AB template, the complexity of processing the business process through encoding is improved.

Summary of the invention

The present disclosure provides a distributed system workflow processing method and a workflow engine system, which can simplify the implementation of a complex system for processing complex business processes and reduce duplication of effort in encoding.

In a first aspect, the embodiment provides a distributed system workflow processing method, which may include: acquiring service information of a distributed system; and setting a plurality of workflows according to the acquired service information, where each workflow includes multiple directed Workflow nodes for acyclic graphs; and periodically monitor the execution status of each workflow node until all workflows are completed.

In a second aspect, the embodiment provides a distributed system workflow engine system, which may include: a service acquisition module configured to obtain service information of a distributed system; and a workflow setting module configured to set according to the acquired service information. Workflows, each workflow includes a plurality of directed acyclic graph workflow nodes; a workflow engine module configured to periodically monitor the execution status of each workflow node until all workflows are completed.

In a third aspect, the embodiment further provides a non-transitory computer readable storage medium storing computer executable instructions for executing the distributed system workflow processing method.

In a fourth aspect, the embodiment further provides a server, the server comprising one or more processors, a memory, and one or more programs, the one or more programs being stored in the memory when processed by one or more When the device is executed, the above distributed system workflow processing method is executed.

In a fifth aspect, the embodiment further provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions, when the program instruction When executed by a computer, the computer is caused to perform the distributed system workflow processing method described above.

A distributed system workflow processing method and a workflow engine system provided by the embodiment are configured to decompose a complex business process into a plurality of workflow nodes including a directed acyclic graph, and periodically monitor each The execution status of the workflow nodes until each workflow node completes the corresponding work, then it is determined that all the workflows are completed, thereby simplifying the implementation of the distributed system processing complex business processes and reducing the duplication of work during coding.

DRAWINGS

The drawings are used to provide a relevant understanding of the technical solutions of the present disclosure, and are used to explain the technical solutions of the present disclosure together with the present embodiments, and do not constitute a limitation of the technical solutions of the present disclosure.

FIG. 1 is a schematic flowchart diagram of Embodiment 1 of a distributed system workflow processing method according to an embodiment of the present disclosure.

FIG. 2 is a schematic diagram of a vertex representation active network of a second embodiment of a distributed system workflow processing method according to an embodiment of the present disclosure.

FIG. 3 is a schematic diagram of the workflow engine monitoring in the third embodiment of the distributed system workflow processing method according to the embodiment.

FIG. 4 is an application field of a fourth embodiment of a distributed system workflow processing method according to an embodiment of the present disclosure. Schematic diagram of the scene.

FIG. 5 is a schematic diagram of an application scenario of a fifth embodiment of a distributed system workflow processing method according to an embodiment of the present disclosure.

FIG. 6 is a schematic structural diagram of Embodiment 1 of a distributed system workflow engine system according to an embodiment of the present disclosure.

FIG. 7 is a schematic structural diagram of a hardware of a server according to an embodiment of the present disclosure.

detailed description

In order to make the objects, technical solutions and advantages of the present disclosure more apparent, the embodiments will be described in detail below with reference to the accompanying drawings. It should be noted that the embodiments in the present invention and the features in the embodiments may be arbitrarily combined with each other without conflict.

The steps illustrated in the flowchart of the figures may be executed in a computer system such as a set of computer executable instructions. Also, although logical sequences are shown in the flowcharts, in some cases the steps shown or described may be performed in a different order than the ones described herein.

The workflow (Workflow) involved in this embodiment may be an abstraction, generalization, and description of complex business processes and business rules between different operational steps in a distributed system. That is, the workflow can use the computer to automatically deliver documents, information, or tasks according to predetermined rules to achieve business goals.

The method in this embodiment can solve the problem that the business process is complicated in a large-scale distributed system. When each workflow node adopts an AB template, the technical problem of processing the business process by encoding is complicated.

The technical solution of the present disclosure will be described in detail below with an optional embodiment. The following alternative embodiments may be combined with each other, and the same or similar concepts or processes may be referred to each other in different embodiments.

FIG. 1 is a schematic flowchart diagram of Embodiment 1 of a distributed system workflow processing method according to the present disclosure. As shown in FIG. 1, the method includes S110-S130.

In S110, the service information of the distributed system is obtained.

