Re: 'PUT' transaction reconsidered (was Re: two-phase send concerns )

• From: Larry Masinter <masinter@parc.xerox.com>
• Date: Thu, 28 Dec 1995 01:15:11 PST
• To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
• Message-Id: <95Dec28.011521pst.2733@golden.parc.xerox.com>

At WWW4, someone in the hallway convinced me that I was wrong about
two-phase send, in that:

a) single transaction PUT was a win for small transactions
b) two-phase PUT might win for large transactions, but only because
   the initial 'is this OK to PUT' 'yes, go ahead' was going to be 
   a small RTT compared to the entire PUT transaction.

So, we might leave the 'is this OK to PUT' as an optional part of the
protocol, and let clients use heuristics (how fast/reliable is the
connection to the server vs how big is the transaction) to decide
whether or not to use the optional part.

Received on Thursday, 28 December 1995 01:22:32 UTC