Re: iSCSI 4.1 & 4.2

[Mark Bakke <mbakke@cisco.com>]

Julian-

We originally were going to just deal with this in the NDT
spec, but since the NDT spec (and the referenced stringprep
spec) affects the character set that is being defined in 4.1
and 4.2, it will be hard to avoid referencing it from 4.1 and
4.2 as well.

--
Mark

Julian Satran wrote:
> 
> Mark,
> 
> I thought that this will be handled within NDT and I will see the end-result.
> I can't make the reference before looking at the list of characters for conflicts and scan
> dead-ends.
> 
> Julo
> 
>   Mark Bakke <mbakke@cisco.com>
>   Sent by: mbakke@cisco.com                   To:        Julian Satran/Haifa/IBM@IBMIL
>                                               cc:        ips@ece.cmu.edu
>   05/01/2002 10:59 PM                         Subject:        Re: iSCSI 4.1 & 4.2
>   Please respond to Mark Bakke
> 
> 
> Julian-
> 
> The specification is at
> 
> http://search.ietf.org/internet-drafts/draft-ietf-ips-iscsi-string-prep-00.txt
> 
> It would be better to reference the specification rather than attempt
> to add the individual characters to iSCSI; the point
> of the separate specification is to leverage the same work that
> the IDN folks are doing.
> 
> --
> Mark
> 
> Julian Satran wrote:
> >
> > Please send me a list of the characters to add or a specification for them. Julo
> >
> >   Mark Bakke <mbakke@cisco.com>
> >   Sent by: mbakke@cisco.com                   To:        Julian Satran/Haifa/IBM@IBMIL
> >                                               cc:        ips@ece.cmu.edu
> >   05/01/2002 09:40 PM                         Subject:        Re: iSCSI 4.1 & 4.2
> >   Please respond to Mark Bakke
> >
> >
> > Julian-
> >
> > Sorry, I've been unavailable the last few days.  What do you mean
> > about the note not being concrete?  The string-prep draft says
> > exactly which additional characters we need for the name-value
> > (all of the international lower-case equivalent characters
> > are valid), and the utf8-value specifies that any UTF-8
> > character is legal (other than \0).
> >
> > Without adding these, we would have to ditch internationalization
> > completely within iSCSI names, but since the DNS folks are going
> > through the pain of internationalizing host and domain names, we
> > should keep internationalization in iSCSI as well.
> >
> > --
> > Mark
> >
> > Julian Satran wrote:
> > >
> > > Mark,
> > >
> > > John Hufferd has solicited twice input from NDT about their additional requirement.
> > > I did not hear back from you.
> > > The attached note does not say anything concrete (are some characters that you need missing?)
> > > either.
> > >
> > > Julo
> > >
> > >   Mark Bakke <mbakke@cisco.com>
> > >   Sent by: mbakke@cisco.com                   To:        Julian Satran/Haifa/IBM@IBMIL
> > >                                               cc:        ips@ece.cmu.edu
> > >   05/01/2002 08:52 PM                         Subject:        Re: iSCSI 4.1 & 4.2
> > >   Please respond to Mark Bakke
> > >
> > >
> > > Julian-
> > >
> > > Text-value is also used for initiator names, target names,
> > > and aliases, which are UTF-8 encoded, so they don't fit into
> > > the current text-value definition.   Here are two more
> > > text value types we should add.
> > >
> > > --
> > > Mark
> > >
> > > Julian Satran wrote:
> > > >
> > > > text-value: a string defined by the regular expression
> > > > "[][a-zA-Z0-9.-+@_/\[\]=]*"
> > >
> > > Used by InitiatorName and TargetName:
> > >
> > > name-value: a UTF-8 string, with its valid character set
> > > defined in draft-ietf-ips-iscsi-string-prep-00.
> > >
> > > Used by InitiatorAlias and TargetAlias:
> > >
> > > utf8-value: any UTF-8 string, terminated by a null character,
> > > Ranges and comma-separated lists are not allowed for keys
> > > that use this type.
> > >
> > > --
> > > Mark
> >
> > --
> > Mark A. Bakke
> > Cisco Systems
> > mbakke@cisco.com
> > 763.398.1054
> 
> --
> Mark A. Bakke
> Cisco Systems
> mbakke@cisco.com
> 763.398.1054

-- 
Mark A. Bakke
Cisco Systems
mbakke@cisco.com
763.398.1054