Re: To address or NAT to address?

[Keith Moore <moore at cs.utk.edu>]

> It seems to me that we may be able to recapture some aspects of end-to-end
> transparency at the application level if addressing issues are focused on
> host FQDNs, rather than IP addresses.

this works to some extent.  it specifically doesn't work for applications
that need to rendezvous with specific processes on specific hosts
(or which need to use specific interface addresses, say for performance
reasons), since a single FQDN often corresponds to multiple hosts.
(or, less frequently, to multiple addresses on a single host).

note also that DNS is often slow, and seems less reliable than IP.
by increasing the reliance on DNS you increase the probability of failure.