SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    RE: Re: iSCSI: Boolean value (yes, no) negotiation



    This seems to be a interesting scenario. 
    
    Mapping this to traditional SCSI model the initiator during message out phase sends to the target its capabilities (support disconnect/reconnect etc) if the target supports the message it acknowledges and the agreement prevails as long as the devices stay powered or else a message reject is sent by the target and the state machine moves on to the next phase.
    
    An inference out of this quote is that the target holds the authority to make final decisions on any proposals that the initiator may choose to make. 
    
    Elaborating the example on which this thread started -
    
    I->T ImmediateData=no
    T->I ImmediateData=no
    
    Irrespective of intelligence either in the target and/or initiator, they should not be soliciting choices otherwise convergence of a proposal might be cumbersome.
    
    Thanks,
    Rakesh
    
    
    Original Message:
    -----------------
    From: Rahul Bhagwat rahulb@veritas.com
    Date: Thu, 7 Feb 2002 00:14:28 +0530
    To: nramas@windows.microsoft.com, ips@ece.cmu.edu
    Subject: Re: iSCSI: Boolean value (yes, no) negotiation
    
    
    In this particular case, the default value for the key is "Yes". Initiator
    should
    be able to express it's preference that it wants to do away with immediate
    date if the target is willing to do so (without forcing it through the AND
    function),
    else the initiator is okay to work with immedidate data. Looks like this is
    not
    possible with the boolean  negotiation. Can we use ImmediateData=? to get
    the
    supported value from the target. If the target responds with yes, the
    initiator can
    either use immediate data (by responding Yes) or not use it (by responding
    No).
    If the target response no, the initiator does not use immediate data. In
    this case,
    the Login need not terminate. Does this give the solution ? Is this a valid
    thing ?
    
    Regards,
    Rahul
    ----- Original Message -----
    From: "Lakshmi Ramasubramanian" <nramas@windows.microsoft.com>
    To: "Rahul Bhagwat" <rahulb@veritas.com>; <ips@ece.cmu.edu>
    Sent: Wednesday, February 06, 2002 11:40 PM
    Subject: RE: iSCSI: Boolean value (yes, no) negotiation
    
    
    > So that means, if the target needs to use ONLY immediate data,
    > it has to fail the login because the initiator said NO to ImmediateData?
    >
    > thanks!
    >  -lakshmi
    >
    > -----Original Message-----
    > From: Rahul Bhagwat [mailto:rahulb@veritas.com]
    > Sent: Wednesday, February 06, 2002 10:07 AM
    > To: Lakshmi Ramasubramanian; ips@ece.cmu.edu
    > Subject: Re: iSCSI: Boolean value (yes, no) negotiation
    >
    >
    > I believe a key is not negotiated thrice as you have pointed out. That
    > is a sender offers a value, receiver offers it's own and that's it. (I
    > think specs has a mention that a key cannot be negotiated twice. This
    > example falls into that category.) The result of the negotiation is key
    > dependent. For example, in this particular key, when initiator sends
    > ImmediateData=no, the negotiation is over as this is an
    > AND function. However it is not an error to send back a response. In any
    > case, the outcome of the result was decided for the initiator when it
    > sent the key first and for the target when it received the key.
    >
    > Regards,
    > Rahul
    > ----- Original Message -----
    > From: "Lakshmi Ramasubramanian" <nramas@windows.microsoft.com>
    > To: <ips@ece.cmu.edu>
    > Sent: Wednesday, February 06, 2002 10:56 PM
    > Subject: iSCSI: Boolean value (yes, no) negotiation
    >
    >
    > > Could someone please clarify Boolean key=value usage
    > > (such as "ImmediateData=yes", etc)?
    > >
    > > For example, the initiator sends to target "ImmediateData=no".
    > > But the target wants ImmediateData. So, it sends back
    > > "ImmediateData=yes".
    > > The initiator, being able to handle it, sends back "ImmediateData=no".
    > > Now, they use immediate data in the PDUs.
    > >
    > > Is this valid? Or, in the above case if the target can't handle
    > > non-immediate data it has to reject the login ?
    > >
    > > thanks!
    > >  -lakshmi
    > >
    >
    
    --------------------------------------------------------------------
    mail2web - Check your email from the web at
    http://mail2web.com/ .
    
    


Home

Last updated: Thu Feb 07 00:18:01 2002
8696 messages in chronological order