Research On Data Service Based On Service Integration Framework

Service-oriented architecture(SOA)provides a flexible service integration method,which has received increasing attention in dealing with the challenges associated with diverse environments.The Service Integration Framework is a DDS-based SOA implementation solution,which provides solutions for the service transformation of components and proposes methods for the management of services,so as to solves the problems involving in the reuse of application functions,the inefficient operation and maintenance of system,and the real-time communication of services.With the continuous expansion of the application scale in the Service Integration Framework system,the amount of heterogeneous business data generated in the systemis also increasing.Traditional data processing methods have been difficult to meet the data processing needs of the Service Integration Framework.Technologies on big data are urgently needed to help users perform data processing work.However,the Service Integration Framework does not propose a model of data-related services,and lacks the specification for how to build datarelated services based on existing DDS and RPC over DDS communication methods.In addition,the service integration framework lacks research on the internal mechanisms of big data processing so that no public interface about data processing has been extracted to simplify the coding work of service developement.In response to the above problems,a research about Data Service based on the Service Integration Framework is conducted,in which the achievements of big data are introduced to the Service Integration Framework,and the methods for data management are proposed.The main contributions of the thesis are:1)A model of Data Service is proposed.The Data Service Model is composed of two parts:data assets and data processing consisting of data mapper and interface adapter.The data assets,which are abstractions of the physical data sets,can be mapped to physical data sets by data mapper during the data processing phase.And the interface adapter can adapt the corresponding interfaces about big data processing for user requests to simplify the development process of data services.2)Methods for managing the data sets in SIF are proposed.By expanding the service registry center,the data management center is designed to manage the data sets in the system,in which a access control model about data sets is proposed to ensure the legality of data access,data service binding rules are formulated to select the most suitable data service for data service consumers,and data set storage rules are designed to make full use of the storage resources of each node of the system and reduce the data processing time.3)Methods of constructing data services based on the publish/subscribe and request/response communication mechanisms of are proposed.Combined with the two communication mechanisms of DDS and RPC over DDS provided by the Service Integration Framework,the design schemes of data collection service and data processing service are proposed respectively so as to be compatible with the original services of the service integration framework,which is convenient for data service consumers to initiate calls.4)A prototype system DSIF is designed and implemented.Common interfaces on data processing are developed in DSIF,and the implementation solutions for data management center and Data Service are also provided in DSIF by expanding the Service Integration Framework.Finally,the Data Service based on the Service Integration Framework is proved to be available by the function and performance tests of the prototype system.