Optionally, the distributed system may refer to a set of multiple servers, which can be used for communication and resource sharing, providing high computing performance and availability, thereby satisfying the requirements of the user for processing the service, and the local server obtaining the service information. The service information may include creating a cloud disk copy service, a clone image service, and other services.

In S120, multiple workflows are set according to the acquired service information, and each workflow includes multiple A workflow node with a directed acyclic graph.

Optionally, the complex business process is decomposed into multiple workflows according to the acquired business information, where the workflow may refer to logic and rules of how the work in the business information is organized together, and is appropriately The model performs representations and performs calculations on the logic and rules. In this embodiment, each workflow may include a plurality of directed acyclic graph-free workflow nodes. Wherein, the directed acyclic graph includes vertices and directional paths connecting the vertices.

2 is a schematic diagram of a vertex representation activity network of a second embodiment of a distributed system workflow processing method. As shown in FIG. 2, a vertex may be used to represent an activity, and a directed path may be used to represent a sequence relationship between activities. This directed acyclic graph is an Activity On Vertices (AOV) network. Illustratively, in the AOV network, if there is a directed path from vertex i to vertex j, and the directed path points from vertex i to vertex j, then vertex i can be called the sender of vertex j, or vertex j is the receiver of vertex i, from which it can be seen that Ci must be active before Cj. As shown in Figure 2, there is a directed path between C3 and C4, and the directed path points from vertex C3 to vertex C4, then C3 is active before C4. Similarly, there is also a relationship between C8 and C9. A directed path, and the directed path points from vertex C8 to vertex C9, then C8 is active before C9; there is also a directed path between C5 and C6, and the directed path points from vertex C5 to vertex C6, then C5 It is active before C6; there is also a directed path between C4 and C6, and the directed path points from vertex C4 to vertex C6, then C4 is active before C6, and so on.

In S130, the execution state of each workflow node is periodically monitored until all workflows are completed.

Optionally, in this embodiment, a separate workflow engine can be designed, and the workflow engine can be a separate process for monitoring the execution of all workflows. For example, in a workflow model based on a directed acyclic graph, the workflow engine periodically searches for a workflow node with a degree of entry 0 in the workflow, and urges the workflow node with the degree of entry 0 to perform work if it works. After the completion, the workflow node is taken out from the workflow model of the directed acyclic graph, and then the workflow node with the degree of entry 0 is continuously searched, and the workflow node with the degree of entry 0 is urged to perform the work, if the work is performed. After completion, the workflow node is taken out from the workflow model of the directed acyclic graph, and is cycled until the workflow node with the degree of entry equal to 0 can no longer be found in the workflow model. The workflow is complete. The degree of entry can refer to the sum of the number of times a certain vertex in the directed acyclic graph is the end point of the directed path in the graph, and the degree of entry 0 can mean that the node in the directed acyclic graph does not act as any directed path. The end point, that is, all directed paths connected to the node, uses this node as a starting point. FIG. 3 is a schematic diagram of the workflow engine monitoring in the third embodiment of the distributed system workflow processing method according to the embodiment, as shown in FIG. 3, in the first phase, The workflow engine detects the workflow node A and the workflow node G with a degree of entry 0, and urges the workflow node A and the workflow node G to perform corresponding work. After the workflow nodes A and G complete the corresponding work, the workflow nodes A and G are deleted from the workflow model of the directed acyclic graph. In the second phase, the workflow engine detects a workflow node B with an entry degree of 0, urging the workflow node B to perform the corresponding work. After the workflow node B completes the corresponding work, the workflow node B is deleted from the workflow model of the directed acyclic graph. In the third phase, the workflow engine detects workflow nodes C and D with an ingress of 0, urging workflow nodes C and D to perform the corresponding work. After the workflow nodes C and D complete the corresponding work, the workflow nodes C and D are deleted from the workflow model of the directed acyclic graph. In the fourth phase, the workflow engine detects the workflow nodes E and F with an ingress of 0, urging the workflow nodes E and F to perform the corresponding work. After the workflow nodes E and F complete the corresponding work, the workflow nodes E and F are deleted from the workflow model of the directed acyclic graph. At this point, the workflow engine can no longer find a workflow node with an entry degree of 0, and the entire workflow is determined to be complete.

