Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

Wikipedia states:

Unicast and anycast addresses are typically composed of two logical parts: a 64-bit network prefix used for routing, and a 64-bit interface identifier used to identify a host's network interface.

Can you use more than those 64 bits to identify a network - or is this not in compliance with any RFC? Also: Can you use more than 64 bits to identify a host?

I know that those questions are probably not of any practical use, since 2 ^ 64 addresses will be enough for most networks. I am just curious.

share|improve this question
2  
It says typically - of course other prefix lengths would be as valid as well. See RFC 4291 section 2.3 for reference - it does not impose any limits on prefixes. –  the-wabbit Dec 18 '11 at 14:34
    
@syneticon-dj thanks for your fast comment. If you turn this into an answer, I'll gladly accept it –  yas4891 Dec 18 '11 at 14:42

3 Answers 3

It is possible, but only recommended in special situations. One example is using a /127 on point-to-point links. I have also seen /48 prefixes on a datacenter colocation LAN, where each customer is allowed to use a /64. Weird, but it happens...

share|improve this answer

It says typically - of course other prefix lengths would be as valid as well. See RFC 4291 section 2.3 for reference - it does not impose any limits on prefixes.

share|improve this answer

Some IPv6 implementations allowed netmasks to be smaller than /64, but the last I saw those died in the 1990's. The modern ones only allow /64 because of RFC and how auto-addressing works. The network part is always 64-bits. When Comcast and company get around to passing out IPv6 to customers, they'll likely be passing out a /64 when they do. When I go to my upstream and ask for an allocation I'll probably get a /48, which I'll then chop into /64's for each subnet I need.

How the addresses end up written in the network documentation is another matter though.

share|improve this answer
    
so what you're saying is: The convention is 64 bits for interface ID and everyone expects it to be, but there is no formal document requiring this. Did I get that right? –  yas4891 Dec 18 '11 at 14:45
    
/64 is what you get asking for a public assignment. However, this is a regulatory decision of the RIRs, not a technical limitation of IPv6. The "formal document" for this is the address allocation policy - like in this example from RIPE –  the-wabbit Dec 18 '11 at 14:50
1  
@yas4891 Correct. In fact many small networks use /112 as that leaves the last hex group for the host ID (which gives somewhat familiar looking IPs for people used to /24 IPv4 subnets). –  Chris S Dec 18 '11 at 15:14

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.