RE: iSCSI: large keys during login?

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

Eddy,

There are many ways to do it. Recall that we had no spanning at all and that was best. But it was not good enough for all values with small PDU's.
The alternative would be to allow at least the sender to lay out data and the send-it out by cutting it in pieces of PDU length without it having to parse.

What you are suggesting requires the sender to build the stream considering the boundaries.

We could go either way but IMHO it is better as we have it now - only the receiver gets more complicated not both receiver and sender.

Julo

Eddy Quicksall <eddyq@ivivity.com> 05/03/2002 11:22 PM Please respond to Eddy Quicksall

To:        Julian Satran/Haifa/IBM@IBMIL, "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>         cc:        andy currid <andy@windriver.com>, ips@ece.cmu.edu, owner-ips@ece.cmu.edu, Bill Studenmund <wrstuden@wasabisystems.com>         Subject:        RE: iSCSI: large keys during login?

Regarding the following sentence in 9.10.4, can we say the key and equal sign most not be spanned? The reason I ask for this is because it makes the parsing of the key much easier if you don't have it cut into two PDUs.

A key=value pair can span Text request or response boundaries (i.e., a

key=value pair can start in one PDU and continue on the next).

 

Eddy

-----Original Message-----
From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
Sent: Monday, April 29, 2002 8:35 PM
To: THALER,PAT (A-Roseville,ex1)
Cc: andy currid; ips@ece.cmu.edu; owner-ips@ece.cmu.edu; Bill Studenmund
Subject: RE: iSCSI: large keys during login?


Pat,

For spanning look at 9.10.4 and 9.11.4.

I assume that the text covers whatever does not fit.


Regards,
Julo

"THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com> 04/30/2002 02:24 AM Please respond to "THALER,PAT (A-Roseville,ex1)"

To:        Julian Satran/Haifa/IBM@IBMIL, Bill Studenmund <wrstuden@wasabisystems.com>        cc:        andy currid <andy@windriver.com>, ips@ece.cmu.edu, owner-ips@ece.cmu.edu        Subject:        RE: iSCSI: large keys during login?

Julian, 
  
I assume you mean that keys can span requests as well. I can't find anything in the draft that says that they can though there also isn't anything that says they can't. Does this only apply to key-value pairs too long to fit in a single response or can a short key-value pair span a request/response when multiple key-value pairs are packed into a PDU? 
  
The draft should clarify. 
  
-----Original Message-----
From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
Sent: Monday, April 29, 2002 3:54 PM
To: Bill Studenmund
Cc: andy currid; ips@ece.cmu.edu; owner-ips@ece.cmu.edu
Subject: Re: iSCSI: large keys during login?


Key can span responses. Bookmarking is for things that have multiple answers - like the SendTargets - Julo

Bill Studenmund <wrstuden@wasabisystems.com> Sent by: owner-ips@ece.cmu.edu 04/30/2002 01:15 AM Please respond to Bill Studenmund

To:        andy currid <andy@windriver.com>       cc:        <ips@ece.cmu.edu>       Subject:        Re: iSCSI: large keys during login?

On Mon, 29 Apr 2002, andy currid wrote:

>
> Hi
>
> In iSCSI draft 12, chapter 10, the Kerberos and SPKM authentication
> mechanisms specify a limit of 64k on their unencoded key values.
>
> Given that these keys are only used during login and, during login,
> the PDU size in use is 8k and, unlike text commands, there is no
> bookmarking, how can you send a key value of this size during login?

I thought that was covered. If you say you can do kerberos or SPKM, you
are saying you can do 64k.

Take care,

Bill