Wildcards in reverse DNS

[Mark Andrews]

> Karl Auer wrote:
> > On Thu, 2007-01-04 at 08:25 -0500, dhottinger at harrisonburg.k12.va.us
> > wrote:
> >> Great for the city.  But if we are an island in the middle of all
> >> ipv4 routers, all the traffic has to be encapsulated in ipv4 packets.
> >> Hence all speed increases are null because everything suddenly
> >> becomes ipv4 instead of ipv6.
> >
> > The islands will join up. It's a chicken and egg thing. IPv4 started
> > out as islands too, don't forget, and it had competition from many
> > other protocols, whereas IPv6 has competition really only from IPv4.
> >
> > People can win a lot from IPv6 without having IPv6 connectivity to the
> > Internet. Autoaddressing, IPSEC, no broadcasts, VAST private address
> > space, etc.
> 
> But what's the point of using it on a Private LAN? It just adds 
> confusion and lack of clarity. Not to mention for privat subnets, it'll 
> just be overkill. What's wrong with using 10.0.0.0 or 192.168.0.0 or any 
> other private subnet?

	There are not enough addresses in 10.0.0.0/8 for some private
	networks.  I'm aware of organisations which have outgrown
	all of the address space from RFC 1918.
 
> I find that the *appearance* of an IPv6 address is confusing in and of 
> it self. It actually look like an IPX address (maybe a cross between an 
> IPX address and a MAC address.) Don't get me wrong, I like HEX, but I 
> think it's nto such a good idea to change such a basic idiom... sice 
> TCP/IP came out, people have used DEC octets, any one wh oworks with 
> networks thinks DEC when thinking of IP addresses... I'd imagine IPv6 
> turns that way of thinking inside out and upside down.
> 
> Wouldn't it be better to have a better solution that could be somewhat 
> compatible with IPv4 so we don't have to get used to something so 
> radically different?

	Nibble is a presentation format.  For ip6.arpa is actually
	make delegation on arbitary bit boundaries easier.  It also
	provides twice as many natural break points than breaking
	on the octet.

	It's also easier to work out which addresses fall within
	a given mask and which don't.
 
> >>   I think the main reason they went with ipv6 was
> >> because of the availibility of ipv4 addresses.  Although NATING would
> >> handle the issue quite well.
> >
> > NAT is a Bad Thing for the Internet. It is a classic
> > treat-the-symptom-not-the-disease response, and while it has saved our
> > bacon for now, the waters are still rising...
> 
> I disagree. If you think NAT is bad then you don't know how to properly 
> use one. For home and busienss LANs, you can have one IP and share it 
> among the whole LAN? How is this a BAD thing? It's a money saver. I 
> suppose that's a BAD thing too.

	NAT works for a limited set of protocols.  There are lots of
	things that just won't work through a NAT.  That is a bad thing.

	If all you want is HTTP and SMTP then NAT kind of works.  Even
	then you have to work around the limitations of the NAT devices
	to make things work in all cases.

	As for being a money saver, it is a money saver in the IPv4
	world where address are scarse.  For IPv6 the default end
	user allocation is a /48 which gives 65536 /64 networks.
	Address space is being allocated to ISP's so that they can
	handout that much addresses space to their clients the
	way they hand out a single IPv4 address today.

> >>   I wouldnt think that every device would need a public ip.
> >
> > Even quite small organisations are running out of *private* address
> > space. IPv6 delivers a vast amount of private address space too...
> 
> Then they have the wrong class (or length) subnet. I dare you find a 
> small organization using a private 10.0.0.0 subnet and is running out of 
> them.
>
> > Don't fall for what Richard Dawkins calls the Argument from Personal
> > Incredulity. The fact that you don't see it, don't get it, don't
> > understand it, don't believe it, down't want it or don't need it
> > doesn't prove anything.
> 
> Thats exactly what you're doing with NAT, while ignoring it's obvious 
> benefits.
> 
> > We cannot imagine what things the future will
> > dream up to do with almost unlimited address space.
> 
> I don't deny it, just probably not with IPv6 as we know it.
> 
> >>   Also IPv4 addresses were handed out quite willy
> >> nilly.  Some institutions own huge blocks of addresses and dont use
> >> them.  I have 3 class C's and only use a fraction of them.  But, I
> >> wont give them up.
> >
> > That's the problem, thanks for being part of it. There are people
> > sitting on largely empty /8 (!) and /16 networks who won't give them
> > up either.
> 
> Sooner or later they will probably need to give part of them up.
> 
> You're also forgetting that private space is completely seperate form 
> public (Internet) space. Most large chuncks of network space are used by 
> ISPs, and hosting companies. Private addresses are only visible to the 
> private network. Anyone can sue thme on their own networks. IE, 
> 192.168.0.0 or 10.0.0.0 can and are used on man MANY private networks.

	And that causes all sorts of problems due to address ambiguity.
 
> >>   Although my ISP is really eager for me to give
> >> some up.  If the internet continues to grow, IPv6 will just be a
> >> stopgap measure.  Those addresses are not infinite.
> >
> > No, and there are already disturbing signs both of incompetence in
> > applying for stupidly large spaces and worse, of incompetence in
> > allocating stupidly large spaces. And of outright land-grabbing. The
> > US Government, for example, wanted an IPv6 /8 network. All for itself.
> 
> How the hell do you eve ndefine how big an IPv6 /8 network is? Deos it 
> equate to an IPv4 /8 or is it everything up to the last octet? The way 
> IPv6, it's eanything but clear and this is one of the many problems that 
> seems to be stiffling IPv6 and why most just don't use it.

	A individual IPv6 address is a /128.  A net the size of a IPv4
	/8 is a /102.
 
> > We will need to be *extremely* profligate with addresses to put a dent
> > in that vast address space. Sadly some people are already being
> > extremely profligate.
> 
> You you don't think any of it at all is at least partly do to the 
> inherently confusing nature of IPv6? (At least when compared to IPv4.) 
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org