The MPLS-OPS Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] RE: penultimate hop popping!!!
As per the RFC3031, LSR 4 is not prohibited from advertising IMPLICIT NULL label to LSR 3, for a FEC which covers a destination on or behind LSR 5. Robert's description is probably a specific method of distributing IMPLICIT NULL label. As per the RFC 3031, if an LSR is egress or proxy egress for a FEC, it may advertise IMPLICIT-NULL to the upstream peer. The following two statements from RFC3031 clarify the above "A LSR MUST NOT request a label distribution peer to pop the label stack unless it is capable of doing so." "An LSR which is capable of popping the label stack at all MUST do penultimate hop popping when so requested by its downstream label distribution peer." --------------------------------- HariKishan Desineni, Software Engineer, IP Access, BU BMSN, Ericsson, 70 Castilian Drive, Santa Barbara, California - 93117 U.S.A Office:(805) 562 6419 Cell:(805) 252 2641 Home:(805) 562 8841 Fax: (805) 685 4465 > -----Original Message----- > From: Robert Raszuk [mailto:raszuk@cisco.com] > Sent: Tuesday, February 19, 2002 4:13 PM > To: Mohammed > Cc: mpls-ops@mplsrc.com; mpls@UU.NET; ahussain@eng.utoledo.edu > Subject: Re: penultimate hop popping!!! > > > Mohammed, > > > LSR 4 knows that it is the egress because its > > adjacent is not LSR. But LSR 3 ( penultimate node for > > that LSP)has to do penultimate hop popping .. but it > > has no way to know about this unless informed by LSR > > 4. > > Well LSR 3 has to do PHP only if LSR 4 advertised to him > implicit-null > label for a given FEC. > > > QUERY : How does LSR 4 inform LSR 3 that it has to do > > penultimate hop popping ... DOES it gives LSR 3 an > > IMPLICIT NULL LABEL. > > It does but only for connected routes (at least in one > implementation > :). So in your case if you are pinging the R5's interfaces > towards the > R4 sure LSR3 will do PHP and in the forwarding table for > this FEC you > will see "Pop tag". > > On the other hand if your destination is on R5 (loopback) > or behind R5 > LSR 3 will not do any PHP. LSR 4 entry in the forwarding > table will be > untagged and by removing the label the packet will be send to LSR5 > unlabeld. > > Also it is probably worth of noting that the difference between "Pop > tag" & "untagged" is that pop tag pops one label on the > stack and sends > resulting labeled or unlabeled packet out while untagged removes all > labels from the stack and if the number of removed labels > is one the IP > packet is send if it is bigger then one packet is discarded. > > Rgs, > R. > > >Mohammed wrote: > > > > Hi all, > > > > I have a query regarding penultimate hop popping. > > Please refer to the following figure .. > > > > > > 0 --- 1 --- 2 --- 3 ---- 4 --- 5 > > ---------- ---------- > > MPLS DOMAIN. > > > > Lets assume that nodes 1,2,3 and 4 form the MPLS > > domain. Lets say node 0 and node 5 are ip nodes. > > > > For traffic generated by node 0 and destined to node > > 5 the path would be 0->1->2->3->4->5 where 1-2-3-4 is > > an LSP. > > > > LSR 4 knows that it is the egress because its > > adjacent is not LSR. But LSR 3 ( penultimate node for > > that LSP)has to do penultimate hop popping .. but it > > has no way to know about this unless informed by LSR > > 4. > > > > QUERY : How does LSR 4 inform LSR 3 that it has to do > > penultimate hop popping ... DOES it gives LSR 3 an > > IMPLICIT NULL LABEL. > > > > Any pointers in this area are welcome. PLease also > > suggest other points that i may be missing. > > > > Thanks to one and all! > > > > Sincerely, > > Abrar. > > > > ===== > > Abrar Hussain. > > Graduate Student/Network Designer. > > The University of Toledo. > > Toledo,OH. > > > > __________________________________________________ > > Do You Yahoo!? > > Yahoo! Sports - Coverage of the 2002 Olympic Games > > http://sports.yahoo.com > > > > ------- > > The MPLS-OPS Mailing List > > Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml > > Archive: http://www.mplsrc.com/mpls-ops_archive.shtml > > ------- > The MPLS-OPS Mailing List > Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml > Archive: http://www.mplsrc.com/mpls-ops_archive.shtml ------- The MPLS-OPS Mailing List Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
|
|