As the population of the galaxy increases and new worlds are settled, timely access for home users and frontier settlements with underdeveloped communications infrastructures is a growing problem. To ameliorate bandwidth issues, a sophisticated array of data caches and virtual intelligence search agent programs are available.
When a user submits a query, it is first routed to the data cache, the user's search agent VI collates mountains of locally-stored data to find the desired material. If the information is not available locally, the query is passed along to neighboring systems, and then outward in an expanding network. VI search agents in those systems replicate the search. If the desired information is found, it is compressed into a "burst" file and queued for transmission to the source system. The burst is assigned a priority based on the number of queries for it; the greater the number of queries, the higher the priority.
When a new solar system is first connected to the net, a selection of the most popular data is installed locally. Though storage hardware is cheap, the capacity required to hold all the data produced everyday by trillions of people on hundreds of worlds is not trivial. It's not economical to store local copies of all the data available on obscure topics just in case.
As colonies mature, older and less-popular chunks of data filter into them as a result of queries and are placed in the local archive. Searches for obscure topics are increasingly likely to produce instant results as the archive grows.