- From: Simon Spero <ses@tipper.oit.unc.edu>
- Date: Wed, 27 Dec 1995 17:55:54 -0800 (PST)
- To: Jeffrey Mogul <mogul@pa.dec.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, http-ng%cuckoo.hpl.hp.com@hplb.hpl.hp.com, mogul@pa.dec.com
Probably should take this to the persist list, or the http-ng mailing list (now at http-ng@cuckoo.hpl.hp.com) The NG approach to HTTP/1.X encapsuation and persistent connections makes this tradeoff easier, since cancellation doesn't require the connection to be torn down. The RTT costs with this approach using optimistic transmission are 0 RTT for the successful case, and approximately 1 in the event of failure (the costs can be greater than one if data is being enqueued at the TCP layer faster than the TCP transmission rate; the excess data is then drained. Simon (defun modexpt (x y n) "computes (x^y) mod n" (cond ((= y 0) 1) ((= y 1) (mod x n)) ((evenp y) (mod (expt (modexpt x (/ y 2) n) 2) n)) (t (mod (* x (modexpt x (1- y) n)) n))))
Received on Wednesday, 27 December 1995 17:57:41 UTC