<html><head><meta http-equiv="Content-Type" content="text/html charset=iso-8859-1"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><div>On 6 Sep 2013, at 04:58, Lorenzo Colitti <<a href="mailto:lorenzo@google.com">lorenzo@google.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div dir="ltr">On Fri, Sep 6, 2013 at 1:38 AM, David Magda <span dir="ltr"><<a href="mailto:dmagda@ee.ryerson.ca" target="_blank">dmagda@ee.ryerson.ca</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">This would be why I would lean towards an DHCP-based solution: you<br>
configure certain subnets/prefixes to have "random" addresses assigned and<br>
others to have MAC-based ones (or 'static-y' reservations). You'd keep the<br>
assignment logs around for some period of time.<br></blockquote><div><br></div><div>You don't need to do this by giving the host different addresses in different prefixes. What you want to do is pick different source addresses based on what destination you're talking to.</div></div></div></div></blockquote><div><br></div>Indeed, I agree it's a source address selection thing.</div><div><br><blockquote type="cite"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">
<div>For example, suppose a given host has both 2001:db8:0:1::12 (static) and 2001:db8:0:1:d723:fa68:bb1a:0743 (privacy). You would want to tell the host to use the static address for your prefixes and the privacy address for everything else.</div>
<div><br></div><div>You can do that using the privacy preference flag defined in draft-ietf-6man-addr-select-opt-11.<br></div></div></div></div></blockquote><div><br></div>I'll have to go back and look now it's been around the IESG, but from memory the privacy flag in that spec is global, not per prefix. In an earlier version it was per prefix, e.g. to 'forbid' privacy addresses on ULAs, but I recall that was reverted based on WG feedback.</div><div><br></div><div>There was an I-D proposing a privacy flag for RAs, which was per prefix, but that was dismissed quite quickly by the WG, see <a href="http://tools.ietf.org/html/draft-gont-6man-managing-privacy-extensions-01">http://tools.ietf.org/html/draft-gont-6man-managing-privacy-extensions-01</a>.</div><div><br><blockquote type="cite"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"><div>Whenever you feel yourself wanting to use ULA and DHCP, bear in mind that's almost always the wrong answer. :)</div>
</div></div></div>
</blockquote><br></div><div>Which is a different thread :)</div><div><br></div><div>Tim</div><br></body></html>