Good Afternoon, For Lab 2, STOPANDWAIT2 has a variable called ber, set to 0.00005, is this the probability of frame corruption? If not, how is frame corruption handled in this topology? I've managed to get the Negative Acknowledgement working (or at least, when a corrupt frame is identified, it sends one, and the message is resent), but every time a message that has a corrupted frame is noted, the resent message will also be corrupt, and I'm not sure if that's on my end yet. Cheers

"Cameron Locke" <21*3*2*2@s*u*e*t*u*a*e*u*a*> wrote: It stands for 'bit error rate' - the probability that one-or-more bits in a frame will be corrupted. Yes, there's an additional problem that needs to be solved. It's easy to detect a damaged frame, and send back a negative-acknowledgement, but what needs to happen if that negative-acknowledgement frame is damaged?

On looking further, I assume ber is Bit Error Rate. That doesn't explain why one failed checksum will always resend into another failed checksum, so I'll keep looking.

I would assume that to avoid the two generals problem, corrupt acknowledgment frames are ignored after a point (for simplicity, if not efficiency, I've been ignoring them entirely and allowing them to timeout). But I'm still not sure why a resent corrupt message is always corrupt. Thanks for the quick reply!