How much latency is considered acceptable?

[a_cutler2]

I'm looking for some insight on latency with an on-prem App Broker implementation. A customer has server space in London for a US based implementation, but is concerned because we would have to accept 120ms of latency between London and where the SCCM resides (houston). I'm not sure how much of an impact this will have on performance, can anybody weigh in on this?

[jdempsey]

That's a very subjective question! 🙂  I couldn't begin to speculate what the performance would look like in that scenario.  We always recommend co-locating App Broker, FNMS, and SCCM in the same data center, and I don't recall having ever done otherwise in any of our implementations.