Only the port that does the actual double-tagging will need to be configured. From a 3750 I just set up for a customer:
system mtu 1998 system mtu jumbo 9000 ! interface FastEthernet1/0/1 description TAGGED UPLINK switchport trunk encapsulation dot1q switchport trunk allowed vlan 2302 switchport mode trunk switchport nonegotiate speed 100 duplex full ! interface FastEthernet1/0/4 description Customer switchport access vlan 2302 switchport mode dot1q-tunnel switchport nonegotiate l2protocol-tunnel cdp l2protocol-tunnel stp no cdp enable The switches need to be able to support Layer 2 Protocol Tunnelling (L2PT). Note this is different from Layer 2 Tunnelling Protocol (L2TP)... thanks Cisco! I don't believe the 2960s fall into this category Cheers Ross -----Original Message----- From: cisco-nsp-boun...@puck.nether.net [mailto:cisco-nsp-boun...@puck.nether.net] On Behalf Of CiscoNSP_list CiscoNSP_list Sent: August-17-12 2:03 AM To: td_mi...@yahoo.com; cisco-nsp@puck.nether.net Subject: Re: [c-nsp] QinQ...inQ? question Thanks Tony - So from a config perspective, can you please provide an example? i.e. From our existing (config) trunk port to carrier - What changes are needed? which ports will need to be changed to dot1q-tunnel? Date: Thu, 16 Aug 2012 22:11:45 -0700 From: td_mi...@yahoo.com Subject: Re: [c-nsp] QinQ...inQ? question To: cisconsp_l...@hotmail.com; cisco-nsp@puck.nether.net Hi, 802.1ad supports an arbitrary number of tags in the header (not limited to 2), so "in theory" it should work. There is obviously MTU considerations to be had (each extra VLAN tag adds another 4 bytes). What I don't know about is equipment support of it and whether this should pose any problems (again, in theory they shouldn't care, and have no need to look at anything after the first vlan header they are switching on). Why not just test it with a couple of extra switches yourself (to simulate your customers gear) and see what happens ? regards,Tony. From: CiscoNSP_list CiscoNSP_list <cisconsp_l...@hotmail.com> To: "cisco-nsp@puck.nether.net" <cisco-nsp@puck.nether.net> Sent: Friday, 17 August 2012 1:09 PM Subject: [c-nsp] QinQ...inQ? question Hi Guys, We typically get QinQ links from our upstreams for p-t-p links between our POPs - Upstream carrier does the QinQ and we simply configure a trunk port to them and tag whatever vlans we need. (i.e. dont need to engage upstream for the vlans we want to use) We have a new POP, with our standard QinQ hand-off from carrier(p-t-p link to another POP)...but we have one of our customers is also requesting QinQ also between the two POPs(From us) - Is this possible? Each end of the link is our equipment (2960 and a 3560) which connects to carrier...each of these switches connect to the customers switches.....note our switches are only doing L2 (All L3 is done on 7200's) Cheers. _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
