Abstract: Routing in ad-hoc net works is a difficult challenge that involves a tradeoff between efficiency and responsiveness. An ad-hoc network routing algorithm must adapt rapidly enough to topology changes to meet the performance demands of users, without over-utilizing network resources. This paper presents the (α, t)-Cluster-frame work which utilizes a distributed dynamic clustering strategy to organize nodes into clusters in which the probability of path failure due to node movement can be bounded over time. The objective of the clustering strategy is to achieve scalability and support robust, efficient routing subject to a wide range of mobility rates. Based on the (α, t)-Cluster scheme, routes within clusters are maintained on a proactive basis; whereas, hierarchical routing between clusters is managed on a demand-basis. Simulation results show that the cluster organization can be effectively adapted to node mobility and that routing is both more robust and efficient than routing in fully proactive, reactive or fixed-hybrid schemes.