Re: iSCSI: Login clarification

[Julian_Satran@il.ibm.com]

JP (is this OK?)

The choice of a message to reject is arbitrary and meant to avoid using a
new message.

BUT - lately I came to believe that the best thing to do is ignore those
messages altogether
and thus avoid some nasty denial-of-service attacks.

If there is no outcry from the WG that is what there is going to be in the
next draft.

Julo


Raghavendra Rao <jp.raghavendra@india.sun.com> on 10/10/2000 19:14:24

Please respond to Raghavendra Rao <jp.raghavendra@india.sun.com>

To:   ips@ece.cmu.edu
cc:
Subject:  iSCSI: Login clarification





Julian,

In the current draft, in section 2.2.3 on iSCSI login, it is stated at the
end of the section:

     Any message sent on a TCP connection before this connection gets into
     full feature phase at the initiator should be rejected by the
initiator.
     A message reaching a target on a TCP connection before the full
feature
     phase will be rejected with an iSCSI check condition.

I'm somewhat surprised by the choice of check condition as a means to
indicate
rejection before the full feature phase begins - If we continue to choose
such
mechanisms, somebody might pollute this space with a need of transport
related/based sense keys or ASC, ASCQs to fully qualify a check condition.

I tend to believe that until the full feature phase begins, the
negotiations
are at the transport level (iSCSI), and as such an iSCSI message to
indicate
rejection might be more appropriate ?


-JP