IPv6 Address Planning

[Iljitsch van Beijnum]

On 16-aug-2005, at 17:08, Lars Erik Gullerud wrote:

> Accepting that you'd need space for 128 reserved anycast addresses  
> on a p-t-p link is not something I'm prepared to do, I prefer to  
> remain unreasonable. :)

I don't think this will cause problems, but it's not impossible that  
at some point a router vendor screws this up and uses one or more of  
these anycast addresses even though the subnet isn't large enough for  
it. I wouldn't worry about this with our collective favorite vendors  
starting with C and J, because they both have reasonably mature IPv6  
implementations and have many ISP customers, but you may want to make  
sure this isn't an issue for a new IPv6 router implementation when  
you evaluate one.

>> it don't juts work, the /128, /127, /126 is a good exampe of that,  
>> due to
>> change a /127 isn't usable over night :}

> /127 still seem to work just fine on my gear, fwiw.

But only because your router vendor doesn't implement the standard (=  
all zeros is all routers anycast address). Others do. /127 is a very  
bad idea.

>> the only way you can be guaranatied it will work are to use /64,  
>> or /112
>> as everyone claim is okay...

> Or, if it turns out "everyone" is actually using longer prefixes,  
> then the vendors will make sure to support it.

Unless I missed something, IPv6 is supposed to be classless, so any  
prefix length shorter than 127 bits should work just fine. So there  
is really no issue here.

(However, RFC 3513 requires interface identifiers to be 64 bit for  
all IPv6 space except ::/3 and stateless autoconfiguration only works  
if the prefix advertised by routers + the interface identifier are  
128 bits combined, so stateless autoconf only works on /64 prefixes.)