The client accessing the MDB web services has to wait for as long as it takes to get the results. This can be very inefficient if complex queries or analyses are requested. Decouple the request and the retrieval of the results, to allow scheduling of query execution to periods of decreased server load. Query result caching: if the same query is requested by different clients, it will be executed only once. Flexible result/documentation delivery. For example, an external web application can register interest in knowing the latest metalloproteins indexed in the MDB, for that it will register a callback service, a method/protocol, and a result format.