Hi Kiran,

That's indeed what is going to happen with Cisco bridges. We tweaked the MST implementation so that we don't send the information for an instance whose mapped vlan don't appear on a given port. For example, there is no vlan mapped to instance 2 allowed on the access port in vlan 10. As a result, we don't send the information for instance 2 over this link. This way, MST 2 does not consider this access link in vlan 10 as a redundant link between the two 3550 and you are sure that the other link (access for vlan 20) will be forwarding. This kind of trick cannot apply to instance 0, that is computed on every port, even access port (thus instance 0 will probably block on g0/2). The goal was to get a behavior closer to PVST so that our users (who are generally used to PVST) are not too confused by MST. I'm not sure if it's a good idea in the long term;-)

If the same scenario was made with two third party bridges, one link would block for all the instances and there would be no connectivity between the two switches for one of the vlan.

Regards,

Francois