- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Mon, 11 Sep 95 15:02:15 MDT
- To: Shel Kaphan <sjk@amazon.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Just to follow up on my own reply to Shel's proposal: Pretty much by accident, I discovered that several people at Boston University, Azer Bestavros and Carlos Cunha, have also been looking at predictive prefetching approaches (which they call "speculative" prefetching). They have published several technical reports (see http://cs-www.bu.edu/techreports/Home.html) but I have not seen anything from them appear elsewhere, including here. So far, they seem to have looked only at purely server-initiated and purely client-initiated models, rather than the hybrid that Shel and I have each proposed. But they have done trace-based simulation studies that apparently show performance gains. -Jeff
Received on Monday, 11 September 1995 15:10:31 UTC