turning off IPv6

[Michael Poole]

John DeCarlo writes:

> On 7/5/06, Tony Arnold <tony.arnold at manchester.ac.uk> wrote:
> >
> > On Wed, 2006-07-05 at 12:53 -0400, John DeCarlo wrote:
> >
> > > Ideally, IPv4 should be disabled out of the box.  It shouldn't be
> > > needed anywhere, and it causes lots of problems today.
> >
> > I presume you mean IPv6! If so, what problems? I have three Dapper boxes
> > and I've not experienced any problems with IPv6.
> >
> 
> Why would anyone presume this?
> 
> Alexander understood my point.  There is no good reason today why people are
> still using IPv4.

I can think of several:

- Common hardware drops essential IPv6 packets (I own and used to use
  a wireless access point that discarded IPv6 route advertisements,
  making it hard to use IPv6 on wireless nodes even on my private
  network)

- ISPs in much of the world only offer native IPv4 access

- Many services only offer native IPv4 access

Sure, IPv6 has some tunneling support that can get around the second
issue, and has backwards compatibility to IPv4 to get around the third
issue, but you are still using IPv4 in both of those scenarios.

> Advocating the use of IPv4 is like advocating use of a CLI only.

The important distinction being that lots of people have video cards
and pointing devices but not so many have native IPv6 connectivity.

There are lots of good reasons to encourage wider use of IPv6, but it
is extremely silly (and technically questionable, due to v4compat and
v4mapped behavior) to say IPv4 should be disabled out of the box.

Michael Poole