We present two server-initiated p r otocols to improve the performance of distributed information systems (e.g. WWW). Our rst protocol is a hierarchical data dissemination mechanism that allows information to propagate from its producers to servers that are closer to its consumers. This dissemination reduces network tra c and balances load amongst servers by exploiting geographic and temporal locality of reference p r operties exhibited in client access patterns. Our second protocol relies on

more »

... peculative service", whereby a request for a document is serviced by sending, in addition to the document requested, a number of other documents that the server speculates will be r equested i n t h e n e ar future. This speculation reduces service time by exploiting the spatial locality of reference p r operty. We present results of trace-driven simulations that quantify the attainable performance gains for both protocols.