<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Hi Arie,<div><br></div><div>I've read that document quite a few times in the past trying to solve this :) about the only difference I do is I use the "legacy" av pair according to the docco, although im sure I tried the sub-qos-policy way too with no difference.</div><div><br></div><div>Essentially I create the child CBWFQ policy I then create a parent policy with a shape average percent and then the child service policy under it, the PPP user is assigned a bandwidth and service policy via RADIUS through authentication, I am able to replicate time after time that saturating the link with the shape average percent set well below the full capacity will only show a "shaping active - yes" when ppp multilink is used.</div><div><br></div><div>I've tried this on 1760/2801/7401/720x with various flavours of IOS in between, is there anything obvious you think I could be missing here, given it works absolutely beautiful with ppp multilink and not so beautiful without it, worth noting that the CBWFQ classes will still match without ppp multilink but the QoS is ineffective given the lack of artificial bandwidth restraint on the ppp session.</div><div><br></div><div>Cheers</div><div><br></div><div>Ben</div><div><br><div><div>On 13/07/2008, at 1:11 AM, Arie Vayner wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div>Ben,</div> <div> </div> <div>Not sure why you really need multilink PPP for shaping to work on virtual-access PPP sessions...</div> <div> </div> <div>Take a look at "Per-Session QoS", which is supported since 12.2(28)SB</div> <div><a href="http://www.cisco.com/en/US/docs/ios/12_2sb/feature/guide/sbbbrs1b.html">http://www.cisco.com/en/US/docs/ios/12_2sb/feature/guide/sbbbrs1b.html</a></div> <div> </div> <div>Arie<br><br></div> <div class="gmail_quote">On Sat, Jul 12, 2008 at 1:05 AM, Ben Steele <<a href="mailto:ben@internode.com.au">ben@internode.com.au</a>> wrote:<br> <blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Not really, the same principles apply, seeing as you are using vrf's are you planning on running MP-BGP with MPLS for this? in my experience EIGRP seems to integrate the nicest out of all the IGP's into this scenario, but anything will work with proper design.<br> <br>Also if you plan to run QoS on your L2TP ppp sessions then you will probably want to look at enabling multilink ppp, last I checked(12.4(18) 7206) this was still required in order to get QoS to work effectively, without it no shaping would occur on the virtual-access(on the lns), I really wish they would fix that up... <div class="Ih2E3d"><br><br><br>Jose Bejarano wrote:<br> <blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Hi,<br><br>we are terminating via L2TP our PPPoE users on a 7206vxr with npe-g1.<br>Customers are then terminating these sessions in a VRF using Radius attributes.<br> Now we are planning to introduce dynamic routing functionality: RIP,OSPF,BGP...<br>Any info/experiences/ideas/tips with using routing protocols over L2TP ?<br><br>Thanks in advance,<br>Jose<br> <br></blockquote><br></div> <div> <div></div> <div class="Wj3C7c">_______________________________________________<br>cisco-bba mailing list<br><a href="mailto:cisco-bba@puck.nether.net" target="_blank">cisco-bba@puck.nether.net</a><br><a href="https://puck.nether.net/mailman/listinfo/cisco-bba" target="_blank">https://puck.nether.net/mailman/listinfo/cisco-bba</a><br> </div></div></blockquote></div><br></blockquote></div><br></div></body></html>