The MPLS-OPS Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] Re: Tag-switching mtu doubt
On Thu, Mar 14, 2002 at 07:09:28PM -0000, Rui Francisco Negrão Valente wrote: > Hi all, > > Imagine the following VPN scenario: > > CE----PE---P---P---PE----CE > > I have FastEthernet in all the links, and tag-switching mtu 1508 at the PEs > and Ps interfaces. > > Due to the label stacking(2 labels) we have additional 8 bytes in the MTU > size. So, at the CPE the maximum datagram size I can send is 1500 (What I > did was pings with 1500 and with the DF=1). > > My question is: > > Why pings fail with datagram size of 1504 from PE to PE? > From what I understand there will be only one label (4 bytes) pushed in the > packet, therefore the mtu size will be 1508... The "tag-switching mtu xx" only affects mpls encapsulated frames. If you check the IP MTU of your interface, it will be 1500 even with the increased tag-switching mtu, so sending IP frames larger than 1500 bytes will fail. The sending PE should drop the packet, but even if that doesn't happen, the last P router will (most likely) do penultimate hop popping and thus have to drop the oversized IP frame. router1#sh int fa8/0/0 | in MTU MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec, router1#sh run int fa8/0/0 | in mtu tag-switching mtu 1508 > Rui Valente /Niels Chr. -- Niels Christian Bank-Pedersen, NCB1-RIPE. Network Manager, TDC, IP-section. "Hey, are any of you guys out there actually *using* RFC 2549?" ------- The MPLS-OPS Mailing List Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
|
|