Quality Of Service Of Component-based Embedded Software

As the development of embedded technology, the application of embedded system has become more wider, at the same time which also brings the growth requirement for embedded software increasingly. Embedded system has strict real-time, limited resources, etc., so the traditional software development technology obviously can't meet these requirements. In order to enhance the development efficiency and reduce the development complexity, cost and time, it is very suitable to employ the component-based software development (CBSD) of embedded technology into embedded system. Now component technology is the fashionable development method, which has the reusable, encapsulation, assemblable, etc., and can significantly improve software quality and development specialization. The paper researches two aspects of component assembly and evolution based QoS for embedded software system, expanses Lw CCM component model and container, and proposed a ELw CCM component modelAccording to the feature for software component development technology, from the user's actual demands of embedded software, research mainly includes the following aspects:(1) Current light weight CCM(Lw CCM) component model and the problems in embedded system application, as well as expanded Lw CCM. First, a QoS meta-model is given under the user's requirement and described part of QoS attributes with XML. Then, it expanded Lw CCM component model and container, proposing a ELw CCM embedded component model, including QoS reflective mechanism module, real-time QoS strategies module, QoS mate-model, QoS evaluation mechanism, QoS substitution mechanism, QoS Observer and QoS adaptive mechanism module. Last, it is discussed the ELw CCM component model characteristics, which effectively support the real-time and resource for embedded system, laying a theoretical foundation for component assembly and evolve.(2) The paper has a research on embedded component assembly based QoS. First, we analyze the component provider providing untust QoS value, according to direct and indirect trustiness, compute all the QoS trustiness which are as weights to revise qos value, and give the user requirement definition of QoS attributes and quantitative method. And then the paper analyzed the current research and deficiency of ant colony algorithm in composition optimization problem. We adopt multi parallel ant colony optimization realize component optimization, With multi parallel search feature, use adaptive information exchange between colonies, compute the weight of the current search shortest path, update local and overall pheromone, as well as describe the algorithm implementation. The method can effectively find a value of assembly component route.(3) The paper discuss QoS-driven component evolve in embedded systm. Fist, we evaluate the QoS attributes with fuzzy logic after composited component and find the evolving components. The paper proposes a model of evaluation and evolution. Then, the component state migrations are analyed. Last, component substitution accomplishes evolve process. Paper proposes a detailed component substitution mechanism, which the new component replaces the old one, QoS monitor checks the new composition component and updated embedded software system and guarantee the rationality of the substitution operation.(4) Take embedded audio order system as a example, through communication, coding, segmentation, synthesis and decoding of five different function component sets verify component assembly based multi-ant parallel colony algorithm, compared with the single ant algorithm. It uses the selection of communication functional components, successfully implemented component substitution, which the given replacement policy is feasible...