Question on how Virtual Connect do loop detection and avoidance.

[chuckk281]

Chris had loop avaoidance question on Virtual Connect;

******************

I am being asked again how VC modules prevent a loop when VC modules are in Active-Active mode and the customer is using Linux NIC bonding.

They have explained that the MAC address for eth0 is visible through VC-B and are asking what Loop prevention mechanism the VC modules are using?

Is the MAC address being transported across the VC module interconnects in this scenario?

They have a LAN defined that includes VC0 and VC1 and eth0.

****************

Lee replied:

******************

There’s an internal loop avoidance mechanism.  This is what is really does to the tags and a visual representation of each scenario.

–      We will never activate 2 uplinks connected to the same vNet

–      vNet can be thought of as vlan in traditional switching

              > We put a second vlan tag on packets traversing the virtual connect networks.

                        - We do not modify the original vlan tag

              > We are a layer-2 device; we cannot pass traffic between vlans

******************

Hope that helps. Questions?