Not tested Nstreme2 myself yet but very interested to hear results, we want to offer this too.
Frequencies more than 1 apart (no interference due to ACR limits) ?
Separate antennas (not mixing them onto one)
CPU usage ?
Length of link (does that count)?
MT people: have you got any (comparitive or otherwise) test results for long distance links, Nstreme and Nstreme2?
I have posted my test results with nstreme2 a while ago. I am planning some more tests on new non-live links (that would not have the real world traffic and thus provide better results).
Stephen: I agree with your comments on nstreme2 requirements. My experiments so far suggest that the lenght of the link is not concern - except as signal attenuates. My longest nstreme2 link so far is about 85km.
CPU usage seems to be lower than that of nstreme1, BUT as I mentioned, I have yet to confirm this with isolated tests.