From: Brian Dennis (bdennis@internetworkexpert.com)
Date: Thu Nov 08 2007 - 04:17:18 ART
It's actually a technical limitation with dense mode groups which is the
default mode Auto-RP uses for the cRP and MA groups. Dense mode packets
can not be received on an interface and then sent back out that same
interface.
--Brian Dennis, CCIE4 #2210 (R&S/ISP-Dial/Security/SP) bdennis@internetworkexpert.com
Internetwork Expert, Inc. http://www.InternetworkExpert.com Toll Free: 877.224.8987 Direct: +1.775.852.3995 (Outside the US and Canada)
Joseph Brunner wrote: > I have heard it's because the c-rp messages can't be forwarded by frame > relay spokes, or other nbma network where every end-point may not have a > full-mesh. (Which is why the HUB must be the MA). So if two people on spokes > of an NBMA network want to receive groups the other site is C-RP announcing, > they need a MA (who has pvc's, etc) to each site to do the C-CRP > "reflection" via the discovery messages. > > Can any else point me to some good rfc or tech doc goodie explaining this. > > Thanks guys! > > -----Original Message----- > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of > William Nellis > Sent: Wednesday, November 07, 2007 11:58 PM > To: Wollmann, Bruno RQHR; ccielab@groupstudy.com > Subject: Re: OT: Multicast Auto-RP > > Going many -> Few then Few -> Many gets around the o^N problem you'll have > in going many -> many... > > ie, why does OSPF work the same way w/ having routers advertise LSA to > DR/BDR only to have them readvertise it? Removes full mesh issues... > > Realistically... i dunno. Probably has to do w/ limited processing power at > the time AutoRP was invented, and being able to have beefier MA's do the > computations. By having RP's just say "hey, I'm an RP", you can have them > focus on doing RP stuff... like decapsulating registers. Then you have an MA > do calculatoins for (what could be many c-RP), then advertise results of > calculations out to all other routers who just follow orders. Otherwise, > every router has to do calculations. May not be too much of an issue > today... but... autoRP came out in the day of the 2500... > > Thats my theory, but yeah, i know what you mean. > ------------------------------------------------------- > r/s > William Nellis IV > nellis_iv@yahoo.com > > > ----- Original Message ---- > From: "Wollmann, Bruno RQHR" <Bruno.Wollmann@rqhealth.ca> > To: ccielab@groupstudy.com > Sent: Wednesday, November 7, 2007 9:29:19 PM > Subject: OT: Multicast Auto-RP > > The document I'm reading right now while studying multicast and auto-rp > says that "all Cisco routers automatically join the Cisco-discovery > (224.0.1.40) group in order to receive group-to-RP mapping information > being multicast by the Mapping Agents in the network". The mapping > agent then advertises the C-RP with the highest IP address as the RP. > > My question is, why bother with this step? Why don't the routers just > join the RP-announcement (224.0.1.39) group and learn the candidate RP's > right from the candidates themselves and then chose the one with the > highest IP address on their own. Why not learn it right from "the > horse's mouth" rather than a mapping Agent? > > Anyone know the answer or have a logical explanation? > > If you want to read what I'm reading for yourself check out page #7 at > ftp://ftp-eng.cisco.com/ipmulticast/training/Module6.pdf > > thanks > Bruno > > _______________________________________________________________________ > Subscription information may be found at: > http://www.groupstudy.com/list/CCIELab.html > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > > _______________________________________________________________________ > Subscription information may be found at: > http://www.groupstudy.com/list/CCIELab.html > > _______________________________________________________________________ > Subscription information may be found at: > http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sat Dec 01 2007 - 06:37:28 ART