CUCM Maximum Node Separation Distance 2

[libellis]

We know that the servers in a CUCM cluster must not have more than a 40msec communication delay between them (80msec round trip). With only transmission delay present, the Cisco SRND puts an upper limit separation distance of 3,720 miles, but warns that in practice this will be much less due to equipment introduced delays (buffers, etc.).

What sort of equipment introduced delays should be expected on MPLS links?

If maximum equipment delay is unpredictable and varies considerably, are there any conservative approaches (e.g. keep distance less than 2,000 miles, etc.)?

 

[trvlr1]

I think you will find it is usable if your round trip is in spec. Distance has little to do with it. A short distance can give you issues if the route is not fast enough, or there are enough interruptions to the flow to slow it down, ie routers switches etc at both ends of your route, plus the telco eqipment and even the bandwidth available. The bottom line is to ping it if you can and see what your time is for the round trip.

 

[libellis]

Delay through fiber is approx. 8 millisconds per thousand miles. At, for example, 3,000 miles separation, there is a built-in 24ms delay. That leaves 16ms for equipment. Also, delay guarantees in service provider SLA's don't seem to encourage node separation either.