VLAN Translation Bridge - VTB

[config t]

Anyone seen a VTB deployed in the field?

I understand the concept (I think). For example, you can turn VLAN 3 traffic into VLAN 30 traffic or whatever else you wanted to 'translate' it to.

But how does it work

[Reggle]

Done it multiple times: over VPLS using Metro switches or ASR, over Data Center LAN, using ASR or Nexus 5600. What exactly do you want to know? Basically it takes a VLAN tag with number X and puts it in VLAN Y.
Any specific deployment you need it for?

[Dieselboy]

Have not used it however we did use QinQ which put one or multiple VLANs inside another VLAN (tunnel). The packet had two 802.1Q tags.
To do this we had small crossover cables between two physical switch ports. One port had the outer VLAN as an access port and the other port had a trunk port. There were some other commands to tell the switch it was qinq.

I mentioned this because technically you could have switchport access vlan 3 on one side, and switchport access VLAN 30 on the other side. This will swap VLANs like you describe. Might throw syslog errors though, and be careful of spanning tree.

Going back to QinQ, we moved up to 4948E switches which allowed us to use one etherchannel to do all customers encapsulation/decapsulation. Previously on the 3750 we needed 2 ports per customer.

I don't know how VTB works but might be something similar ?

[RoDDy]

Not sure if this is the same thing but I am familiar with QinQ tunneling and VLAN translation.

I work for a service provider and we use QinQ to basically preserve the use of VLANs across the network and also to give customer traffic a level of separation to cut down on the number of broadcasts they would be getting.

For example we have a Service VLAN or S-VLAN which is used for HSI (high speed internet). The traffic is encapsulated with an 802.1Q tag lets say VLAN 300 (service vlan assigned for delivery of HSI). Each customer has a unique VLAN assigned to their router/ONT called the customer vlan (C-VLAN). Traffic destined to the internet leaves the customers device with the C-VLAN tag heading towards the OLT (optical line terminal) gets the S-VLAN tag added in front of the C-VLAN tag and heads towards the core. On returning to the customers device from the other direction, the traffic reached the OLT with both tags and the S-VLAN is stripped, leaving just the C-VLAN, so the traffic can be sent in the C-VLAN.

In the case of VLAN translation, instead of adding the S-VLAN tag in front of the C-VLAN tag for traffic coming from customers, what happens is that it actually replaces the C-VLAN tag completely and leaves the S-VLAN tag in its place. For traffic returning with just a S-VLAN tag, it would remove the tag and leave the traffic untagged i believe.

Not sure if this explanation helps in anyway. Maybe i don't understand the question fully

-Roddy

[config t]

Totally forgot about this post..

My life was hell up until about a month ago.

The idea was to implement a VTB at the top of the Layer 2 campus architecture because the customer neglected to migrate to a mandated VLAN schema as per the standardized design we were tasked with implementing, so VTB was to be a temporary workaround until they got it together.

I say "was" because the project was nixed. Turns out attempting to install a 10Gb network on top of an un-tested, un-certified, un-documented, poorly terminated, and in most places not even labeled.. (deep breath) fiber infrastructure was a bad idea. Jesus I still have nightmares about that project.

Anyway.. I'm still unsure of how it was supposed to be configured. They made another fatal decision of keeping their existing VLAN scheme (all VLAN's go everywhere! Yay!) before we could implement it.

[deanwebb]

All VLANs go everywhere? Really? Just a sec, let me get something that will help with that...

[Rllavona13]