Previous Grid researches envision Grid computing a feasible, option for high performance computing. Recent researches focused mostly on the large scale Grid computing. GIS(Grid Information Service) in Grid middleware provides services for user queries...
Previous Grid researches envision Grid computing a feasible, option for high performance computing. Recent researches focused mostly on the large scale Grid computing. GIS(Grid Information Service) in Grid middleware provides services for user queries to discover available resources as well as to get general information on resources. As the arrival rate of queries becomes larger, which is usually proportional to the increase of the Grid size, it is crucial to implement high performance GIS to make large scale Grid systems in reality. In recent Grid systems, GIS is implemented by using LDAP-based distributed directory systems.
This paper presents a study made to evaluate the performance of LDAP-based distributed directory systems. At first, a performance model is presented respectively for the data distribution and for the data replication scheme. Next, a performance analysis is made to understand the performance trade-off between the two schemes in pursuit of a practical implementation model for high performance Grid Information Systems.