Paper #: 3.0.8
Title:  Speculative Data Dissemination and Service to Reduce Server Load, Network Traffic and Service Time for Distributed Information Systems

1. Problems
-better dissemination of data through new server-initiated protocols
-client-side "speculative service" vs. server-side

2. New Idea and Strengths
-clusters of servers analyze usage data to find more popular files. these files are stored in a proxy server that is the first server requested by clients. by placing the more popular files here, a significant increase in response is gained.
-files are assigned probabilities of being viewed according to the same usage data, and the most likely "next" files are sent to the client's proxy, in the hopes that when these files are requested they will already be locally cached.

3. Weaknesses and Extensions
-doesn't really address the problems in bandwidth created by all this speculative sending of files
-they only examine the usage data for their own site, which is an educational site with relatively few updates. it remains to be seen how this model would translate to commercial sites, or sites where the primary content is constantly in flux.
-based on personal observations, I am unsure how much the benefits gained in performance by speculative caching will outweigh the bandwidth increase caused by this. it may be faster for typical behavior, but it could be slower in cases of more random behavior.
-the final idea of a combination of client-side and server-side speculation being collaborative was good. continuing in this thread, I think this service could be extremely beneficial to frequently visited sites. A user could mark certain pages in their Favorites listing so that the server will provide cached files for those sites only, and for random browsing servers will not pre-fetch.