Re: radical solutions

[Ran Atkinson <rja@extremenetworks.com>]

At 06:18 15/02/01, Erik Nordmark wrote:

>I understand how GSE-like (and NAT-like) technology helps a lot with
>renumbering for a site that only initiates communication.
>
>But I fail to see how this helps the other end; if a site
>has N RGs then the DNS would return N different addresses for a given
>node each having a different RG (but the same identifier). 
>
>In order to provide good redundancy we would need some mechanism 
>for the initiating party to quickly find a working RG from that set.
>(unless we want to update the DNS each time a customer<-> isp link fails, or 
>unless we want rfc 2260 style local repair between the isps serving
>the customer).

        Folks interested in GSE should take the above comments
as input into how to edit/revise the GSE documentation, IMHO.
In the absence of current GSE documents to discuss, it seems
best that we not assume GSE works on this list -- though GSE
folks should by all means form a design team and work on these
items there, IMHO.

Ran
rja@extremenetworks.com