Eliminate information inconsistency for X.500 directory

This paper introduces a method of avoiding information inconsistency in X.500 directories supporting real services. The basic X.500 directory mechanism does not define the relationships between the attribute values of the different entries. If some attribute values are related to other attribute values of different entries, information inconsistency sometimes occurs. To extended the directory service, we have created an index sub-tree, which define relationships in the directory information tree (DIT). In a distributed environment the index sub-tree sometimes causes the information inconsistency. To avoid this inconsistency, we propose a method that holds user-defined relationships, scans a relationship tree and corrects any inconsistency automatically. This method works as a external process, so it yields short time delays. To evaluate the effect of the delay, we propose a criteria based on statistical analysis. The criteria maintains the information so that user-experienced delay is small. We also discuss information reliability, a very important factor for real directory services. We have encompassed this factor in the criteria