Research on agent mechanism for heterogeneous distributed databases

对异构数据库，特别是半结构化数据库的检索方法的研究很多，但实际操作的项目并不多。本文对分子生物学数据库的通信方式和运行环境进行了探讨。本研究对象选择了被引用最多的四个数据库Genbank、Prosite、PIR和PDB，并将这四个数据库的文件存储在我们实验室的计算机中。代理程序使用我们以前的工作结果进行编码，并具有XML格式的接口。访问这些数据库的程序以Java applet的形式提供。用户可以使用这些程序访问数据库。对象之间的通信由HORB对象实现。代理可以移动数据库并收集必要的数据。对于这种智能体控制，我们研究了IBM公司发明的aglets。此外，作为研究成果的一个应用，我们研究了在教育机房的每台PC机上收集个人学习记录的基础，在局域网套接字环境下保证用户认证的方法，以及如何将这些记录用于教育。

There are plenty of research on heterogeneous databases, particularly on retrieval method from semi-structured databases, however, the number of projects for actual operation was not so many. In this research, communication method and environment for operation of molecular biology databases are discussed. The four most referenced databases, i.e., Genbank, Prosite, PIR and PDB, are selected for this research target and the files of these four databases are stored in computes in our lab. Agent programs are coded using our previous work results and have XML format interface.The programs which access these databases are provided as Java applets. Users can use these programs to access the databases. The communication between objects is implemented by HORB objects. Agents can move around databases and collect necessary data. For this agent control, we studied aglets which is invented in IBM.Furthermore, as an application of our research results, we studied the basis for colleting personal learning record in every PC in the educational computer lab, the way to assure the user authentication in a LAN socket environment, and how to utilize those records for education.