Real-time Dynamic Scheduling Services In Distributed Systems - Design And Implementation Of The Corba Environment

The coexistence of heterogeneous entities in the large scale distributed systems makes it more and more difficult to develop distributed applications. CORBA provides the ability of interoperability among those entities. In recent years more and more real-time applications bring into distributed system, to guarantee their real-time requirements, it is necessary to make ensure the real-time performances of operating systems, application layer, underlying networks and middlewares in the system. For middlewares, not only need be changed to a real-time one (e.g. Real-Time CORBA), but also need import Dynamic Scheduling Service. This thesis does some research on the Real-Time CORBA Dynamic Scheduling Service from theoretical and practical point of view.Firstly, OMG's CORBA specification and the architecture of CORBA are described, and then the features of Real-Time CORBA are discussed in detail. The knowledge of real-time scheduling disciplines and the scheduling algorithms used in real-time CORBA environment are introduced. Real-Time CORBA dynamic scheduling service is not the scheduling service that must use the dynamic scheduling disciplines (e.g. EDF, LLF), it is the one whose underlying infrastructure must be able to satisfy real-time requirements in a dynamically changing environment. After analyzing the limitation of static scheduling service, we give a detailed declaration to dynamic scheduling service, and also describe the syntax and semantics of the various interfaces specified by real-time CORBA2.0 dynamic scheduling service.Based on the above knowledge, we put forward a dynamic scheduling service framework, which includes the scheduling module and dispatching module. Their co-operation makes resource allocation process of the whole real-time CORBA system successful accomplishment.Through systemic tests, the result shows this dynamic scheduling service is capable enough to meet the application's real-time requirements, and also can adapt to the dynamic change of the application environment.