RE: iSCSI: NOPs on discovery session

["Julian Satran" <Julian_Satran@il.ibm.com>]

Somesh,

NOPs are used for more than keep alive. They carry number acknowledgements 
and are done by the iSCSI layer while TCPs keep alive being kept at the 
TCP level does not say anything about the layers above (usually works 
between  brain-dead machines).  In general the higher the level of a 
"keep-alive" the more functional it is.

Julo




"Somesh Gupta" <somesh_gupta@silverbacksystems.com>
Sent by: owner-ips@ece.cmu.edu
27-10-01 03:31
Please respond to somesh_gupta

 
        To:     "Ayman Ghanem" <aghanem@cisco.com>, <ips@ece.cmu.edu>
        cc: 
        Subject:        RE: iSCSI: NOPs on discovery session

 

Ayman,

It is not completely clear what the objective of the
NOP command is? TCP itself has a keep-alive "feature"
that can be used to ensure that the system at the
other end is alive. So you don't need anything else
to ensure that the "target is alive".

However, we could use it as an opportunity to add
another feature in the protocol and reinvent TCP
twice.

Somesh

> -----Original Message-----
> From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
> Ayman Ghanem
> Sent: Thursday, October 25, 2001 8:29 PM
> To: ips@ece.cmu.edu
> Subject: iSCSI: NOPs on discovery session
> 
> 
> Julian,
> 
> I am not sure if this came up before, but we also need to include 
> NOP-OUT to
> the set of commands allowed on the discovery session. An 
> initiator may keep
> the discovery session active, and send NOP-OUT to ping the target, or in
> response to a target ping.
> 
> -Ayman
> 
>