Re: iSCSI: Case-sensitivity in iSCSI names

[Mark Bakke <mbakke@cisco.com>]

Glen Turner wrote:
> 
> John Hufferd wrote:
> >
> > I want to jump on this bandwagon too.  This seems to be exactly the right
> > approach.
> 
> I'm showing some caution.
> 
> It seems that NAMEPREP still requires casing tables.  Where
> and how are these maintained?

NAMEPREP casing only needs to be done when the iSCSI name 
is assigned, and it can be simplified in some cases, such
as when names are generated using only ASCII.  Since the
assignor of the names completely controls which character
sets are used, it can include the appropriate normalization
code:

- If the names are just generated using lower-case (in any
  character set) plus digits, no normalization is required.

- If the names are generated from some other all-ASCII string,
  tolower() normalizes and isalnum() verifies.

- If the names are generated from more general, internationalized
  text, either the equivalent of tolower() and isalnum() appropriate
  to the character set may be used, or the NAMEPREP procedure
  that Bob sent in his last email can be used.

In any case, it is at the complete discretion of the naming
authority that generates the name whether to use one of the
above methods.  It can be kept very simple (first or second
method), or include case tables to make it more user-friendly.

User-friendly naming conventions are an option, and they always
cost something.
 
> Does the proposal lead to language-based variants of
> iSCSI products.  At the moment I can safely buy a
> hard disk drive from the USA and use it anywhere in
> the world: IDE, FC or SCSI.  That should be so for
> iSCSI.

That's still the case.  An iSCSI name that is not in the
NAMEPREP-normalized form MUST NOT be used within the iSCSI
protocol.  Therefore, all compares on iSCSI names are opaque
and byte-for-byte, so it doesn't matter where the device
is manufactured, or in what locale it is used.

--
Mark7

> Mark wrote:
> > Anyway, if this is to be used for domain names, we ought to use it,
> > too.
> 
> Except that no-one is suggesting embedding DNS in a platform
> that won't be maintained after sale.
>
> Before support one way or the other I'd like to be
> clearer about the boundaries of iSCSI naming.

Hopefully, the above helps.

> Regards,
> Glen
> 
> --
>  Glen Turner                                 Network Engineer
>  (08) 8303 3936      Australian Academic and Research Network
>  glen.turner@aarnet.edu.au          http://www.aarnet.edu.au/
> --
>  The revolution will not be televised, it will be digitised

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