[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Modified IPv6 to unmodified IPv4
On 2007-10-23 10:42, Stephen Sprunk wrote:
Thus spake "Brian E Carpenter" <brian.e.carpenter@gmail.com>
That doesn't cover us for the day when we have true IPv6-only apps
with no reptilian memory of IPv4. I think I'd rather avoid the
application having any knowledge of IPv4 at all. If we're going to
do this horrible thing, let's do it 100%. (As Iljitsch knows, I'm
playing with some thoughts on this, but they aren't quite ready for
public ridicule yet.)
Will code like that ever exist? The changes to the sockets API (e.g.
getnameinfo() and getaddrinfo()) has made it so that modern apps aren't
any more v4-aware than they're v6-aware; they've become
version-agnostic, not separately aware of v4 and v6, unless they have a
specific reason to know.
Correct. I think the tricks that have to be played are below the
socket interface. (Although there is the residual problem of applications
that ignore RFC 1958, RFC 2101 and RFC 2775 by storing IP addresses for
future reference, but that is probably insoluble.)
The problem I think you're indirectly referring to is NAT awareness, and
most of the problem there is going to exist in v6 with stateful
firewalls even if we do manage to avoid creating a NATv6 world. ALGs
are simpler when you only have to inspect packets, not mangle them, but
they're not going away, and end hosts/apps can't get away from coping
with administratively-imposed reachability problems.
Unfortunately I believe you are correct; this has to be an exercise
in damage limitation.
Brian