A model for a Distributed Directory System

this document presents the spectrtcatton of a distributed algorithm for the name resolution problem In distributed directory systems. The algorithm essenttally performs the search1ng In the Name Tree, also mantatntng a distrbuted cache or network addresses, structured as a tree, to reduce the cost of the search. in terms or remote accesses. The algorithm is based on abstract models for directories as we11 as for networks, whicn are also described in the present paper. A first experimental distributed implementation of a directory system on a network or IBM PCs is going to, start based on the work presented In this paper.