Back end architecture based on transformed inverted lists-A surrogate file structure for a very large data/knowledge base

A detailed analysis is presented of transformed inverted lists (TIL) and an inverted surrogate file structure. A parallel back-end architecture, based on TIL, is described for the management of a very large data/knowledge base. The approach relies on an inverted list-indexing scheme that is performed on the surrogate files instead of the usual database inversion applied in conventional information retrieval systems. Equations to estimate the storage overhead and query response time are presented, showing the method to be space- and time-efficient.<
>