Ran,

Thanks.  I was reaching a conclusion based on what someone said at the last 
6man session.  Given that opaque is not what I thought, your suggestion makes 
sense to me.

Bob

On Dec 6, 2011, at 11:35 AM, RJ Atkinson wrote:

> 
> Bob,
> 
>  My read of the relevant RFCs is that the 
> DHCP Line-ID option specifies the use of the "NVT"
> character set.  
> 
> The character set specification is implicit, applying 
> to all DHCP options, rather than having been written
> explicitly for this single DHCP option.  One has 
> to really study the full set of DHCP RFCs 
> (i.e. the full daisy-chain of RFC citations from
> the References backwards) to see this, but it is
> consistent with every DHCP implementation that 
> I can identify.
> 
>  I can't find ANY DHCP implementations that permit 
> "arbitrary bytes" to be entered into the Line-ID option 
> field. I can't think of any DHCP implementation that
> permits, for example, LINEFEED or CARRIAGE-RETURN
> to be entered into ANY DHCP-related field.  Many
> have problems with inputs other than printable
> characters from ISO-8859, for example.  
> 
>  The word "opaque" in the DHCP Line-ID option
> specification refers to the packet processing
> for the DHCP packet, not to the contents of
> the field or how that field is entered 
> (presumably via a CLI or perhaps web page someplace) 
> or how that field is to be displayed to humans.
> 
> The DHCP Line-ID option does not use the word 
> "opaque" to indicate absence of a character set.
> That is a mis-reading of the DHCP RFC set.
> 
> As a former ISP person, I can assure that the
> Line-ID values absolutely DO get displayed
> to operators and others in network operations.
> It is important to be able to configure the
> Line-ID values and to view them in order to
> troubleshoot real networks (e.g. a DSL deployment).
> 
> So the assumption that the DHCP Line-ID option
> does not have any specified character set is
> not actually correct.
> 
> Now, if one's goal is *strict* compatibility with
> the DHCP Line-ID option, then the NVT character
> set (originally specified in very early RFCs)
> ought to be specified.  
> 
> I think, however, it would be reasonable to choose 
> some alternative such as US-ASCII, ISO-8859-1, 
> or (maybe -- possibly more risky given existing
> DHCP implementations -- but more globally suitable)
> UTF-8.
> 
> Yours,
> 
> Ran
> 
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [email protected]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------

--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to