|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] RE: Keep-alive traffic (was iSCSI: more on StatRN)Glen, The timeout for a connection should be much longer than the time required to plow through a buffer. Again this buffer should be kept lean using a separate flow control from that of TCP. I see little advantage in attempting to process a connection probe out of sequence. Doug > Y P Cheng wrote: > > > You are referring to TCP implementation with software. In an > > TCP-Offload-Engine (TOE) adapter, there is no queue. Every > incoming frame, > > including the ping is served at the speed of the wire... > > The BSD socket API won't allow this sort of look-ahead for a > standard TCP segment. > > So for iSCSI Ping and Ping Response to be worthwhile to allow > software implementations to implement fail-over schemes, the Ping > and Ping Response should use TCP's Urgent feature to allow > out-of-order delivery of the Pings to the socket held open by the > iSCSI target? > > So maybe we need to add: > > Ping and Ping Response SHOULD be marked by TCP as urgent data [RFC793]. > > Do we need to allow for that BSD4.2 mis-interpretation of the TCP > urgent pointer? Does iSCSI have a Null command that we can require > as padding before the Ping commands? > > -- > Glen Turner Network Engineer > (08) 8303 3936 Australian Academic and Research Network > glen.turner@aarnet.edu.au http://www.aarnet.edu.au/ > -- > The revolution will not be televised, it will be digitised >
Home Last updated: Tue Sep 04 01:06:32 2001 6315 messages in chronological order |