Research On Reliability Of Middleware Based On Components

With the swift and violent development of information technology, computer is applied more and more widely in different fields. And popularization and extension of the computer network impel distributed computing to expand constantly to various kinds of applications even more. It becomes an important problem for the development of distributed computing technology to satisfy the high reliability demand of various kinds of applications under network environment, especially critical applications. As an effective way, fault-tolerant support is offered to satisfy the high reliability demand of distributed applications, but it is very difficult to develop distributed applications directly, because if developed directly as so, not only numerous application logic should be considered to have the vested functions realized, but also complicated fault-tolerant logic should be taken into account for providing fault-tolerant support to applications. So it becomes the focus of attention to strip fault-tolerant logic out of various kinds of applications, set up unified fault-tolerant support platform,and then offer fault-tolerant support to applications. With the development of middleware and Web technology, three or multi-layer distributed applications become the mainstream of application development. It is middleware that solves the problem of interconnection for application software under the environment of different operating systems, helps to realize sharing of different heterogeneous information resources in network, strips out some special logic functions of web,communication management and distributed applications and make them integrated to form an effective support platform, offering unified support to distributed applications. In addition, application based on middleware is realized as the form of components. With the framework offered by middleware, application is developed by means of component. At the same time development based on component promotes the reasonable division of functions that middleware offered, gives middleware flexible expansion ability. In this way, it becomes an effective measure for solving above-mentioned problem to strip fault-tolerant logic out of various kinds of applications, make it finished by middleware, and then provide fault-tolerant support to distributed applications by fault-tolerant middleware with unified form. At first in this text correlative conceptions about component and middleware are introduced and analyzed, then a survey of software reliability was carried on. On this basis, combining with some up to date fault-tolerant technologies, a fault-tolerant middleware model FTBM is brought forward. Finally the system structure, relevant algorithms and all sorts of administrative technologies of FTBM are described in detail.