A distributed system workflow processing method provided by the embodiment provides a plurality of workflows according to the acquired service information by acquiring service information of the distributed system, and each workflow includes multiple directed acyclic graphs. The flow node, and periodically monitors the execution status of each workflow node until each workflow node completes the corresponding work, and determines that all workflows are completed. The method is to decompose a complex business process into a workflow of a plurality of directed acyclic graph workflow nodes, and periodically monitor the execution status of each workflow node in the workflow network workflow until each work The flow node completes the corresponding work and determines that all workflows are completed. This simplifies the implementation of complex business processes for distributed systems and reduces duplication of effort in coding.

Optionally, in S120, the method may further include: respectively, according to the predetermined process of the acquired service information, numbering the workflow and the workflow node, and setting an execution sequence of the workflow node, and establishing a description A database of workflows and workflow nodes.

Optionally, the workflow and the workflow node are separately numbered according to the predetermined process of the obtained service information, where the predetermined process of the service information may generally refer to a resource that the service needs to acquire and related steps for implementing the effect, The number can be a letter or a number, but each number is not repeated, it is a unique identifier; and the order in which each workflow node performs work is arranged, and the time of execution work and its detailed data can be recorded in the storage system. Establishing a database describing the workflow and the workflow node, such as a workflow node as a sender, and an address, a parameter, an execution state, and an execution progress of the workflow node as the recipient corresponding to the sender, and the like The workflow engine can monitor each workflow node to perform work in order, so that it can distinguish which process the workflow node is triggered by, and can also determine the execution result of the previous workflow node, and the workflow node. After the execution is successful, the next workflow node is triggered, and thus all the workflows can be completed quickly and orderly.

Optionally, the database may include a workflow table, an execution table, and a sequence table.

The workflow table may include the number and execution status of all workflow nodes.

The execution table may include the number of the workflow node and the number, execution content, execution status, and execution progress of the associated workflow.

The sequence table may include coding and order relationships of workflow nodes that perform work.

Optionally, a plurality of databases may be designed to describe the AOV workflow, including adopting a Workflow table, an Action table, a Sequence table, etc., wherein all Workflow nodes may be saved in the Workflow table. Number and execution status, in the Action table, the number of the workflow node and the number, execution content, execution status, and execution progress of the workflow node may be saved, and the sequence and the order of the workflow node that performs the work may be saved in the Sequence table. Relationship, that is, through the description of the above database, defines "What should each workflow node do?", so that the workflow engine can query the database to know the upstream and downstream workflow nodes corresponding to each workflow node, and The request message is sent to the workflow node that needs to perform work until all workflows are monitored, which simplifies the coding process and makes it easier to implement these complex business processes.

At the same time, using the database to save the Workflow table, Action table and Sequence table, defines the entire process of the workflow, if there is a need to restart the workflow, whether it is in the sender, receiver or workflow engine, can be executed Restart the operation without affecting the final execution of the workflow.

Optionally, in S130, the method may further include: periodically scanning the workflow table according to a preset time period, and sequentially tracking each workflow node to perform work according to an order relationship between the workflow nodes in the sequence table. The state, and feedback the execution progress of the workflow node, urging the workflow point solution of the unfinished work according to the execution progress of the workflow node until each workflow node completes the corresponding work, and determines to complete all the workflows.

Optionally, a loop timer may be set, the workflow table is periodically scanned, a workflow node that has not been completed is found, and a command message is sent to the workflow node. After receiving the command message, the workflow node may want to cycle the timer reply. A response message or call the Finish Action function interface to inform the workflow engine that the workflow node has completed execution, the workflow engine will feedback the execution progress of the workflow node to the next workflow node, and the next workflow node will The execution state of the workflow node is saved in the Action table. If the response of the workflow node is not received, the workflow engine is Periodic monitoring is performed to prevent loss of messages. In addition, if the workflow node performs time-consuming operations such as resource allocation, the workflow node may first respond to the current execution status, for example, by using the progress percentage, thereby improving the user experience.

Optionally, the Action table may further include: a function name that processes success, failure, and timeout conditions.

