<div class="gmail_quote">On Wed, Apr 29, 2009 at 4:25 PM, Matthew Moyle-Croft <span dir="ltr"><<a href="mailto:mmc@internode.com.au">mmc@internode.com.au</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div style=""><div class="im"><br><div><div>On 29/04/2009, at 2:31 PM, Matthew Moyle-Croft wrote:</div><blockquote type="cite"><div style=""><div><br></div><div>However, a per-user VLAN might be nice too if the aggregation can be sorted out (imagine BIG vlan rewrite switches!).</div>
<div><br></div></div></blockquote><br></div></div><div>There are precedents for doing this - some countries moved to this model from ATM PVC per customer.</div><div><br></div><div>However, it does strike me that a per vlan model would burn a /30 for each customer rather than a /32 - in a world where IPv4 addresses might become quite valuable for a time do you really want to increase a cost by four times just to get rid of PPPoE? </div>
</div></blockquote><div><br>I can't speak for other vendors, but on Redback* you can use a common DHCP pool with subscribers spread over multiple VLANs.<br><br>Another interesting issue that I have seen is where multiple CPEs have the same MAC address ('cheap' CPEs in particular). If the subscribers are in individual VLANs you can handle it ok but if you have a common shared VLAN it can cause some headaches :)<br>
<br>Cheers,<br><br>Rob<br><br>* Day job disclaimer, speaking for myself on this list.<br></div></div>