2 ms·
Yeah that is basically what I was imagining, but there's no failure condition that matches that in reality. You might get unlucky and fail to converge for 10 i
by chubot 12y ago
Yeah that is basically what I was imagining, but there's no failure condition that matches that in reality. You might get unlucky and fail to converge for 10 iterations, but not for say 10,000.
It is sort of Byzantine in that the "system" is inserting very specific message losses with knowledge of the algorithm.