A Study On Service Governance Platform Based On Process Mining

With the rapid development of cloud computing, Service Oriented Architecture (SOA)has developed in depth. More enterprises have chosen SOA to develop their enterpriseapplications, because SOA can maintain the best consistency between the business processmodel and the applications. Service is the core of SOA. Therefore, integrated servicegovernance has become the key of applying the SOA to build enterprise applications.Service governance covers different phases of service life cycle and is mainly dividedinto the design-time service governance and run-time service governance. Design-time servicegovernance concerns about the service design phase. It cannot dynamically adjust theallocation strategies of service resources in time, according to the changes of businessrequirements in runtime. At the same time, run-time service governance concerns about theservice monitoring phase to monitor a single service. It is lack of the monitoring analysis ofthe relationship between services in the actual business process.In order to solve the problem in the existing service governance methods, which theycannot synchronously coordinate service resources in time based on the changes of thebusiness requirements and the business process model, one integrated service governancemethod and platform is put forward. It is based on the process mining technology andintegrates the business requirements, the business process model and the service executioninformation. It focuses on how to analyze the service execution information through theprocess mining technology then to get the changes of the business requirements and thebusiness process model. And then it synchronously coordinates service resources, guide thequality management of services, guide the cooperation between services and improve therelationship between service provides and service consumers according to the above analysisresults. The main research work is as follows:(1) Service Governance Platform based on Process Mining Platform framework includes data layer, mining layer and application layer from bottomto top. Data layer includes the service governance meta-model put forward by this paper,which is the data support layer for the platform to carry out the process mining. Mining layeris the core of the platform and responsible for applying the process mining techniques toanalyze the log file in control-flow view and case view to generate service governancestrategies integrating actual business requirements, business process and service executioninformation. Application layer provides interfaces for the platform roles to interact with theplatform.(2) Service Governance Meta-Model based on Process MiningAccording the research about the process mining and service governance, a servicegovernance meta-model is established covering the whole process of process mining, serviceexecution information and the business data in runtime. It is suitable for the process miningtechnology and one general model which can be extended according to specific scenarios.Service governance feedback sub-model formulates the corresponding service governancestrategy template and is the basis for the platform to get more targeted service governancestrategies for tenants and service.(3) Service Governance Method based on Process MiningBased on the case and control-flow view in process mining technology, one servicegovernance method is established. It describes the information mining and statistical process,and defines the data models generated in the process. It gives out the process of generatingservice governance strategies based on the data models and the service governancemeta-model.(4) Prototype of Service Governance Platform based on Process MiningBased on ProM Framework, the platform prototype is developed. Finally, a logisticsinformation platform of the transportation was used as a case study. And the servicegovernance method is compared with the existing service governance methods.The Platform prototype system implementation and verification results show that theplatform has a certain theoretical reference value and application prospect.