Optionally, when the workflow node that is the sender sends a command to the workflow node that is the receiver, the command may be executed successfully, or may fail, or may expire after the command is issued, and the response may not be received. For example, if the storage resource is insufficient, the application for the cloud disk will fail, or the storage process may exit abnormally. The sender has not received the response message, and the above execution table also includes functions for processing success, failure, and timeout. The name can be used to perform the complex business process flow when the workflow node receives the success, failure, and timeout messages according to the corresponding function name in the execution table.

Each of the workflow nodes has an attribute as a sender or a receiver. The distributed system workflow processing method provided in this embodiment can be implemented when corresponding to different application scenarios.

FIG. 4 is a schematic diagram of an application scenario of a fourth embodiment of a distributed system workflow processing method according to an embodiment of the present invention. As shown in FIG. 4, in application scenario 1, when an AB template is used, when a workflow node is sent as a When it is the attribute of the receiver, the workflow directly defines the "who commands what to do", that is, the work that each workflow node needs to perform, and the workflow engine sends the request message directly to the responsible work. The workflow node, in order to prevent the loss of the message, also periodically tracks the workflow node until the workflow node responsible for performing the work is completed, wherein the workflow engine does not see the execution method of the workflow node. The execution method of the workflow node may be a third-party library function, or a program execution body such as sending a message to another process or a thread, or creating another workflow. The workflow can be ended after the execution of the workflow node responsible for performing the work.

FIG. 5 is a schematic diagram of an application scenario of a fifth embodiment of a distributed system workflow processing method according to an embodiment of the present invention. As shown in FIG. 5, in application scenario 2, when two workflow nodes are required as senders and receivers respectively, At the time of this, the workflow engine can act as an intermediary between the two workflow nodes. The workflow engine can first find the workflow node that performs the work, instead of sending the monitoring request as the sender node of the sender. The workflow node as the receiver sends the execution progress and the final result to the workflow engine. After the workflow node as the receiver answers the final result (success or failure), the workflow engine can forward the result to the sending. The workflow node, ultimately as the hair The sender's workflow node ends the workflow node to continue operation.

Therefore, the distributed system workflow processing method provided in this embodiment can implement complex service processing better, whether it is applied to the application scenario of the AB template or the complex application scenario 2, so that the message delivery is more reliable.

Optionally, FIG. 6 is a schematic structural diagram of Embodiment 1 of a distributed system workflow engine system according to the embodiment. As shown in FIG. 6, the system may include:

The service obtaining module 10 is configured to obtain service information of the distributed system.

The workflow setting module 20 is configured to set a plurality of workflows according to the acquired service information, where each workflow includes a plurality of directed acyclic graph workflow nodes;

The workflow engine module 30 is arranged to periodically monitor the execution state of each workflow node until all workflows are completed.

The distributed system workflow engine system provided by this embodiment may include: the service acquisition module acquires service information of the distributed system, and the workflow setting module sets multiple workflows according to the acquired service information, and each work The flow includes a plurality of directed acyclic graph workflow nodes, and the workflow engine module periodically monitors the execution status of each workflow node until all workflows are completed. The system is to decompose a complex business process into a plurality of workflow nodes including a directed acyclic graph, and periodically monitor the execution status of each workflow node until each workflow node completes the corresponding Work, determine that all workflows are completed, which simplifies the implementation of complex business processes for distributed systems and reduces duplication of effort in coding.

Optionally, the workflow setting module 20 may be further configured to respectively number the workflow and the workflow node according to a predetermined process of the acquired service information, and set an execution sequence of the workflow node, and establish a description. A database of workflows and workflow nodes.

The system provided in this embodiment can perform the foregoing method embodiments, and the implementation principle and technical effects are similar.

Optionally, the workflow engine module 30 may be further configured to periodically scan the workflow table according to a preset time period, and sequentially track the workflow node performing work according to the order relationship between the workflow nodes in the sequence table. Status, and feedback the execution progress of the workflow node, urging according to the execution progress of the workflow node until all workflows are completed.

Optionally, the execution table may further include: a function name that processes success, failure, and timeout conditions.

Optionally, the embodiment further provides a non-transitory computer readable storage medium storing computer executable instructions for executing any of the above distributed system workflow processing methods.

Optionally, FIG. 7 is a schematic structural diagram of a hardware of a server according to the embodiment. As shown in FIG. 7, the server includes:

A processor 410 and a memory 420; may also include a communications interface 430 and a bus 440.

