|
It has hindered the development of a computer system to cope with software complexity crisis, in 2001, IBM proposed a "autonomic computing" concept, whose goal is to enable the computer system to have the capabilities of self-management.First of all, autonomic computing framework for C/S-architectured distributed application systems and software model are given on the basis of research results in related fields both at home and abroad, and its work mechanism in more detail is given. Secondly, for the difficult problems of software updates and adjustments existing in a C/S-architectured distributed application system, combined with the continued consistency of distributed systems , an optimized automatic update method for client software is presented, and some of the key issues encountered in the automatic update,such as selection of updating sources , consistency, a measure of the Consistency of Demand Intensity, the updated timing issues,are given,and also solution strategies to deal with them are gvien.Thirdly, the concepts of autonomous computing system model are described,and strategies to implement autonomic computing properties based on this model are given, such as achieving self-healing by the state rollback, modifying BUG automatically combined with knowledge base to achieve the code-level self-healing, and monitoring the system operating environment dynamically combined with knowledge base to achieve self-optimization.Finally, on the basis of "Network-based Social-security Audit System"(N-SAS) of China’s Golden Auditing Project of National Audit Office, a autonomic computing system is built in a real C/S-architectured distributed application system environment.The method proposed in this thesis is verified by experiment, and analysis is conducted to achieve the desired goals.
|