IGD currently targets relational databases, ACeDB databases, ASN.1 files, or tagged-ASCII files. No additional heterogeneity or cooperation assumptions are required regarding IGD's component MBDs, except that their data and metadata are available electronically as files or through a query language. Documentation on or knowledge of the MBD semantics are also assumed.
SRS and LinkDB build index files for retrieving information from structured or semi-structured ASCII files and do not involve databases per se. Thus SRS and LinkDB assume that MBD data are available as ASCII files, and that entities of interest (objects, records, ``database entries'') are represented within these files as contiguous strings.
We are not aware of the assumptions made by Entrez and Docking-D.