Re: ISCSI: Login Parameters - Defaults?

["Mallikarjun C." <cbm@rose.hp.com>]

I agree with David's recommendation that all keys that one
cares for must be explicitly negotiated, but -

>if the other side happens to use a different
> default,....

I don't believe that'd be a compliant implementation in such a case. 

I believe the required level of compliance on all defaults is "MUST".
If not (and coupled with the truth in David's recommendation), I don't 
particularly care if we do away with defaults.....

Thanks.
--
Mallikarjun

Mallikarjun Chadalapaka
Networked Storage Architecture
Network Storage Solutions Organization
Hewlett-Packard MS 5668 
Roseville CA 95747
cbm@rose.hp.com


----- Original Message ----- 
From: <Black_David@emc.com>
To: <kevin_lemay@agilent.com>
Cc: <ips@ece.cmu.edu>
Sent: Saturday, April 27, 2002 10:15 PM
Subject: RE: ISCSI: Login Parameters - Defaults?


> Kevin,
> 
> Part of what's going on here is that a robust implementation
> minimizes assumptions about what the other party to the communication
> will do - this is part of the usual IETF admonition to be
> conservative in what is sent and liberal in what is accepted.
> If the default value is for convenience/preference and your
> implementation can cope if the other side happens to use a different
> default, then there's no need to negotiate that key.  OTOH, if us
> of the default is important to your implementation for some reason,
> it is incumbent upon you to negotiate it.  It is not robust
> to assume that the other side will use the default in the absence
> of negotiation and then complain that it's the other implementation's
> fault that things went wrong when it didn't use the specified
> default -- it was your responsibility to negotiate the key if
> you cared about its value.  Yes, this results in potentially
> verbose negotiations, but it reduces the possibility for confusion
> about what's supposed to happen as a result.
> 
> Thanks,
> --David
> 
> -----Original Message-----
> From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
> Sent: Thursday, April 25, 2002 3:33 PM
> To: LEMAY,KEVIN (A-Roseville,ex1)
> Cc: ips@ece.cmu.edu
> Subject: RE: ISCSI: Login Parameters - Defaults?
> 
> 
> 
> I am trying again - If you are happy with the dafualt don't do anything. If
> you want to change it you must negotiate - not just answer to an implied
> offer. 
> 
> Julo 
> 
> 
> "LEMAY,KEVIN (A-Roseville,ex1)" <kevin_lemay@agilent.com> 
> 04/25/2002 10:04 PM 
> Please respond to "LEMAY,KEVIN (A-Roseville,ex1)" 
>         
>         To:        Julian Satran/Haifa/IBM@IBMIL 
>         cc:        ips@ece.cmu.edu 
>         Subject:        RE: ISCSI: Login Parameters - Defaults? 
> 
>        
> 
> 
> Julian,
> 
> But defaults are an implicit offer. If I do not offer the key the I am
> assuming the default.
> 
> Can you re-word the sentence to incorporate what you just explained in the
> e-mail... This is much clearer, at least to me.
> 
> Kevin
> 
> 
> -----Original Message-----
> From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
> Sent: Thursday, April 25, 2002 11:47 AM
> To: LEMAY,KEVIN (A-Roseville,ex1)
> Cc: ips@ece.cmu.edu
> Subject: Re: ISCSI: Login Parameters - Defaults?
> 
> 
> 
> It means only that regardless of the fact that a default is available if you
> state first a key=value you are the originator and the other party MUST
> respond unless specified otherwise (as in booleans). In other words there is
> no such think as an implicit offer of the current or default value. 
> 
> Julo 
> 
> 
> 
>                 "LEMAY,KEVIN (A-Roseville,ex1)" <kevin_lemay@agilent.com> 
> 
> 
> 04/25/2002 07:01 PM 
> Please respond to "LEMAY,KEVIN (A-Roseville,ex1)" 
> 
> 
>        
>        To:        Julian Satran/Haifa/IBM@IBMIL 
>        cc:        ips@ece.cmu.edu 
>        Subject:        ISCSI: Login Parameters - Defaults? 
> 
>       
> 
> 
> Page 64 in iSCSI v12 says,
> 
> "All negotiations are stateless and explicit (i.e., the result MUST be
> based only on newly exchanged values). There is no such thing as implicit
> offers. If an explicit offer is not made then a reply cannot be expected."
> 
> It is unclear to me how this is supposed to work. Does this mean that I must
> negotiate every key pair? This is confusing because chapter 11 still
> contains default values.
> 
> If I do not negotiate a parameter, is it assumed to be set at the default or
> must I negotiate every parameter ?
> 
> If I must negotiate every parameter, then the "default" value should be
> removed or renamed to "suggested setting".
> 
> Thanks,
> 
> Kevin Lemay
>