The idea of this article is based on the parallel indexing concept (Taniar & Rahayu, 2002) in which an indexed object residing in a BS is either fully, partially, or not replicated to others BSs. Therefore, every server contains either partial or all indexes of other servers. In our proposed approach, whenever the requested results return from neighboring cells, we append the resulting items to the current cell. This implies that when the next user sends a request, the current cell needs to look up its own index first to verify if the data is in its local storage. If the data is not present, the current server sends a request to the neighboring cells on behalf of the client; otherwise, the current server directly sends the requested query to the client. We have evaluated our proposed approach and showed that the access time can be reduced by a factor of two.The next section of this article describes some related work. We then describe our proposed work and the simulation model, and we compare the performance of our proposed technique to other techniques. Finally, we conclude the article and suggest future work.