The paper considers the data search problem in distributed P2P networks. Since no central search engines are available, other effective methods must be developed to avoid a complete search of all nodes each time. We investigate mechanisms and introduce the main structure and functionality of such a decentralized cooperative search engine. It is working on the basis of so called ants. This search engine reduces the search time for information, which are needed by more than one node in a peer-to-peer network community. Therefore, the ants are able to switch in a special manner between different behavior strategies to search, concentrate and return the respective information. The various strategies are derived from ant colonies and were simulated in a small world community environment.