new 6bone pTLA prefix proposal, comments by 4 March 2002 please

[bmanning@karoshi.com]

 broken or not, its part of the protocol design. "fixing" it
 will take more years than has been expended on IPv6.
 


> 
> On Mon, 18 Feb 2002 bmanning@karoshi.com wrote:
> 
> > 	there are areas/places where renumbering fails miserably,
> > 	notably within DNS, SNMP, NTP and anywhere applications
> > 	depend on knowing the whereabouts of remote systems, -by
> > 	address-.  Many applications use the IP address to reduce 
> > 	the delay in a DNS lookup. These applications are sensitive
> > 	to wholesale renumbering, often to to point that they have
> > 	no idea how broadly the hardcoded IP address has spread.
> 
> Imho, an application doing this is fundamentally broken.  One day this
> will come back to bite, especially under IPv6.  I believe that avoiding
> renumbering to cater for apps that did the wrong thing in the first place
> would be a big mistake.  Besides, dns data is cached, so the first lookup
> might be slower but the rest should come off a local dns server.
> Rob
> 
> -- Robert Brockway B.Sc. email: robert@timetraveller.org  ICQ: 104781119
>    Linux counter project ID #16440 (http://counter.li.org)
>    blake: up 40 days,  8:00, 11 users,  load average: 1.00, 1.00, 1.00
>    "The earth is but one country and mankind its citizens" -Baha'u'llah
>