The averaging algorithm proposed by Lamport et al. works for a completely connected network of clocks. Will such an averaging algorithm for clock synchronization work on a cycle of n clocks, of which...

The averaging algorithm proposed by Lamport et al. works for a completely connected network of clocks. Will such an averaging algorithm for clock synchronization work on a cycle of n clocks, of which m can exhibit two-faced behavior and n > 3m? Assume that each link in the cycle allows bidirectional communication.