| With the rapid development of both Internet and enterprise information technologies, the application systems implemented in enterprises are always based on different platforms or tools, which makes the whole software system larger and more complex. Thus, how to perform effective service management in such distributed envrioments becomes an important issue to be tackled.SCA is known as an effective way to integrate the decentralized business functions by making use of common components. As one of the most mature open source SCA frameworks, Tuscany provides a lot of support on the standardization of service component and its calling conventions, which is widely used in business information technology systems. In this work, we choose the Tuscany container as the service management container. Based on this container, we design a service management and deployment platform including distributed SCA service unified configuration, deployment, and distant monitoring. Then, we consider the case that multiple server nodes are in a distributed environment, and each server contains different data and different services using Tuscany container. Focusing on the situation where large number of user requests occurs in a short time, we design a dynamic load balancing algorithm in order to efficiently organize the processing power of each node and allocate the large number of heterogeneous resources. This algorithm can maximize the throughput of the cluster and utilization of resources.Next, combining distributed concurrent environment of the Internet of Things, we present a detailed study of Tuscany of its service invocation mechanism and analyze problems of the existing instantiation mechanisms. Then we design instance pool model to manage the life cycle of multiple instances which is responsible for handling the generation, initialization, operation, dispatch and destruction of instances. Finally, we also do experiments to test the performance of the schema. The results show that it works well. |
PDF Full Text Request