Network design for the implementation of text searching using a multicomputer

摘要

This paper discusses the design of a nearest neighbour document retrieval system that has been implemented on a multicomputer network. This network is built from transputers, high performance microprocessors that have been developed specifically for the construction of highly parallel computing systems. The system uses a two-stage retrieval algorithm in which an initial text signature search is used to exclude large numbers of documents from the detailed and time-consuming pattern matching search. The latter is implemented using a processor farm, so that documents matching at the signature level can be examined in parallel to determine whether they are, in fact, a good match for the query. The logical and physical processes required for the implementation of such a two-stage search are described and the development of an efficient network for the pattern matching stage of the search is discussed.