The processor 410, the memory 420, and the communication interface 430 can complete communication with each other through the bus 440. Communication interface 430 can be used for information transmission. The processor 410 can call the logic instructions in the memory 420 to perform the distributed system work processing method of the above embodiment.

In addition, the logic instructions in the memory 420 described above may be implemented in the form of software functional units and sold or used as separate products, and may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the embodiment may be embodied in the form of a software product stored in a storage medium, including a plurality of instructions for causing a computer device (which may be a personal computer, a server, or The network device or the like) performs all or part of the steps of the method described in this embodiment. The foregoing storage medium may be a non-transitory storage medium, including: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. A medium that can store program code, or a transitory storage medium.

Finally, it should be understood that those skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by executing related hardware by a computer program, and the program can be stored in a non-transitory computer. In reading a storage medium, the program, when executed, may include a flow of an embodiment of the method described above, wherein the computer readable storage medium may be a magnetic disk, an optical disk, a read only memory (ROM), or a random access memory. (RAM), etc.

Industrial applicability

The present disclosure provides a distributed system workflow processing method and a workflow engine system, which can decompose a complex business process into workflows composed of a plurality of workflow nodes based on directed acyclic graphs, and periodically monitor work. The execution state of the workflow node in the flow network workflow until all the workflows are completed, which simplifies the implementation of the complex system processing complex business process and reduces the duplication of work during coding.

Claims

A distributed system workflow processing method, applied to a server, comprising:

Obtaining business information of the distributed system;

Setting a plurality of workflows according to the acquired business information, each workflow comprising a plurality of directed acyclic graph workflow nodes;

The execution status of each workflow node is periodically monitored until the plurality of workflows are completed.
The method of claim 1, after the setting a plurality of workflows according to the acquired service information, further comprising:

The workflow and the workflow node are respectively numbered according to the predetermined process of the acquired business information, and the execution order of the workflow node is set, and a database describing the workflow and the workflow node is established.
The method of claim 2, wherein the database comprises a workflow table, an execution table, and a sequence table;

The workflow table includes the number and execution status of all workflow nodes;

The execution table includes a number of the workflow node and a number of the workflow to which it belongs, an execution content, an execution status, and an execution progress;

The sequence table includes the encoding and ordering relationships of the workflow nodes that perform the work.
The method of claim 3 wherein said timing monitors an execution state of each workflow node until all workflows are completed, including:

Periodically scanning the workflow table according to a preset time period, sequentially tracking the execution status of each workflow node according to the order relationship between the workflow nodes in the sequence table, and feeding back the execution progress of each workflow node. Tracking based on the progress of the workflow node until all workflows are completed.
The method of claim 3 wherein said execution table further comprises: a function name for processing success, failure, and timeout conditions.
A distributed system workflow engine system comprising:

a service acquisition module, configured to obtain service information of the distributed system;

a workflow setting module, configured to set a plurality of workflows according to the acquired service information, where each workflow includes a plurality of directed acyclic graph workflow nodes;

The workflow engine module is configured to periodically monitor the execution status of each workflow node until all workflows are completed.
The system of claim 6 wherein said workflow setting module is further configured to:

After setting a plurality of workflows according to the acquired service information, the workflows and the workflow nodes are respectively numbered according to a predetermined process of the service information, and the execution order of the workflow nodes is set, and the description is performed. A database of streams and workflow nodes.
The system of claim 7 wherein said database comprises a workflow table, an execution table, and a sequence table;

The workflow table includes the number and execution status of all workflow nodes;

The execution table includes a number of the workflow node and a number of the workflow to which it belongs, an execution content, an execution status, and an execution progress;

The sequence table includes the encoding and ordering relationships of the workflow nodes that perform the work.
The system of claim 8 wherein

The workflow engine module is further configured to periodically scan the workflow table according to a preset time period, and sequentially track the status of each workflow node performing work according to the order relationship among the workflow nodes in the sequence table, and feed back The execution progress of each workflow node is tracked according to the execution progress of the workflow node until all workflows are completed.
The system of claim 8 wherein said execution table further comprises: a case function name for processing success, failure, and timeout.
A non-transitory computer readable storage medium storing computer executable instructions for performing the distributed system workflow processing method of any of claims 1-5.