Originally posted by: stephen2615
We are running two fabrics between our data centres and each fabric has 2 x 2 Gpbs ISL's on Brocade directors not running trunking mode. It seems obvious that our Cisco ONS DWDM solution is using a network with alternate paths as I believe one of the ISL's is about 8 km in length while the other is about 35 km. Our networking person (IP that is) does not think there is any problem with different lengths because he does not have to worry about latency in his 10 Gpbs networks. Besides he can throw together a port channel and forget about it.
The 35 km length is already pushing the latency boundaries with sync TC especially when Exchange sometimes has 15 ms random write response times without any of the overhead of TC. I hope the mismatched lengths in one fabric don't cause the problem I am seeing.
I have a horrible suspicion that the fabric with different lengths is causing havoc with our TC replication but I never seem to get any decent answer on it. Would anyone care to comment on this?