The large unstructured text collections demand full-text search capabilities from IR systems. Current systems typically allow users only to connect to a single database (or site) at a time. An IR system should be able to provide multiple users with concurrent, efficient access to multiple text collections distributed across remote sites. However, high resource demands of networked IR systems limit their performance as the size and the number of data collection increase. In general, the size of inverted file index for fulltext search is either as big as that of original text collection or more. Our goal is to represent a new index management method, which is fast enough in practice while the performance is not limited by the size and number of data collections.