Research And Implementation Of Supporting Workflow Distributed Execution And Dynamic Adjustment

In recent years,with the rapid development and application of advanced information technologies such as "cloud computing,artificial intelligence,big data,Iot and Internet" in all walks of life,there has been a trend that everything is service.A large number of services are interconnected,forming the service Internet.The service Internet has built complex service solutions in various ways and means such as crossdomain,cross-network,online and offline services,and provides various convenient services for a large number of users.The industry calls this phenomenon "big service".With the continuous development of information technology,workflow has become a key technology for aggregating services in various fields.Since services in different fields are independent of each other,the workflow in the service Internet has certain particularities: a workflow consists of several independent and autonomous processes.Service Fragments are assembled.The execution of this workflow in a large service environment often involves the invocation of services in different fields.Different process service segments need to be independently executed on different process engines,and the engines provided by each organization have their own Constraints,so only lightweight,low-code methods can be used to achieve control.In view of the above background and problems,this paper proposes a lightweight workflow distributed execution method.This paper focuses on the research on distributed execution management and control of workflow,focusing on realizing its characteristics of stability,high availability and high flexibility.Based on the design concept of "process control process",it makes full use of the control ability of the business process itself,and constructs the meta-process through process abstraction to realize the control of distributed workflow execution.As a special workflow,"meta-process" can carry enough information to cooperate with service agents to automate the distributed scheduling and execution of workflows,and data state migration between nodes.Compared with the existing workflow distributed execution method,this method has the following advantages: 1)Based on the non-invasive design concept,it has high portability and can be widely used in various process engines.2)Do not pursue an overly complex architecture,and indirectly achieve scheduling control through "metaprocess" to reduce the complexity of business logic.3)In view of the instability in the actual application environment,the dynamic adjustment of the workflow model is supported.Finally,this paper implements a workflow distributed execution management system based on this method.By simulating a large number of program instances and applying it to the system,it is confirmed that all cases can be executed correctly,and the feasibility of the method is verified.