clustering schemes may cause the cluster structure to be completely rebuilt over the whole network when some local events take place, e.g. the movement or “die” of a mobile node, resulting in some cluster head re-election (re-clustering). This is called the ripple effect of re-clustering. In other words, the ripple effect of re-clustering indicates that the re-election of one cluster head may affect the structure of many clusters and arouse the cluster head re-election over the net- work [18]. Thus, the ripple effect of re-clustering may greatly affect the performance of upper-layer protocols. In addition, most schemes separate the clustering into two phases, cluster formation and cluster maintenance, and assume that mobile nodes keep static when cluster formation is in progress. This is because for the initial cluster formation of these schemes, a mobile node can decide to become a cluster head only after it exchanges some specific information with its neighbors and assures that it holds some specific attribute in its neighborhood. With a frozen period of motion, each mobile node can obtain accurate information from neighboring nodes, and the initial cluster structure can be formed with some specific characteristics. However, this assumption may not be applicable in an actual scenario where mobile nodes may move randomly all the time.