Re: [OSPF] draft-venkata-ospf-dynamic-hostname-02.txt

[Carlos Pignataro <cpignata at cisco.com>]

Please find two additional small comment inline.

On 2/29/2008 7:15 PM, Subbaiah Venkata said the following:
> Hi Anton,
> 
> Great questions. Some of them we already thought but deferred for WG 
> consensus.
> Answers inline.
> 
> On Fri, Feb 29, 2008 at 4:21 AM, Anton Smirnov <asmirnov at cisco.com 
> <mailto:asmirnov at cisco.com>> wrote:
> 
>        Hi all,
>        have a comment on this draft (OK, it's not WG document yet but hope
>     that's not a problem).
>        The document doesn't stipulate that Dynamic Hostname TLV can be
>     present at most once. 
> 

In addition to the scope clarifications from below, we can also 
explicitly mention one TLV per RI LSA.

> 
> Yes. I think it is better to keep it simple by mandating one Dynamic 
> hostname TLV per originator.
> This saves unnecessary flooding and network bandwidth.
>  
> 
>     Also, if implementation chooses to announce both
>     AS and area scope RI LSA then it is not clear if it is allowed to
>     announce different names.
> 
> 
> Again, one TLV per originator. If a router is originating at least one 
> AS scope LSA
> (for example, an ASBR originating AS external LSAs) then it should 
> generate AS scope
> Dynamic hostname TLV. Otherwise it should originate area scope TLV.
>  
> 
> 
> 
>        I personally would prefer 1:1 mapping between ID and router's name
>     to be required but some people might disagree with me. Either way
>     document should specify if multiple names are allowed or not.
> 
> 
> We will include some wording on these in the next revision of the draft.
>  
> 
> 
>        If we are to send text string then document should specify what
>     exactly is understood by 'text'. Is it ASCII or UTF-8 encoding? Are
>     symbols like 0x1 or 0xAF allowed?
> 
> 
> We discussed and deferred this for WG consensus.
> I prefer ASCII text unless there is a strong reason for UTF-8 or UNICODE.

We'd like to get thoughts and opinions on this, as likely arguments for 
both encoding in US-ASCII and UTF-8 charsets can be made. E.g., a 
hostname as a globally/widely visible name would use US-ASCII [RFC2277], 
etc.

Thanks,

--Carlos.

> 
> Thanks for the review.
> 
> Venkata.
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> OSPF mailing list
> OSPF at ietf.org
> https://www.ietf.org/mailman/listinfo/ospf

-- 
--Carlos Pignataro.
Escalation RTP - cisco Systems
_______________________________________________
OSPF mailing list
OSPF at ietf.org
https://www.ietf.org/mailman/listinfo/ospf