Linux source address selection vs. EUI-64

Brian E Carpenter brian.e.carpenter at
Sun Nov 14 04:35:37 CET 2010

On 2010-11-14 04:32, Eugen Leitl wrote:
> On Sat, Nov 13, 2010 at 03:36:57PM +0100, Geert Hendrickx wrote:
>> For individual hosts (esp. in a VPS environment), assigning a /64 or
>> larger makes little sense to me, a /96 is more than enough.
> Is that an official recommendation? I currently have a single /56,
> which I would like to distribute over several thousands customers,
> each on a virtual server, with currently one static IPv4 address.

It is certainly not. /64 should really be the longest. Firstly
it's assumed by SLAAC and secondly you can't possibly know whether
a customer wants to subnet. A /56 would be a better default per
customer, in case they want multiple subnets. How can you possibly
tell what a customer will do if they hook up building automation
systems for example.

This is why we designed IPv6 in the first place!

*Please* see
which is on the way to being a recommendation.


> I can see how a /96 for each customer would be more than enough
> (especially that since hundreds of customers are going to share
> the same MAC, due to the low-overhead virtualization technology
> used). 
> Presumably, in above context it would be a sane thing to parcel 
> the /56 into /64, each for one physical server's distinct NIC 
> MAC, and from each /64 a /96 for each virtual server?

More information about the ipv6-ops mailing list