[tahoe-dev] [tahoe-lafs] #867: use ipv6

Randall Mason clashthebunny at gmail.com
Sun Feb 17 22:03:52 UTC 2013


On Sun, Feb 17, 2013 at 3:07 AM, David-Sarah Hopwood <
david-sarah at jacaranda.org> wrote:

> On 16/02/13 21:05, Eugen Leitl wrote:
> > On Sat, Feb 16, 2013 at 03:14:54PM +0200, Randall Mason wrote:
> >> On Fri, Feb 15, 2013 at 9:06 PM, Greg Troxel <gdt at ir.bbn.com> wrote:
> >>
> >>> Also, fe80:: addresses should probably be ignored, as they are meant to
> >>> be used only on a single link.
> >
> > Please do not forget http://en.wikipedia.org/wiki/Cjdns and related,
> which
> >
> > cjdns addresses are the first 16 bytes (128 bits) of the SHA-512 of the
> > public key. All addresses must begin with the byte 0xFC, which in IPv6
> resolution,
> > is a private address (so there is no collision with any external
> Internet addresses).
>
> The prefix FC00::/7 (i.e. first byte 0xFC or 0xFD) is reserved for
> "unique local addresses" (see
> http://en.wikipedia.org/wiki/Unique_local_address
> and RFC 4193). However, I don't think that the arguments that Greg Troxel
> put
> forward apply to such addresses, in general. It's quite plausible that a
> Tahoe-LAFS grid could be made up entirely of nodes that are inter-routable
> using FC00::/7 addresses.
>

A little test seems to confirm this:

> $ ping6 fc00::1234
> PING fc00::1234(fc00::1234) 56 data bytes
> --- fc00::1234 ping statistics ---
> 4 packets transmitted, 0 received, 100% packet loss, time 3000ms

works, whereas:

> $ ping6 fe80::1234
> connect: Invalid argument

falls over hard.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://tahoe-lafs.org/pipermail/tahoe-dev/attachments/20130218/4869ea84/attachment.html>


More information about the tahoe-dev mailing list