By sin on May 21, 2009
An ordering index is used for searches involving ">=" or "<=" operators. An attribute value is normalized using attribute's ordering matching rule to generate the key. Since the normalization techniques is same for both equality and ordering matching rules, both the indexes end up having the same data. This is something which can be avoided by new design which is already in place for extensible indexes. Hopefully, OpenDS 3.0 will avoid using the separate ordering indexes when not needed.