RE: iSCSI - long key values

["Wheat, Stephen R" <stephen.r.wheat@intel.com>]

I don't know why that is any different than receiving a long list of
key=value pairs that you want to hold onto until you have them all, prior to
operating on any of them, given you don't know that they'll arrive in any
given order that will yield one's ability to process them "on-the-fly"
without possibly having to undo actions.


At the very least, the text has to go somewhere anyway.  To "remember" where
to keep copying incoming text shouldn't be overly burdensome for the
Initiator.

Stephen

-----Original Message-----
From: Dave Sheehy [mailto:dbs@acropora.rose.agilent.com]
Sent: Wednesday, September 19, 2001 3:59 PM
To: ips@ece.cmu.edu
Subject: RE: iSCSI - long key values



> So, am I missing something here?

How do you avoid overflowing either side? Right now, each side only has to
buffer one PDU worth of data (per connection). Your proposal increases that
requirement to at least 2 PDUs.

Dave Sheehy