1: \begin{abstract}

2: Herman's algorithm is a synchronous randomized protocol for achieving

3: self-stabilization in a token ring consisting of $N$ processes.  The

4: interaction of tokens makes the dynamics of the protocol very

5: difficult to analyze.  In this paper we study the expected time to

6: stabilization in terms of the initial configuration.

7: 

8: It is straightforward that the algorithm achieves stabilization almost

9: surely from any initial configuration, and it is known that the

10: worst-case expected time to stabilization (with respect to the initial

11: configuration) is $\Theta(N^2)$.  Our first contribution is to give an

12: upper bound of $0.64N^2$ on the expected stabilization time, improving

13: on previous upper bounds and reducing the gap with the best existing

14: lower bound.  We also introduce an asynchronous version of the

15: protocol, showing a similar $O(N^2)$ convergence bound in this case.

16: 

17: Assuming that errors arise from the corruption of some number $k$ of

18: bits, where $k$ is fixed independently of the size of the ring, we

19: show that the expected time to stabilization is $O(N)$.  This reveals

20: a hitherto unknown and highly desirable property of Herman's

21: algorithm: it recovers quickly from bounded errors.  We also show that

22: if the initial configuration arises by resetting each bit

23: independently and uniformly at random, then stabilization is

24: significantly faster than in the worst case.

25: \end{abstract}