With the rapid increase of Internet users

it is more and more important for Internet information retrieval systems to process large amount of users' query efficiently and rapidly.In this paper

we propose a model

Fastresponse

of distributed information retrieval systems for Internet applications.Currently

the distributed information retrieval systems on the Web have problems in the aspects of availability

reliability

scalability and intelligent query service.In Fastresponse

we propose some methods for the management of Web server to resolve the problems above.Thus

we can better process the users' query.We describe the features of Fastresponse.The system adopts a two-level scheduling policy to arrange the works of servers and documents.The system uses the standby servers that are dynamically selected when the system requires.The system monitors the running servers' states in the same cluster and among the clusters.The system automatically regroups the servers according to the use's request.