Re: updating GSE for the new millennium

[Brian E Carpenter <brian@hursley.ibm.com>]

Indeed. Getting back to GSE, I think the question we have
to ask about any solution is: how would we bootstrap
the Internet, or a large portion of it, if the power went
off everywhere for a while? If the solution isn't robust
in that case (e.g. due to circular dependencies) then it
doesn't measure up. As Tony said, you can arrange things so
that DNS-based GSE doesn't create circularity, but only if you 
think about it.

   Brian

Tony Li wrote:
> 
> Root name servers are also a service deserving of anycast
> addresses.
> 
> Tony
> 
> |    Joe is absolutely on the right path.
> |
> |    > contributed the following to RFC 1958:
> |    >
> |    >    3.11 Circular dependencies must be avoided.
> |    >
> |    >       For example, routing must not depend on look-ups
> |    in the Domain
> |    >       Name System (DNS), since the updating of DNS
> |    servers depends on
> |    >       successful routing
> |
> |    And if you change this to:
> |
> |    >       For example, core default free routing must not
> |    depend on look-ups in
> |    >       the Domain Name System (DNS), since the updating
> |    of DNS servers depends
> |    >       on successful defaul free routing
> |
> |    Long ago there was a discussion about whether or not DNS
> |    servers were
> |    things that needed to be multihomed - the answer was that
> |    they were not.
> |
> |    One does far better with multiple DNS servers on unique PA
> |    addresses.
> |    DNS already deals with failover, etc...
> |
> |    ]       ON HUMILITY: to err is human. To moo, bovine.
> |         |  firewalls  [
> |    ]   Michael Richardson, Sandelman Software Works, Ottawa,
> |    ON    |net architect[
> |    ] mcr@sandelman.ottawa.on.ca
> |    http://www.sandelman.ottawa.on.ca/ |device |    driver[
> |    ]
> |    panic("Just another Debian GNU/Linux using, kernel
> |    hacking, security guy"); [
> |
> |